2023:Review User Guide

From Grooper Wiki
Revision as of 11:13, 17 May 2023 by Rpatton (talk | contribs)
WIP This article is currently under construction.

It was written for Grooper version 2022. Its guidance is mostly applicable to version 2023 as well.

Be aware:

  • Functionality was improved in 2023. There may be slight differences from information provided in this article.
  • Screen shots were taken from the 2022 UI. There may be slight differences from what you see in version 2023.

This article is for document review users using the Grooper Web Client to review Grooper's automated processing results. The Grooper Web Client allows users to connect to a Grooper dashboard over the internet via a web server. This allows end-users to process review based steps in a Batch Process in a web browser, without the need to install Grooper on their own machine.

Previous Versions

Grooper 2022

About

Welcome to the Grooper Web Client! The Grooper Web Client allows users to process documents using a web browser.

In the following sections, we will give end-users guidance on how to navigate the Web Client user interface and use it to process Batches to review documents. We will discuss the following topics:

Web Client UI

First, let's look at how to navigate the Grooper Web Client interface.

Access the Grooper Web Client by entering the URL provided by your Grooper administrator.

  • You may be prompted to enter user account credentials.
  • If you are not prompted, Windows may have automatically passed on your logon credentials.
⚠

The Grooper Web Client DOES NOT support Internet Explorer.

The following browsers are supported:

  • Microsoft Edge
  • Google Chrome
  • Apple Safari

Other modern browsers may work but have not been fully tested, such as:

  • Mozilla Firefox
  • Opera Web Browser

Upon entering the URL, you'll land at the Web Client's homepage. This page is divided into four main sections:

  1. Navigation Links
  2. Repository Info
  3. Recent Events
  4. Context Toolbar

Navigation Links

The Navigation Links section is the main way you'll get around in the Web Client. It contains a variety of links for Grooper users, including:

Design - Used to access and edit the repository.

  • From here, designers can add, delete, or change any part of the repository. This is where Grooper is configured for whatever documents need to be processed. Generally you want to restrict access to this area to individuals who will be building the Grooper repository.

Batches - Used to access a list of all current Batches in production.

  • From here, reviewers can see and select Batches in process. They can also filter Batches by a variety of search criteria, use a search function to search for Batches by keyword, and process user attended review activities.

Tasks - Used to access a list of review tasks ready for reviewers.

  • This is another way for end-users to select and start review based work via the Web Client. Only review tasks ready for processing will be presented to the reviewer. Reviewers can also filter review tasks by Batch, Batch Process, Step, or Queue.

Imports - Used to access a list of recent import jobs.

  • This is a way to view how the Import Watcher is working. You can see a description of the Import Job, information regarding the "Provider" of the import, listed priority levels, date submitted, and more.

Jobs - Used to access a list of the recent workflow.

  • This is an organized list to show what jobs have been submitted, worked on, and completed. Each entry lists detailed information such as the "Job ID", Batch Process being used for the job, which step of the Batch Process the job is currently in, and other important information.

Stats - Used to access a variety of statistics on Grooper's workflow.

  • This is a good way to see how work is going. You can perform various queries and retrieve statistical information that can give an overview of current work efficiency in Grooper.

Learn - Used to access Grooper University courses at learn.grooper.com.

  • This is an external resource for Grooper designers who have an active training subscription.

Wiki - Used to access our wiki site at wiki.grooper.com

  • If you're reading this you've already found our Grooper wiki! This is an external resource containing articles about a variety of Grooper topics.

Repository Info

The Repository Info window provides some "at a glance" processing statistics and information about your Grooper Repository.

A Grooper Repository is the environment in which processing resources are created and executed. This includes the Batches of documents themselves, the Batch Processes used to process them, and components used in the Batch Process such as Content Models.

This data displayed in the Repository Info window subdivided into three sections:

Totals

  • This is a running total of various aspects of the Repository, including the total number of published Batch Processes, total tasks in current and previous Batches in production, and total number of "nodes" (the processing objects Grooper architects create in Grooper Design Studio).

Tasks

  • This displays numbers regarding the review based activities for Batches in the Repository, including those ready for processing, those currently being worked on, and those that were previously completed.
  • This can give end-users a quick view of tasks awaiting review.

Nodes

  • This displays the total number of specific types of Grooper objects in the Repository.
  • This information will be most useful for Grooper architects working in Grooper Design Studio.

Recent Events

The Recent Events window is Grooper's event log.

This panel displays information regarding different processing events. This includes audit trails of processing events, such as Batch creation, task steps in a Batch Process submitted for processing, and Batch completion. This also includes warnings and error messages, giving you information about errors processing steps of a Batch Process.

This panel can be useful to track down information or a sequence of events if you're troubleshooting a problem.

FYI If you're familiar with the thick client version of Grooper Design Studio, this is essentially the same event log you see when selecting the root node of your Grooper Repository.

Context Toolbar

The Context Toolbar is a navigation bar providing various utility in the Web Client.

Depending on the context (which page you've navigated to), this menu will change slightly. However, please note wherever you are in the Grooper Web Client, clicking the house icon will always take you back to this home screen. You can also use the Context Toolbar to navigate to the Design, Batches, Tasks, Imports, Jobs, and Stats pages.

Click here to return to the top

Switching Grooper Repositories

Depending on the size and scope of your operation, you may be working out of multiple Grooper Repositories. If you are, you may need to switch between Grooper Repositories to access documents ready for processing in one or the other.

To do this, you'll use the Repository button on the homepage's Context Toolbar.

  1. First, the Grooper Repository you're currently working in is always displayed at the top of the homepage.
  2. To switch Repositories, click the Repositories button.

A new window called "Change Repository" will pop up.

  2. Select the Repository you wish to switch to from the list.
  3. Click "OK".

Upon making your selection you will switch to the selected Repository, granting you access to all the Batches and processing assets contained therein.

  2. You'll see the Repository listed at the top of the homepage has changed to the selected Repository.

Performing Review Tasks: The Batches and Tasks Pages

Documents come into Grooper either by scanning pages or importing files into a Batch. A Batch is the fundamental container of work in Grooper. It holds your documents as they are processed through Grooper. Along with the container comes a list of processing instructions called a Batch Process.

So a Batch is really two things:

  1. A container of documents in various states of processing.
    • These are represented as Batch Folders and Batch Pages contained in the Batch Root Folder.
  2. A step by step list of instructions of what to do with those documents.
    • This is the Batch Process.

A Batch Process will consist of automated tasks called Unattended Activities, as well as review-based activities requiring user intervention called Attended Activities. For end-users, most of your work will be centered around document review tasks (or Attended Activates). In these activities, you will review the automated work Grooper has done in previously in the Batch Process. For example, you may be reviewing the classification decisions Grooper made or reviewing Grooper's data extraction to ensure all data was captured accurately.

Different organizations will utilize human review to varying degrees. Depending on the use case, Grooper may be able to automate more work without the need for human intervention. However, as good as Grooper can be at making document processing decisions, no computer software can beat the human brain. Review tasks are well suited for situations where you need to ensure the accuracy of Grooper's results in one way or another. You play a critical role in verifying Batches are processed accurately through the steps of a Batch Process.

So, how do you get started?

There are two ways users can start processing review tasks in a Grooper Repository, either using the Batches or Tasks pages. Either is acceptable. These present two different ways of displaying available work in Grooper. We will start by reviewing the Batches page.

Batches Page

The Batches page will present a user interface to select Batches currently in production within the Repository. Users will be able to see the Batch's progress and process any human attended Review activity.

To get to the Batch page, click the Batches icon on the Grooper Web Client homepage.

  1. In the Navigation Links panel of the homepage, click the Batches link.

This will bring up the Batches interface. The first thing you'll see is a list of Batches currently in process.


You can sort the Batch List by the following properties:

  1. Batch
    • This column lists the name of the Batch. Often, this name will be related to the Batch Process used with a timestamp tacked onto the end.
  2. Process
    • This column lists the Batch Process assigned to the Batch. These are the step-by-step processing instructions given to the Batch.
  3. Step
    • This is the current step in the Batch Process being applied to the Batch.
  4. Activity
    • This is the current step's Activity type.
    • FYI: You can name a step whatever you want in a Batch Process. Many steps simply share the Grooper activity's name. However, for Review tasks in particular, you'll often find they are given a more descriptive name, describing the type of review you're going to do for that step.
  5. Status
    • This describes the state of the Batch's current step. This can be Working if the step is currently processing, Ready if the step is able to be processed and just waiting for a user to start it, or Paused if the whole Batch has been placed in a paused state, preventing any steps from being processed.
  6. Priority
    • This is the priority assigned to the Batch. Higher priority steps will consume system resources before lower ones, effectively processing first.
  7. Created
    • This is simply the date and time the Batch was created.
  8. Created By
    • This is the Grooper user who created the Batch.

If you have a particularly large number of Batches, you can narrow down what you're looking for using the search box or the filter utility.

  1. In the search box you can free search any text in the Batches, Process, Step or Activity columns.
  2. Or, you can select the Filter icon, to filter out Batches by certain criteria.
  3. This will bring up a window to filter out your selection based on Status, Process, Step or Activity.
  4. Click the arrow next to the property heading you want to filter by.
  5. Check the box next to the specific value you want to filter by.
  6. Click "Save" to execute the filter or "Cancel" to cancel.

Now that we've gotten the lay of the land, you're probably asking yourself how do I actually start doing work in Grooper? How do I start reviewing documents?

  1. First, select a Batch from the list.
  2. The "Progress" tab displays the current progress of the selected Batch.
  3. Each rectangle represents a step in the Batch Process.
  4. The step's name is listed under the rectangle.
  5. These numbers indicate how many tasks have been processed for the given step.
    • In this case there were 16 out of 16 total tasks completed in the "Separate" step.
    • You may have noticed that 16 tasks were completed for the "Separate" step, but only 15 were completed for the "Recognize" step. There are a total of 15 pages in this Batch. After Grooper determines how to separate the documents, there is an additional step that performs the separation.

What color the step is will indicate something about the steps processing status.

  1. Blue indicates the step's tasks were completed successfully (or without error).
  2. Grey indicates the step is ready for processing.
    • Either it's waiting its turn for steps before it, the Batch has been "paused", or in certain circumstances the step was skipped.
Red will indicate one or more tasks in the step have failed to process for one reason or another.
Green will indicate one or more tasks in the steps are actively being processed.

For end-users doing review work in Grooper, you will be processing steps with the "Review" activity type that are ready for processing.

  1. For these four batches, all are currently at a Review step in their process.
  2. The Batch we will be working with is listed as "Ready"
  3. With the Batch selected we can see in the "Progress" tab the step is grey, which also indicates it is ready for processing.
    • Both a step's Status listed as Ready and its color being grey mean the same thing. It's just two different ways of visualizing/understanding it's ready to go.
  4. To start the Review module, simply double click the Batch.

This will bring up the Review activity module to perform one kind of review or another, be it classification review, data review, image processing review or another. In Grooper, the different kinds of review applications are displayed as "Views". For example, the type of review this step is doing is classification review. The user is presented a "Classification Viewer" in order to verify each document in the Batch is classified correctly.

We will discuss how to use this "Classification Viewer" and the other "Review Views" later in the #Review Views section of this article.

For now, we're going to simply exit the review module.

  1. To exit without saving your work, click the "Stop" button to return to the Batches page.
  2. You can also use the Context Toolbar to navigate to a different page.

FYI If you're familiar with the Grooper Dashboard application in the Grooper thick client, the Batches Page should look very familiar to you. The interface is nearly identical, just with a different graphical skin.

There are, however, some things you can do in the thick client that you cannot do in the web client.

In the current version of the Grooper Web Client, you CANNOT:

  1. Process unattended activity steps in a Batch Process.
• You must have an Activity Processing service running to process unattended tasks.
  2. Perform "ad hoc" file imports using an Import Provider.
• Your Grooper designer will need to set up an Import Watcher service to create Batches from an import source.
  2. Use the "Tasks" tab to search for all ready tasks, completed tasks, working tasks, or tasks in an error state.
  3. View the processing statistics in the "Processing History".

Furthermore, you may note the "Batch Viewer" tab is missing in the Grooper Web Client. However, you can open a similar viewer by first pausing a Batch, then double clicking it to open it.


Tasks Page

The Tasks page is different from the Batches page in that it only presents users with Batches with Review steps currently ready for processing. Users can pick and choose which Batch they want to review, or they can set up a task filter and start processing all Batches it returns in order the Batch's age.

To get to the Tasks page, click the Tasks icon on the Grooper Web Client homepage.

  1. In the Navigation Links panel of the homepage, click the Tasks link.

This will bring up the Tasks interface. The first thing you'll see is a list of Batches with Review steps ready for processing.

  1. The list of Batches is always sorted by Age with the oldest Batch listed first the the newest created Batch listed last.
  2. You can select also the Filter icon, to filter out Batches by certain criteria.
  3. This will bring up a window to filter out your selection based on Queue, Process, Step or Batch (the Batch's name).
  4. Click the hamburger icon at the end of the property heading you want to filter by.
  5. Select the specific value you want to filter by.
    • For example, we could select a particular Batch Process which would give us a list of only Batches with that Batch Process
  6. Click "Save" to execute the filter or "Cancel" to cancel.

To start reviewing Batches, you have two options.

  1. You can select a single Batch from the list by double clicking it.
  2. You can click the "Play" icon to start reviewing all Batches in the list that match your filter.
    • Once one Review task is completed, the next Review task in the list for the next Batch will automatically open.
    • This is a handy way to start feeding yourself review work, without manually selecting each Batch every time you complete a Review task.

Just as we saw using the Batches page, this will bring up the Review activity module to perform one kind of review or another, be it classification review, data review, image processing review or another. For example, this is the exact same "Classification View" module for the exact same Batch we saw earlier. The document review is identical whether you open the Review step using the Batches page or the Tasks page. The only difference is how you get there.

The individual "Review Views" will be discussed in the #Review Views section of this article.

For now, we're going to simply exit the review module.

  1. To exit without saving your work, click the "Stop" button to return to the Batches page.
  2. You can also use the Context Toolbar to navigate to a different page.

FYI This interface and how you interact with it is similar to using the Grooper Attended Client thick client application. Prior to the Grooper Web Client, the Grooper Attended Client was used to filter production Batches with Review steps ready for processing and start processing them. Just as is the case with the Tasks Page, the Grooper Attended Client would feed end users review tasks, Batch after Batch until no work was left in the queue.

The biggest difference between the two interfaces is the task list in the Grooper Web Client. This is the same list you would access using the Open Task button in Grooper Attended Client.

Otherwise, the two interfaces are functionally same. All the same filtering capabilities you have in Grooper Attended Client are present in the Grooper Web Client.

Click here to return to the top

What is a Document?

Before continuing, lets take some time to cement some Grooper terminology we've been using as well as some of the icons you'll be seeing through the rest of this article.

As we've mentioned previously, a Batch is the fundamental collection of work in Grooper's document processing. It is essentially two things:

  1. A container of documents in various states of processing.
  2. A step by step list of instructions of what to do with those documents, or its Batch Process.

We often use the term "document" loosely. It can be an overly generic term for the stuff in the Batch that Grooper is doing stuff to. However, from Grooper's perspective a "document" is a very specific thing represented in a specific way in a Batch. So what is a document really?

Grooper has two objects to represent items in a Batch:

  • Batch Folders
  • Batch Pages

So, anything in a Batch is either a folder or a page.

A "document" is just a special kind of folder. In the most basic sense, a "document" is a folder with content. That content can be child Batch Pages or a digital file (like a PDF) attached to the folder.

This is Grooper's normal representation of a Batch as a hierarchy of Batch Folders and Batch Pages.


At the top is the Batch Root.

  • This is always represented by a folder icon and named after the Batch itself. The Batch Root is truly just a folder. Just like any other folder, it contains items. It's just a special folder in that its at the top of the folder hierarchy, containing all items below it.


Batch Folders will be represented by a folder icon.

  • So both "Folder (1)" and "Folder (2)" are Batch Folders.


Batch Pages are represented by thumbnails of the page's image.

There's a big difference between "Folder(1)" and "Folder (2)".

  • "Folder (1)" is a document (or a "document folder").
  • "Folder (2)" is not (It's just a simple folder).

Why? "Folder (1)" has content. It contains two Batch Pages, "Page 1" and "Page 2". We can expand the folder's contents using the arrow button to the left of the folder icon.

"Folder (2)" has no content, making it a regular old folder.

FYI You'll often hear Grooper users talk about a parent/child relationship. A parent/child relationship describes how items (called "objects" or "nodes") are related at different levels in a hierarchical structure, such as our Batches. In this case, the pages (which are at Level 2 of the Batch hierarchy) are children of the document folder "Folder (1)" (which is at Level 1 of the Batch hierarchy). "Folder (1)" is the parent of its child pages. Folder (1) is a child of the Batch itself (which is the root or Level 0 of the Batch hierarchy).

Simple enough, right?

Next, let's talk about classification. A classified document is a document folder who has been assigned a Document Type from a Content Model.

Grooper architects design Content Models to determine what makes one kind of document distinct from another and how to get information from them. These "different types of documents" are distinguished as Document Types created in the Content Model. By assigning a document folder a Document Type, Grooper then can use the logic defined in the Content Model to extract data from it.

Proper document classification is often critical to the process downstream. So, it's paramount to make sure Grooper assigned a document the right Document Type. One of the things you may be doing in Grooper is executing a classification review module to do just that.

However, be aware, once a document is classified, the items in your Batch are going to look a little different.

Here, "Folder (1)" has been classified. It's folder name has changed to "Federal W-4 (1)". Why? It was assigned a Document Type named "Federal W-4".

  • Notice the icon changed as well, from a folder icon to a document icon.


"Folder (2)" is still not a document, just a folder. It has no content.


"Folder (3)" is a document, just an unclassified one. It does have content, but no Document Type assigned to it.

  • Its name remains the generic "Folder" name, and its icon has not changed.


So, a document is a special kind of folder, and a classified document is a special kind of document.

  • Documents are folders with content.
  • Classified documents are documents that have been assigned a Document Type.

⚠ If you're importing files (such as PDFs or TIFF files), rather than hooking Grooper up to a scanner to bring in content, please pay attention to this next part.

The two main ways to get content into Grooper is by scanning pages directly into a Batch or importing files (such as PDF or TIF documents) from a file system.

If you are importing document files, Grooper will create a Batch Folder for every file imported, and attach that file to the folder. Things will look a little different than what we've described so far.

Here we have three Batch Folders created for three PDF files imported into a new Batch. Absolutely no processing steps have been executed for this Batch.

However, for each folder...

  1. You'll see the document icon instead of the folder icon for each item.
  2. The folders are named "Document (#)" instead of "Folder (#)".
  3. The file imported for each folder is attached to the folder and listed under its name.

Are these folders documents? Yes

  • While these folders do not have child content, like pages, they have attached content in the PDFs attached to each folder.

Are these documents classified? No

  • Despite sharing the same icon as a classified document, these documents are not classified.
  • They will not be classified until they are assigned a Document Type and their name changes from "Document (#)" to "Document Type Name (#)"

To sum up:

  • All documents are folders. Not all folders are documents.
  • Documents are folders with content.
    • Content can be child pages (or documents).
    • Content can be files attached to the folder.
  • Classified documents are documents who have been assigned a Document Type.

Review Views

In this section, we will demonstrate the various document review applications in Grooper and how to use them.

When you start processing Review steps in a Batch, you're going to see one or more different "Views" into the Batch. These Review Views present the Batch in different ways, best suited for the type of work you're doing. In these Views, you will verify Grooper's work during automated steps of a Batch Process and use the review modules to manually edit a document if Grooper made a mistake.

There are currently five Review Views available in Grooper:

Classification Viewer
  • You will use this to verify how Grooper classified a document during the Classify step. You may also use this view to verify how pages were separated into document folders during the Separate step.
Data Viewer
  • You will use this to verify how Grooper extracted data from a document during the Extract step.
Thumbnail Viewer
  • You will use this to review individual page images. Most commonly, this is used to verify how pages were processed by an IP Profile (for example, during the Image Processing step) or otherwise ensure the pages are ready for OCR during the Recognize step.
Folder Viewer
  • This is a fairly generic Batch viewer. This is most often added as a secondary Review View so that the user has an option to navigate to folders using the standard folder/page hierarchy view.
Scan Viewer
  • If you are using Grooper to scan paper documents with a scanner hooked up to your workstation, you will use the Scan Viewer to do so.

Document Viewer Tips

The Document Viewer is a common element among all Review Views. It will always occupy the right-most panel of the Review screen. It's how you, the user, can inspect a document or page selected in a Batch.

Before we get into each of the individual Review Views and how to use them, let's familiarize ourselves with the Document Viewer. This will include quality of life advice, such as how to zoom in and out of a page's image.

Zooming In and Out

By default, the image will be zoomed to a Width view. The image will fill the viewer based on the width of the document.

The zoom view is indicated by the Zoom icon at the top of the Document Viewer.


There are three ways to zoom in or out of a document's image.

  1. Click the Zoom icon at the top of the Document Viewer and select your preferred view.
  2. Hold the Ctrl key and use the mouse wheel to zoom in and out of the image more granularly.
  3. Use keyboard shortcuts to select a zoom view or zoom in or out.

1. The Zoom Icon

  1. Click the Zoom Icon.
  2. You will get a drop down list of a variety of different zoom modes to choose from:
    • Fit
    • Fit Height
    • Fit Width
    • Full (100%)

2. Mouse Wheel to Zoom

You can also use the mouse wheel to zoom in and out of the image.

  1. Be sure your cursor is hovered over the image.
    • If you don't, you'll end up controlling the zoom view for your entire browser window.
  2. Press and hold the Ctrl on your keyboard and either:
    • Scroll forward on the mouse wheel to zoom in.
    • Scroll backward on the mouse wheel to zoom out.
  3. The Zoom icon will change to a magnifying glass to indicate a custom Zoom is being used.

You can zoom in up to 300% of the image's size and zoom out up to 5% of its size.

3. Keyboard Shortcuts

Alternatively, you can use the following keyboard shortcuts to control the zoom view:

Zoom Setting Keystroke
Width W
Hieght H
Fit F
Full 1
Zoom In I
Zoom Out O

Resizing Panels

You may also resize the Document Viewer panel. This can be particularly helpful when using the Data Viewer to review extracted data.

For example, we can't see all the the extracted table data here. There are columns hidden out of view.

We can resize the Document Viewer panel to see more of the Review Viewer panel, using our mouse.

  1. Hover your cursor between the Review Viewer and the Document Viewer.
    • You will see the narrow gap between the two panels change to an orange color.
  2. Click and hold the left mouse button.

  2. Move your mouse right to narrow the Document Viewer or left to widen it.
  3. With the Document Viewer narrowed, we can see all of the columns of the extracted table data.

Rendition Views

The Rendition Views menu is found at the top right of the Document Viewer. Just click the icon to access the drop down menu. This allows users different views of the document or page's content. Depending on the circumstance, review users may find one Rendition View most helpful to complete their Review task.


The Rendition Views are as follows:

  1. Child Rendition
  2. Attachment Rendition
  3. Text Rendition

Attachment Rendition

If you ingested documents into a Batch by importing files (such as PDFs) from a file system, you will be able to access the Attachment Rendition. When files are imported into Grooper, a document folder is created for each file, and that file is attached to the folder.

  1. The attached file listed here is the original imported file attached to the document folder.
    • In this case it's a PDF file named "08.pdf".
  2. Selecting the Attachment Rendition will display this attached file.
  3. For multipage documents, you can use the page navigator to navigate between pages.

Child Rendition

The Child Rendition will display a document's content, as composed of its child objects. For example, if a folder has child pages, the document is the sum total of all its pages.

  1. Expanding out this folder shows it has two child page objects.
    • FYI: In this case, an activity called Split Pages was applied to the document folder. This created a page for each page in the attached PDF. The attached PDF was a two-page file. So, we ended up with two child pages in the folder.
  2. Selecting the Child Rendition will display the folder's content, as comprised of its child objects and their images.
    • In this case, a document formed from the two pages in the folder.
  3. If there are multiple child pages, you can use the page navigator to navigate between pages.

Text Rendition

The Text Rendition will display a document's OCR or extracted native text data.

  1. Selecting the Text Rendition will displays the document's Grooper generated full text data.
  2. Instead of an image, every line of text is displayed in the Document Viewer.

Click here to return to the top

Classification Viewer

The Classification Viewer allows Grooper users to review document classification. Grooper classifies documents using logic defined in a Grooper Content Model. Document Types are added to the Content Model to distinguish one type of document from another. Grooper is able to tell one Document Type from another by using trained examples of the documents, assigning rules for classification, or some combination of the two. Most typically, a document is assigned a Document Type during the Classify step of a Batch Process (although there are other ways depending on the Batch Process and how documents are ingested to a Batch).

Starting the Review Step

In the Classification Viewer you will visually verify the Document Type Grooper assigns is correct. You will either manually assign documents a Document Type if Grooper was unable to classify the document or change the document's Document Type if Grooper misclassified the document.

  1. We will select this Batch to review Grooper's document classification during the Classify step.
  2. As you can see, the step's name is "Classification Review"
  3. The steps activity type is "Review"
  4. And most importantly, its status is "Ready", indicating it's ready to be processed.
FYI A Grooper designer can name a Batch Process step whatever they want, but the activity type for review steps, regardless of the Review View, will always be Review.

Most often, the Grooper designer will name the review step after the kind of review that's being done or the Review View being used. However, be aware if the Grooper designer does not provide a custom name, the Review step will simply be named "Review".

When you open the Classification Viewer module, this is what you'll see. The Batch's documents are presented in the typical folder hierarchy viewer.

  1. Your job will be to select document folders and ensure the correct Document Type was assigned.
  2. Document Types are listed in the Document Types Viewer panel below the Batch Viewer panel.
    • In this example, we will be reviewing invoices. We've created a Document Type for each invoice's vendor.
  3. The document type will be listed in the folder's name.
    • For example, this document's name is "Nama (2)". It was assigned a Document Type named "Nama" (or the "Nama" Document Type).
  4. If a document was not classified, it will be flagged.
    • This is indicated by the red flag icon next to the folder.
    • Furthermore, the folder's name will remain the generic "Document".

Reviewing Document Classification

  1. To start reviewing, select a document folder.
  2. This will bring up the document in the Document Viewer panel.
  3. The document's classification results will be displayed in the Document Types Viewer.

  2. This document as assigned the "Fairdeal" Document Type.
  3. Why? Grooper determined it to be most similar to the "Fairdeal" Document Type based on the Content Model's classification logic.
    • In this case it scored an 87% similarity rating.
    • Put another way, Grooper is 87% confident this is a "Fairdeal" document.
  4. While there is some similarity to other Document Types, they are less than the "Fairdeal" Document Type's similarity.
    • Grooper will always assign the document the Document Type whose similarity is highest.

Grooper's calculation of these similarity scores are based on a variety of things, such as training algorithms and extraction rules. While Grooper tries to emulate what a human does when it looks at a document and makes a decision as to what it is, it's purely mathematical in nature. If the score is highest, its that Document Type from Grooper's perspective.

You, as a human being, are intuitive. You can make cognitive connections a computer simply can't. So, your job is to look at the document and make sure Grooper got it right.

Is this an invoice from Fairdeal Services?

  1. Yes. Grooper got it right. You can see the company's logo.
  2. You can see the invoices remittance address is addressed to Fairdeal Services.
  3. If you're familiar with invoices from this company, you will notice patterns in how the document is structured, how information is visually laid out on the page.
    • Whatever the use case is, you will use your knowledge of the document set to decide what the document is, and therefore what Document Type should be assigned, often within a split-second for each document.

Your job for the document is done. You've verified its Document Type is correct.

  2. You can move on to check the next document.
    • You may use your mouse and click on the next document.
    • You can also use the Up and Down arrow keys on your keyboard to navigate from one document to the next.

Correcting Document Classification

So what happens when things go wrong?

  1. Notice "Document (5)" has a flag next to it.
    • It has not been assigned a Document Type.
    • Also, the folder's name being "Document" is another indication it hasn't been classified.
  2. Why? It's not similar enough to any Document Type for Grooper to confidently classify the document.
    • FYI: By default, a document must score a similarity rating of 60% for a Document Type to be assigned. However, this can be adjusted. In your environment, your Grooper designer may have lowered that to allow a document to be classified below that threshold.
  3. This document should have been assigned the "Risiti" Document Type as it is an invoice from Risiti Construction.

So, we need to fix this and manually assign the Document Type. There are two ways to do this.

Option 1: Right Click and Assign Document Type

  1. Right click the document you want to classify.
  2. Select Assign Document Type.
    • Or, you can use a keyboard shortcut by selecting the document and pressing Ctrl + Shift + A on your keyboard.

This will bring up the "Assign Document Type" window.

  2. Press the hamburger button at the end of the Content Type property.
  3. Select the appropriate Document Type from the Content Model.
    • In our case, we've selected "Risiti"
  4. Click the EXECUTE button to assign the Document Type.

FYI You can also use the search box to search for a Document Type by name. Simply start typing in the search box

Upon applying your selection, the Document Type will be assigned to the document.

  2. The document's name has changed to "Risiti"
  3. The "Risiti" Document Type is now selected in the Document Types Viewer.

FYI You may have noticed the flag remains on the document after manually assigning it a Document Type.

Depending on how the Classification View is configured in Grooper Design Studio, you will either be allowed to complete your review with flagged documents or you will not be able to complete the task until all flags are resolved.

If you can't complete review until flags are resolved, you will need to remove the flag.

To remove a flag from the document:

  1. Right click the document.
  2. Select Clear Flag.
    • Or, you can use a keyboard shortcut by selecting the document and pressing Ctrl + Shift + L on your keyboard.

Option 2: Use the Document Types Panel

A quicker method of manually classifying a document may be to simply select the right Document Type from the Document Types Panel. We will use the next document in our Batch to illustrate this.

Another common problem is Grooper misclassifying a document.

  1. This document was classified as an "Ankara" Document Type.
  2. It should have been classified as a "Biha" Document Type, but its similarity score was too low.
    • "Ankara" scored an 89%. "Biha" scored an 87%. 89 is greater than 87. So, "Ankara" won out.

Rather than right clicking the document in the Batch and selecting a Document Type from a dropdown list, you can also simply double click the right Document Type in the Document Types Panel.

  1. Double click the Document Type in the Document Types Panel.
  2. The document will be assigned that Document Type.
    • So, our document changed from "Ankara" to "Biha".

Option 2.5: Better Utilizing the Document Types Panel

You should continue checking all document folders to ensure they've been classified correctly. We have one more problem in our Batch to resolve.

  1. Check out "Document (8)".
    • This document is flagged and unclassified. This should have been assigned the "Rechnung" Document Type, but it wasn't. It was not classified whatsoever.
  2. However, it scored a very high 92% similarity to the "Rechnung" Document Type, and it's also the most similar Document Type.
    • What gives? Why wasn't it classified?
  3. The problem is its not different enough from the next most similar Document Type.
    • "Rechnung" scored 92%. "Standard" scored 91%. That's only a 1% difference in their similarity.
    • In effect, this is "too close to call". Grooper has erred on the side of caution and not classified the document, leaving it up to the reviewer to determine which Document Type is correct.
FYI By default, Grooper requires at least a 2% difference in Document Type similarity. However, this minimum difference can be increased or decreased in Grooper Design Studio.

So, we need to manually classify the document. This gives us an opportunity to demo a handy keyboard shortcut.

  1. Select the document you want to classify and press the Tab key on your keyboard.
  2. This will move you to the Document Types Search Box. Start typing the Document Type you want to select.
  3. Once you've narrowed down which Document Type you're looking for, simply press enter to assign the document the selected Document Type.
    • FYI: You can also use the Up and Down arrow keys in the Document Types Viewer to select Document Types as well.

This is particularly useful if you have a large Content Model with dozens or hundreds of Document Types.

⚠

Bug Alert: At this time, hitting tab may not move the selection to the search box. We are currently working on a fix for this keyboard shortcut. In the meantime, just click on the search box to begin typing.

Completing the Review Step

  1. Once all documents have been reviewed, you're ready to complete the task.
  2. To do this, click the "Complete Task" icon in the Context Toolbar.

You will be presented with a Confirmation window to verify you're ready to complete the review task.

  2. Click the OK button to complete the task.

  2. This will complete the Review step in the Batch Process
  3. Grooper will continue processing the next steps in the Batch Process.

⚠ When in any Review View, you will have three buttons in the Context Toolbar.
  1. Complete Task
  2. Stop Task
  3. Delete Task


The Stop Task button will close the Review task. This will exit the Review View and return you to the previous page.

  • If you Stop Task, changes to the Batch ARE SAVED.
  • This means if you stop review work on a Batch, you (or another reviewer) can pick up where you left off.


    • The Delete Task button will delete the current task, typically meaning it will delete the current Batch you are reviewing.

  • There is no "undo delete" in Grooper. If you Delete Task, you will delete the Batch without going back.
  • DO NOT press the Delete Task button unless you are absolutely sure you want to delete the Batch forever.
    		•

    Completion Criteria

    The Classification Viewer may be configured so that certain criteria must be met in order to complete the review task. If so configured, either or both of the following conditions must be satisfied:

    • All document folders must be classified.
    • All flags on document folders must be removed.

    If this completion criteria has been enabled, and a Batch has documents that are flagged and/or unclassified, you the Classification Viewer will notify you in two ways:

    1. A yellow exclamation mark will appear next to the Classification Viewer' tab.
    2. The Complete Task button will be greyed out.
      • This button will be unclickable until the completion criteria is satisfied. This is Grooper's way of ensuring all documents have been reviewed before the task is completed.

    Click here to return to the top

    Shortcuts

    Shortcut Keystrokes Description
    Shared Folder and Page Commands
    Flag Item Ctrl + L Places a flag on the selected folder/page. Users may select pre-generated flag messages or enter their own custom message.
    Clear Flag Ctrl + Shift + L Removes a flag on the folder/page.
    Goto Flagged Ctrl + G Selects the next document/page in the Batch with a flag. If there are no subsequent documents with flags in the Batch, it will cycle back to the first document with a flag.
    Delete Del This will delete the selected folder/page. CAUTION!!! There is no "undo" in Grooper. If you delete an item, it will be gone forever.
    Rename F2 Renames the folder/page. Be aware, this does not classify a document folder. It only changes the folder's name.
    Cut Ctrl + X Cuts a selected folder/page in the Batch.
    Copy Ctrl + C Copies a selected folder/page in the Batch.
    Paste Ctrl + V Pastes a copied or cut folder/page to the selected folder location in the Batch.
    Move Down Ctrl + Down Moves the selected folder/page down in the Batch.
    Move Up Ctrl + Up Moves the selected folder/page in the Batch.
    Append to Previous Ctrl + P For folders, this appends any of a selected folder's children (pages or folders) to the folder before it. Effectively this will delete the selected folder and move any of its pages/folders to the bottom of the previous document/folder.

    For pages, this will move the selected pages to the bottom of the previous folder above.

    Prepend to Next Ctrl + Shift + P For folders, this prepends any of the selected folder's children (pages or folders) to the folder after it. Effectively this will delete the selected folder and move any of its pages/folders to the bottom of the next document/folder.

    For pages, this will move the selected pages to the bottom of the next folder below.

    Merge Selected Ctrl + M Merges selected folders/pages into a new document. This will create a folder, prompt you to assign it a Document Type, and move the selected folders/pages into the new folder.
    Folder Specific Commands
    Assign Document Type Ctrl + Shift + A Opens a window to select a Document Type for the selected document.
    Remove Level Ctrl + U Deletes the folder and moves any child objects (pages or folders) to the folder's level in the Batch. For example, if there was a document folder at Level 1 in the Batch with a single page in it (at Level 2). The folder would be deleted and the page would be moved to Level 1 in the Batch.
    Create Folder Ctrl + F Creates a new empty folder.
    Page Specific Commands
    Rotate Left Ctrl + Left Rotates the page 90 degrees to the left (counter-clockwise).
    Rotate Right Ctrl + Right Rotates the page 90 degrees to the right (clockwise).
    Split Folder Ctrl + S Splits a document into a new folder at the selected page. This applies specifically to document folders with multiple pages. Imagine you have a five page document folder at Level 1 in the Batch. You select page 3 and apply the "Split Folder" command. This will cut pages 3 to 5 from the document folder and place them into an unclassified folder at Level 1. You'll end up with two folders created out of the original (One containing pages 1 and 2. One containing pages 3 to 5) both at the same level in the Batch hierarchy (Level 1).

    Data Viewer

    The Data Viewer is used to review the data Grooper collects from each document during the Extract step of a Batch Process.

    The Extract activity applies the logic set up in a Content Model to find and return data from a document. This extraction logic is defined by configuring Data Models. Data Elements are added to the Data Model for each piece of information you want to collect.

    There are three types of Data Elements. Data can be collected as either Data Fields, Data Tables, or Data Sections (or "fields", "tables", and "sections" for short).

    • Fields are for the most basic kinds of information listed on a document.
      • This is what's called "single instance" data. Think a social security number on a W-2 form. There will be one single social security number filled in for the whole document. There is a single instance of this information (hence the term "single instance"), collected as a single value for the field.
    • Tables are necessary to collect information listed in a table formed by rows and columns on a document.
    • Sections can be tools to group data into a category, sub-divide a document into smaller units, or establish "multi-instance" sections (more on what this means later).

    As a reviewer, it's your job to check Grooper's results for each of these Data Elements after the Extract activity collects them. This is precisely what the Data Viewer is for. There's a lot of things that can go wrong in the wide world of document processing. Optical Character Recognition (OCR) can convert a document's image to digital text. However, it's not perfect. Rarely will your OCR results be 100% accurate. If the document's underlying text data is imperfect, your data extraction will probably not be accurate. There might be other problems with the extraction logic's ability to find and return data. This is especially the case for document sets with a lot of variety. If a document has a data structure that has not been properly modeled in the Data Model's design, there's a good chance Grooper will fail to return the data at all or only return partial data. Regardless why the error occurred, you, the reviewer, are the last line of defense to ensure accurate data is captured for each document.

    Starting the Review Step

    In the Data Viewer you will verify the data Grooper extracts from each document is correct. If what Grooper extracts does not match up with what's on the page, you will edit the result using text box editor.

    1. We will select this Batch to review Grooper's data extraction results obtained during the Extract step.
    2. As you can see, the step's name is "Data Review".
    3. The step's activity type is "Review".
    4. And, most importantly, its status is "Ready", indicating it's ready to be processed.
    FYI A Grooper designer can name a Batch Process step whatever they want, but the activity type for review steps, regardless of the Review View, will always be Review.

    Most often, the Grooper designer will name the review step after the kind of review that's being done or the Review View being used. However, be aware if the Grooper designer does not provide a custom name, the Review step will simply be named "Review".

    When you open the Data Viewer module, this is what you'll see. This is a different view into a Batch than we've seen so far. It's designed specifically to give us information about the data collected for each document.

    1. Instead of using a folder hierarchy, you can navigate through the documents in the Batch using the Folder Navigator at the top of the Review Panel.
      • There are eight document folders in this Batch. We are on the first document in the Batch. So we are at folder "1" of "8", indicated by "1 / 8".
      • You may use the single arrow buttons to go to the next or previous document.
      • You may use the double arrow buttons to go to the first and last document.
      • You can also type the number of the document you want to select in the number box.
    2. The document's classified Document Type and folder number is listed next.
      • Pro Tip: If you need to reclassify the document at this point, you can right click this heading and choose "Assign Document Type" to change its Document Type. Be aware changing a document's Document Type will clear its extracted data. However, you can also right click this heading and select "Extract" to re-run Grooper's data extraction.
    3. The document's extracted data occupies the rest of the Review Panel. The various fields, tables and sections established in the document's Data Model are listed here with their extraction results placed in editable text boxes.

    Basic Navigation

    When selecting a field, you should notice a few things:

    1. The field's name is bolded.
    2. The field's extracted value is to the right in a textbox.
      • The selected field's textbox is also outlined, whereas the rest are not.
    3. The extracted value is highlighted in green on the document.
    4. If any context was used to find the value (such as a label), it will be outlined in blue on the document.

    Your job as the reviewer is to look at the extracted value on the left and make sure it matches what's on the document on the right.

    • In this case it does, we can move on to the next field.

    Most review users will use the Tab key on their keyboard to quickly enter the next field.

    1. Press the Tab key to enter the next field or click on the field you want to review.
    2. If the value was extracted, Grooper will automatically go to that field's value on the document.
      • For multipage documents Grooper will take you to the page where the value is.
      • If you have the Document Viewer zoomed in, like we do here, it will center the value in the viewer.

    2. Keep pressing the Tab key or click on each field, reviewing all fields and table cells, until you've reached the last one on the document.
      • At this point, if you press the Tab key, Grooper will save any changes you've made to the document's data, and take you to the next document in the Batch.

    The next document has something wrong with its data. Grooper has several visual cues designed to indicate it's found data errors for the current document, or the Batch as a whole.

    1. The yellow exclamation mark indicates invalid documents. There is something wrong with the data for at least one document. This makes the document "invalid" and in need of validation. Your job is to review any data errors and manually correct them or verify they are correct.
      • This could mean a required value was not extracted, something requires manual validation, Grooper extracted data that did not match a field's expected type, or a custom validation event placed an error on a field.
    2. By default, Grooper will not allow you to complete the review task, until all data errors are resolved. The Complete Task button will remain grey and unclickable until all data errors have been resolved.
      • FYI: Depending on the Batch Process, you may need to complete the review task with errors present. In those cases, your Grooper designer will configure the Data Viewer so that you will be able to complete with errors unresolved.
    3. The yellow document icon in the Folder Navigator indicates there are "invalid documents". The number displays how many documents have data errors.
      • You can click this icon to navigate to the next invalid document.
    4. The red warning icon indicates there are data errors in the selected document's data. The number displays how many fields or table cells are in an error state.

    1. Click the warning icon to get more information about the errors present.
    2. This will toggle a list of every field or table cell with an error and their corresponding error message.
      • In this case it's telling us the "Invoice Total" field's "Value is required".
    3. Any field in table cell in an error state will be highlighted red.
    4. When you select that field, the error message will pop up next to it.


    Next, we will dig into some of these common data errors and how to resolve them by discussing how to review fields, tables and sections in the Grooper Web Client.


    Reviewing Data Fields

    We will start our journey into data review by looking at how to review fields. We will use the same set of invoice documents we reviewed for classification previously. It's fairly common in a review workflow to go from reviewing document classification to reviewing data extraction. First, you review Grooper's work to make sure the documents are classified correctly. Once Grooper knows what kind of document it's working with, it knows what data its looking for and how to find it. Now that Grooper has extracted the data, we can use the Data Viewer to verify it collected all the data required and collected it accurately.

    FYI It should be noted document Data Models have a high degree of configurability. Obviously, unless you're processing invoices, the specific data elements you will be reviewing in your environment will be different. You may have hundreds of data points to review on a single document. You may have just a few. That all depends on the business requirements for your document set and what your organization deems appropriate to extract from them.

    However, the basics remain the same across all use cases. Grooper will extract information from the document, populate that data into fields and tables, and you'll review the results based off what you a human can see on the document.

    Required Fields

    Commonly, an organization will deem certain data critical for document processing. Certain fields must therefore be extracted in order for the work to be considered complete. In Grooper, we satisfy this requirement by making a field "required". This will place the field (or table cell) in an error state if no value was extracted at all. In the Data Viewer, Grooper will alert you that the required value is missing, and will require you to manually enter it before review is completed.

    In the case of this document's Data Model, three fields are required:

    • Invoice Number
    • Invoice Date
    • Invoice Total
    1. We have navigated to the second document in the Batch.
    2. The "Invoice Number" and "Invoice Date" fields extracted just fine.
    3. The "Invoice Total" field did not. It is empty or "blank".
      • Grooper will highlight any required fields that are empty in red.
    4. When you enter that field's textbox, Grooper will pop up a message indicating the problem, "Value is required."

    1. All we need to do is enter the value for this invoice total as it appears on the document.
    2. Type the value into the field's textbox and press Enter or Tab to move to the next field.
    3. You will see the error warning disappear because the data error was resolved.
    4. Press the Save button to save any changes made to the document's data.

    Data Model Differences

    Before looking at more problems, please be aware Data Models can be (and often are) different for individual Document Types. For the most part, we're working with a "flat" Content Model. All the Document Types share the same Data Model, meaning we're looking for the same data elements for each one. However, in your environment, each Document Type may represent more diverse kinds of documents and require their own individual Data Models with their own specific fields and tables. Or, your Document Types may all share some data elements, but have some addition fields unique to the individual Document Type.

    This is the case with our "Envoy" Document Type. For the most part, the data we want to collect from this Document Type is the same as the rest. However, just for the "Envoy" Document Type we want to collect the purchase order number listed on the invoice. For whatever reason, we'll pretend have a business need for the PO number from this vendor, but none of the rest.

    1. The top half of the review screen is occupied by the "parent" Data Model's fields. These are the ones shared by all Document Types.
    2. Then, we have the additional "Envoy" Data Model's elements we can review as well.
      • In our case it's a single "PO Number" field in a section named "Additional Details"
    3. We review the field just like we would any other field, and continue to the next document.
    FYI Pro Tip!

    Most users find tabbing through fields with the Tab key is the easiest way to review a document's fields in the Data Viewer.

    If you are on the last field of a document (such as this one) and press the Tab key, it will save the document and take you to the next one in the Batch.

    Data Element Overrides (and Required Validation)

    Another way Data Models can differ from Document Type to Document Type is through "Data Element Overrides" (sometimes just called "overrides" for short). This allows Grooper designers to change how fields, tables and sections behave for a specific Document Type while still maintaining a parent Data Model shared by multiple Document Types.

    We're going to use another common review feature to demonstrate this. There may be some data that is not only required to be present, but extremely important Grooper extracted accurately. Your Grooper designer may designate this as a field that requires validation. So, even if it's accurately extracted, the field will stay in an error state until the user clears it.

    For the "Ankara" Document Type, we've decided the "Remit To Address" requires manual validation. We've set up an override so that just this Document Type requires validation for this field. For the rest of them, we'll just take what Grooper gives us.

    1. Fields requiring validation will always be in an error state until the field is reviewed.
      • Grooper will give you an error message saying "This field must be reviewed"
    2. In our case, Grooper did extract this address accurately. What's on the document is what's in the extracted field.

    So, how do we proceed? We have to get rid of the error or Grooper will consider this an "invalid" document.

    To clear the error, you must "confirm" the field is valid.

    1. Right click in the field's text box.
    2. Select Confirm
      • Or, you can use the keyboard shortcut F6

    2. This will confirm the value is correct, and the textbox's color will change to green.
    FYI You may have noticed there are still data errors present on this document. The total number of errors dropped from "4" to "3".

    The remaining three errors pertain to the extracted table data. We will circle back to these issues in the next section when we discuss reviewing table extraction in Grooper.

    Rubberband OCR

    Our next document has a lot of problems with it. Grooper was not able to extract much from this document at all.

    1. Even the "Invoice Number" field is inaccurate.
    2. Only part of the invoice number was successfully extracted.

    2. We'll just type in the right invoice number.

    However, typing can be time consuming. There is a handy feature that can save time called "Rubberband OCR". We're going to use Rubberband OCR to capture the next two fields.

    1. Right click the field's textbox you want to fill with Rubberband OCR.
    2. Select Rubberband OCR.

    2. Using your mouse, lasso the text on the document you want to capture.


    2. Grooper will fill in the field with whatever text data falls within the selected region on the document.
    ⚠ BE CAREFUL!!! Grooper grabs whatever text data falls within the zone as obtained by the Recognize activity. If text was obtained by OCR it may or may not be 100% accurate. Minor errors in OCR can cause characters to be misrecognized. Furthermore, the more degraded the document's image is, the more likely OCR results are to be inaccurate.

    Even when using Rubberband OCR, you should always visually verify the rubber banded text matches what's on the document.

    There's also a keyboard shortcut for Rubberband OCR.

    1. Enter the field's textbox you want to fill with Rubberband OCR.
    2. Press the F4 key.

    2. Using your mouse, lasso the text on the document you want to capture.


    2. Grooper fills in the field with the lasso'd text.
    FYI If you use Rubberband OCR in a field that already has text present, it will overwrite the text with the rubber banded text.

    "Valid" Doesn't Mean Accurate

    1. If we click the "Invalid Documents" icon (or press Ctrl + F8 on the keyboard) Grooper takes us back to this document.
      • As mentioned before, this document has errors pertaining to its extracted table data, which will discuss in the next section.

    This may lead you to believe the rest of the documents are fine, and there is nothing wrong with their data.

    However, "valid" does not mean always mean accurate.

    You should always take care when reviewing documents to touch every single one, verifying each field even if Grooper does not flag it as erroneous.

    "Valid" from Grooper's perspective means some very specific things. This includes:

    • A required field or table cell is missing.
    • A field set to require manual validation has not been validated.
    • A field's value does not match it's expected type (For example, Grooper expects a currency value to be extracted and a word is extracted instead. We'll see an example of this later on)
    • A custom validation expression throws an error (There will be an example of this in the #Advanced Techniques: Validation and Calculation Expressions section of this article)

    If Grooper fails to extract a non-required field or the underlying OCR data is inaccurate, you may still need to edit the results.

    1. For example, on the last document in our Batch, the "Remit To Address" was not collected.
    2. But, Grooper did not flag any data errors for this document. It is, technically, a valid document.
      • For the "Rechnung" Document Type the "Remit To Address" was not a required field and is not set up to require validation.
    3. That said, there is a remittance address on the document. So, we really should collect it.
      • It would be inaccurate to say there is no "Remit To Address" on this document.

    Reviewing Data Tables

    Tables differ from fields in that they are "multi-instance" data. You might have a single row for one table on one document. The next might have a hundred. There are multiple potential instances of a row, meaning there will be multiple instances of a column's value (one for each row).

    By and large, the review experience for tables is the same as for fields. Just instead of reviewing a single field's value, you're reviewing multiple cell values for all the columns and rows of the table. The same basic principles of validating Grooper's extraction against what's on the document apply.

    That said, there are a few things specific to table extraction you'll need to know.

    Visualization

    First, let's look at what Grooper is showing you when you enter a table cell for review.

    1. We have selected this cell to review. The "Price" column's value in the fourth row of this invoice's line items table.
    2. The blue highlighted region is showing you where Grooper determined the selected cell's column to be.
    3. The yellow highlighted region is showing you where Grooper determined the selected cell's row to be.
    4. The green highlighted region is where Grooper determined the selected cell's value to be.

    Removing Table Rows

    Extracting tabular data from documents can be difficult because they can be so variable. Your Grooper designer will need to model the table's structure in one way or another for a variety of different documents. A table on one document might look totally different from another, and more often than not, there's no telling how many rows there will be.

    One thing that can happen during extraction is either not enough rows are captured, or too many rows are captured. You will need to add or remove rows accordingly.

    1. In the case of this document, we've collected three additional rows in error. These are not part of the document's table and need to be removed.
    2. Furthermore, three cells in these erroneous rows are in an error state. We have to resolve these errors before we can complete our review.
      • This is because our "Price" column is set up to expect decimal values (specifically currency values). However, we've captured letters instead of numbers. This is one of the ways Grooper will let you know it extracted something it didn't expect (if its able).

    We just need to get rid of these rows entirely.

    1. To remove a row, first right click the dash mark at the beginning of the row.

    2. Select Delete Row.
      • Note you can also insert rows, duplicate a selected row, and move rows up or down.

    2. Continue removing rows until all extra error rows are gone.
      • FYI: You may also use the keyboard shortcut

    Adding Table Rows

    What do you do when Grooper fails to extract a table at all? You need to add some rows and fill in the cells manually.

    1. This invoice definitely has a line items table, detailing goods and services purchased.
    2. However, for whatever reason, Grooper was unable to extract any data for any row.
      • We don't even have a blank row to start filling. So, we need to add one.

    1. To add a row, right click any of the table's column headers.
    2. Then, select Append Row.

    2. This will add a single, blank row to the table.
    3. You may continue to use the Append Row button to add any more necessary rows, right click the dash mark next to the row and press the Insert Row button, or use the keyboard shortcut Ctrl + Shift + Ins to insert additional rows.

    2. Fill in the cell values and you're good to go!
    3. Don't forget to save your changes.

    Again, "Valid" Does Not Mean Accurate

    Attention to detail is paramount when reviewing a document's data, regardless whether it's field-level data or tabular data. It may even be more paramount when reviewing table data, as tables can be densely packed with information that is easily glossed over if you're not paying close attention.

    1. At this point there are no more invalid documents in the Batch. The Invalid Documents icon is gone.

    So we're good to complete our review, right? Wrong!

    Always remember, a "valid" document does not necessarily mean its data is 100% accurate.

    There's actually an error in the extracted table data due to an OCR error.

    1. This value should be $7.09 not $9.00
    2. Due to an OCR error, the extracted text was "09" (which was then formatted as $9.00 when formatted to a currency decimal value).
      • Obviously, this isn't right and needs to be fixed. This is a good example of how OCR errors can cause inaccurate data to be collected. Without eagle eyed reviewers like you, this is the type of thing that can be missed corrupting your data set downstream.

    1. With this cell edited, we have completed our review for this Batch.
    2. We have saved our changes to the current document.
    3. We can now complete our review task by pressing the Complete Task button.

    So far, we have discussed how to review two of the three Data Elements Grooper uses to collect data, Data Fields to collect field level data and Data Tables to collect table data.

    In the next section, we're going to discuss the remaining Data Element, the Data Section and how it can be used to subdivide a document into smaller sections. There are at least a couple things you'll need to be aware of if your document set's Data Models utilize sections to capture document data.


    Reviewing Data Sections

    Data Sections (or just "sections" for short) are a method of subdividing extraction in one way or another. They are used in three main ways in Grooper:

    1. As an organizational tool.
      • Fields and tables can be placed in a section to group them logically as similar kinds of data.
    2. To restrict extraction to a smaller subsection of a document.
      • Imagine you have a four-hundred page document. What you want to extract is on one paragraph on one page somewhere in the document. If you can first narrow down extraction to just that paragraph, it often makes extraction much simpler and ultimately more accurate.
    3. To extract repeating data fields from multiple repeating sections.
      • This pertains to "multi-instance" sections. Think about a form you've had to fill in where two or more parties have to enter in the same information, maybe their name, address, date of birth, and other information. You will know the fields for each individual, but need a method to capture the fields for multiple individuals (in other words, multiple instances of that data). In these cases, a "multi-instance" section will allow you to do this.

    As a reviewer, you may see any combination of sections used in these ways.

    1. Sections for Organization

    Spoiler alert! You've already seen this.

    1. When reviewing our invoice documents, you may have noticed this title "Header Details".
      • This is actually a name of section in our document's Data Model!
    2. The field's seen here are the section's "child" data elements.
      • Essentially, the "Header Details" section is just a container to put data fields in with a label to help identify them.


    Placing these fields in this kind of a section has no impact on extraction. This was purely used as a container to place data elements of a similar kind. In our case, data listed in the header of an invoice (as opposed to the line item information found in a line item table).

    FYI If you want to get real technical, a field in a section's full name is SectionName.FieldName

    You may see this reflected in the reviewer when evaluating data errors.

    See here the "Invoice Total" field in error is listed as "Header Details / Invoice Total".

    2. Sections for Subdividing a Document (Single Instance)

    In this document, we have a few sections performing different functions. The first function is to organize the information extracted in such a way that it is easy to read.

    1. This phone number is not inside of a section (there is a reason for this that we will discuss in a bit).
    2. Here we have a section for personal info.
    3. This section is further devided with a section for Address Info.
    4. Further down we have Dependents and Previous Employment.

    These sections also have an important function in extraction. On this document we have two phone numbers that need to be extracted: 555-555-5555 and 555-555-0124. Notice that they both have the same label of "Phone #" on the document. When Grooper runs extraction, by default it returns the first item that it finds on the page from top to bottom. Sections can help us extract both the number for Document Processors, Inc and the number for Johnny B Grooper.

    1. You can see that we are extracting the Document Processors, Inc phone number.
      • Here the extractor is just returning the first result on the page.

    2. The phone number for Johnny B Grooper is being extracted within a section.
    3. If you look closely, there is a pink box around a portion of the document. For the "Personal Info" section, Grooper is only looking inside this pink box for anything extracted within the section.
      • Since the Document Processors, Inc phone number is outside of the selected section, Grooper will skip over that and only consider what is inside the pink box. So, we are able to get Johnny B Grooper's phone number with no problems.

    3. Sections for Subdividing a Document (Multi-Instance)

    Sometimes, documents have repeating sections of fields listed in a single document where the same set of information is collected repeatedly throughout the document. In these cases, a field by itself isn't going to cut it. A field is designed to collect one piece of information. If that field is repeated multiple times, you're missing out on all the times it's listed.

    Sections can be configured as "single instance" sections or "multi instance" sections. So far, all we've seen are single instance sections. Multi-instance sections allow you to subdivide a document into multiple sections of similar data.

    A "multi-instance" section will allow us to divide up the document into multiple sections. Then we can extract (and review) the fields for each section.

    1. If you look on our document here, we have three sets of information for the "Previous Employment" area of the document. The "Employer", "Dates of Employment", and "Job Title" fields are repeated three times with different values for each set.
    2. We have created a multi-instance section here at the bottom.
    3. If you click the arrows next to "Previous Employment" you'll scroll through each set of data.


    Deleting Section Records

    Grooper finds these section instances in a variety of ways. Just like Grooper can produce false positive field results, it can produce false positive section instances as well. Part of your job in reviewing section data will be determining whether or not a section record is valid at all and should just be deleted.

    1. This document produced three total section instances/records.
      • We have navigated to the third section record.
    2. Only two of these sections are valid. The third just has nonsense entered into the fields.
    3. We can tell this record is invalid because the information in the fields does not match what is expected by the field label.

    So, we've got some deleting to do. This data is "junk". We want to get rid of it before we export these documents and their data to some back-end system, like a database or content management system.

    1. To delete a section record, right click inside of the section you want to delete.
    2. Click Delete Record.
      • Alternatively, enter any field in the section and use the Ctrl + Del shortcut.

    2. when the "Delete Record" window pops up, click "APPLY".

    2. The section (and all it's fields) have been deleted. You can see our section record count has dropped. Now we have two total records instead of three.

    Adding Section Records

    And now for the flip side of the coin, adding section records. Just like Grooper can fail to extract a field for whatever reason, so can it fail to extract a section instance. In these cases, you're going to need to add a section record.

    1. For this document, Grooper found two section records. However, three section records are filled in on the document.
    2. Grooper missed this one because someone marked up the document with a marker.
      • Grooper often uses key words and labels as anchors as to where a section starts and stops. If OCR can't read the text data, neither can Grooper.

    There are two ways to add a section record.

    • You can append a record to the end of the section records.
    • You can insert a record before a selected section record.
    1. To append a section record, right click the section's header title.
    2. Click Append Record.

    1. To insert a section record, right click any of the section's fields.
    2. Click Insert Record.
      • Or, enter a field's textbox and use press Ctrl + Shift + Ins on your keyboard.

    If you happen to insert the section record in the wrong location/order, you can move a section record by using the Move Previous and Move Next commands.

    1. In our case, the missing section should be the third section but we inserted it rather than appending it. So, it's now section "2 / 3".
    2. We can right click any field in the section to move this record to the right spot.
    3. We will select Move Next to move this record to position "3 / 3".
      • Or we can use the keyboard shortcut Ctrl + Shift + Right.

    Last but not least, any added section's field or table values will need to be entered manually to complete its review.

    Completing the Review and Completion Criteria

    Once you are finished reviewing the documents extracted data you will complete your review by pressing the Complete Task button.

    ⚠ When in any Review View, you will have three buttons in the Context Toolbar.
    1. Complete Task
    2. Stop Task
    3. Delete Task


    The Stop Task button will close the Review task. This will exit the Review View and return you to the previous page.

  • If you Stop Task, changes to the Batch ARE SAVED (as long as you pressed the "Save" button or otherwise saved changes to a document's data in the Data Viewer by tabbing out of the final field on one document to the first in the next).
  • This means if you stop review work on a Batch, you (or another reviewer) can pick up where you left off.


    • The Delete Task button will delete the current task, typically meaning it will delete the current Batch you are reviewing.

  • There is no "undo delete" in Grooper. If you Delete Task, you will delete the Batch without going back.
  • DO NOT press the Delete Task button unless you are ABSOLUTELY SURE you want to delete the Batch forever.
    		•

    Completion Criteria

    The Data Viewer, just like anything else in Grooper, has some degree of configurability. This gives Grooper designers the flexibility to architect solutions to best fit the needs of your business case. Be aware your Grooper designer may configure Data Viewer in one of two major ways:

    • Either you will be able to complete your review with invalid documents...
    • Or, you will not.

    You should be aware this is set up at the Batch Process level. You should not assume, just because you can complete a review task, the Batch has no errors that need to be manually validated. Some Batch Processes may be configured to allow you to complete the Review step with unresolved data errors. Others may not.

    This Batch's Batch Process was configured so that Review could not be completed with invalid documents' data errors unresolved.

    1. There are invalid documents present.
    2. The Complete Task button will remain greyed out and unclickable until the errors are validated.

    This Batch's Batch Process was configured so that Review can be completed with invalid documents' data errors unresolved.

    1. There is a data error present on this document.
    2. Despite the presence of invalid documents, the Complete Task button is lit up and can be pressed.


    Be aware, there are reasons you may want to (or more to the point need to) complete Review with data errors present. In that case, your Grooper designer may allow you to do so. However, do not assume just because you can complete the Review task, there isn't anything left to review.

    Please fully review your documents data as dictated by your organization before completing your review.

    Click here to return to the top

    Shortcuts

    Shortcut Keystrokes Description
    Navigation Shortcuts
    Go to Next Field Tab Navigate from current field/table cell to the next one. If you are on a document's last field/table cell, this will save any changes and navigate to the next document in the Batch.
    Go to Previous Field Shift + Tab Navigate from current field/table cell to the previous one.
    Go to Next Error Ctrl + N CURRENTLY INOPERABLE IN BETA. Navigate to next field/table cell with a validation error.
    Go to Previous Error Ctrl + P CURRENTLY INOPERABLE IN BETA. Navigate to previous field with a validation error.
    Go to Next Invalid Document Ctrl + F8 Navigate to the next invalid document. This will cycle back to the first invalid document in the Batch if no subsequent documents are invalid. This will do nothing if there are no invalid documents left in the Batch.
    General Validation/Review Shortcuts
    Confirm F6 Confirm the current field/table cell is valid.
    Rubberband OCR F4 Allows the user to populate a field/table cells value by selecting a portion of text on the document with the mouse.
    Rubberband Zone F3 Allows the user to place a geometric zone for a field/table cell by selecting a region on the document with the mouse. This is often used for document redaction or to highlight something on the document for a secondary review.
    Table Shortcuts
    Append Row Ctrl + Ins Adds a row to the end of the table.
    Insert Row Ctrl + Shift + Ins Inserts a row above the currently selected row.
    Delete Row Ctrl + Del Deletes the currently selected row.
    Duplicate Row Ctrl + D Duplicates the currently selected row.
    Move Row Up Ctrl + Shift + Up Moves the currently selected row up one row (i.e. If you moved the third row up, it would make the third row the second row)
    Move Row Down Ctrl + Shift + Down Moves the currently selected row down one row (i.e. If you moved the third row down, it would make the third row the fourth row)
    Section Shortcuts
    Append Record Ctrl + Ins Adds a section record to the end of a multi-instance section
    Insert Record Ctrl + Shift + Ins Inserts a section record before the currently selected section record.
    Delete Record Ctrl + Del Deletes the currently selected section record.

    Advanced Techniques: Validation and Calculation Expressions

    ‼

    COMING SOON!! This portion of the article is under construction.

    Advanced Techniques: Database Lookups

    ‼

    COMING SOON!! This portion of the article is under construction.

    Advanced Techniques: Rubberband Zone

    ‼

    COMING SOON!! This portion of the article is under construction.

    Thumbnail Viewer

    The Thumbnail Viewer allows a review interface for individual pages in Grooper. Most commonly, this is used to review the results of an IP Profile applied to scanned pages (for example, during an Image Processing step in a Batch Process).

    Commonly, users will review whether pages were not appropriately oriented or de-skewed by an IP Profile, flag pages that need to be re-scanned for one reason or another, or delete superfluous pages, like blank pages.

    Starting the Review Step

    In the Thumbnail Viewer you will review individual pages. This Review View is generally used early on in a Batch Process to ensure documents "look good", in one way or another, and are ready to be processed by OCR.

    1. We will select this Batch to review the page-by-page results of the Image Processing step.
    2. As you can see, the step's name is "Image Review".
      • This is common to name the type of review work your doing in a Thumbnail Viewer. You are reviewing how each page's image looks, visually.
    3. The step's activity type is "Review".
    4. And, most importantly, its status is "Ready", indicating it's ready to be processed.
    FYI A Grooper designer can name a Batch Process step whatever they want, but the activity type for review steps, regardless of the Review View, will always be Review.

    Most often, the Grooper designer will name the review step after the kind of review that's being done or the Review View being used. However, be aware if the Grooper designer does not provide a custom name, the Review step will simply be named "Review".

    When you open a Thumbnail Viewer, this is what you will see. Keep in mind, this review interface is designed for reviewing pages. The Batch's folder structure is hidden. The user is simply presented a list of pages in the Batch

    1. On the left side of the screen, you will select a page's thumbnail.
    2. You will use the Document Viewer on the right, to review the selected page's image.
    3. Page numbers are listed below the page's thumbnail.


    For the purposes of this exercise, not much was done to these page's during the Image Processing step. This will give use several issues to address using the Thumbnail Viewer.

    Sorting Pages

    The Thumbnail Viewer allows you to sort pages by a variety of different image qualities, such as the page's overall intensity or dimensions. This can be a useful tool to aid the reviewer in a variety of different ways.

    You can sort images by the following qualities:

    • Physical order (the numerical order of the pages in the Batch from first to last or last to first)
    • Intensity (the overall "brightness" of the image)
    • Width
    • Height
    • Aspect Ratio
    • Flag status (whether or not a Grooper flag is present on the page)


    1. To sort pages, Click on the sorting icon in the top righthand corner of the Thumbnail Selector.
    2. Select the icon that corresponds with the sort option you prefer (if you hover over each icon you will get a tooltip that lets you know which icon is which sorting category).
      • For example, we'll choose Sort By Intensity.

    Intensity measures the "brightness" of an image. From a practical standpoint, text is represented by black (or darker) pixels contrasted against the document's background which are white (or brighter) pixels. Sorting by intensity is a quick way to group all blank (or mostly blank) pages together, as they will be the most bright (since they have more white pixels). This gives the reviewer a quick way to delete blank pages, if necessary.

    After sorting by intensity, the images are laid out from "least intense" to "most intense".

    1. The "least intense" images are grouped at the top.
      • In other words, the pages with the most text or other printed material.
    2. The further down in sequence you go, the more intense the images get, with the "most intense" images at the bottom.
      • In other words, all the blank pages.

    You can also flip the sort order of the documents. For example, if we choose to Sort By Descending, the most intense (lighter) images will be the first in the Batch and the least intense (darker) images will be last.

    1. Click the three line icon in the top right of the Thumbnail Selector.
      • This will change between "sort by ascending" or "sort by descending".
    2. The brightest page is listed first.
    3. The darkest page is last.

    FYI Please note this does not change the pages' order in terms of their sequence in the Batch. It just changes the viewing order in which pages are presented.

    Even though "Page 1" appears to be towards the end of the Batch, it's actual position is still the first page in the Batch.

    If you need to go back to the original view of the Batch, with pages listed in sequential order, all you need to do is sort by "Physical Order".

    1. Click the sorting icon in the rop right of the Thumbnail Selector panel.
    2. click the Physical Order icon.

    Now the pages are listed in sequence as they are ordered in the Batch itself (You might have to change the sort order from "Descending" to "Ascending" again.


    Next, we're going to look at some common problems you may encounter and how to resolve them within the Thumbnail Viewer.

    Resolving Common Problems

    The Thumbnail Viewer allows users to delete superfluous (or "junk") pages, make adjustments to individual pages, or flag them for one reason or another. We will discuss the following common Thumbnail Viewer uses:

    • Deleting "junk" pages
    • Manually rotating pages
    • Flagging pages
    • Applying an IP Profile

    Deleting "Junk" Pages

    Blank pages are junk pages. They contain no meaningful information. More likely than not, you'll just want to get rid of them. This is where sorting by intensity can come in rather handy.

    FYI Grooper does have ways to automate blank page deletion. However, there my be reasons why Grooper fails to delete blank pages. For example, if the page is "nearly" blank. Furthermore, the following advice will apply to deleting what would otherwise be considered "junk", pages you don't want as a part of your final document set for one reason or another.

    In this case, we've already sorted by intensity. So, the brightest pages are first in the list.

    1. To delete a single page, select and right-click its thumbnail.
    2. Then select, Delete.
      • Alternatively, just select the thumbnail and press the Delete key on your keyboard.

    2. When the "Delete" window pops up, click "EXECUTE".
    ⚠ There is no "undo delete" in Grooper!

    Please be absolutely sure you're ready to delete a page before deleting it.

    1. You may also multi-select pages by holding down the Shift key to select all pages' thumbnails in sequence.
      • Or using the Ctrl and clicking multiple thumbnails.
    2. All selected pages will be deleted.

    You may also be asked to evaluate the content of a page to determine whether or not it is "junk".

    1. Someone drew a smiley face on this document.
    2. But, this page has meaningful data. We wouldn't want to delete it just because someone doodled on it.

    1. This page, on the other hand, is a "bleed through image".
      • It's a scan of the back page where the marker bled through.
    2. We would want to go ahead and delete it. It carries no meaningful information.

    1. "Page 6" has a hand-written note on it.
    2. It looks like someone wrote this note on the back page of a document for a reason. This could be meaningful information. Most likely, this information should be kept intact. I would say this page should not be deleted.
      • This is where you the human comes into play in human review. A computer can't make these kinds of decisions. It's up to you to determine what counts as good information that should be kept and a junk page that should be trashed (likely according to some guidelines your organization sets out for you).

    Manually Rotating Pages

    Another common problem you may encounter are pages that are oriented incorrectly. They may be scanned in upside down, or oriented as a landscape image when they should be a portrait image. Grooper does have methods to detect a page's proper orientation, but sometimes that fails. For example, some documents have horizontal text along both the horizontal and vertical axis of a page. Which is the "proper" orientation in these case? You, as a reviewer, may need to make that decision.

    In cases like this, you'll need to manually rotate a page to the correct orientation.

    1. We have some pages that are in a landscape orientation but should have a portrait orientation.
      • We're going to sort by aspect ratio to group all these landscape pages together.
    2. Click the sorting icon in the top right of the Thumbnail Selection panel.
    3. Select the "Aspect Ratio" icon.

    2. If needed, click the sort order icon to change whether the landscape pages show at the top or bottom of the batch.

    2. To rotate a page, right-click its thumbnail.
    3. In our case, we want to rotate it left (or counter-clockwise) to correct the page's orientation (so, the top of the page is at the top and the bottom is at the bottom).
      • Note there are also keyboard shortcuts as well, Ctrl + Left to rotate left and Ctrl + Right to rotate right.

    2. This will rotate the image 90 degrees to the left (or counter-clockwise), fixing the problem.

    As with deleting pages, you can select multiple pages by holding down the Shift key to select all pages' thumbnails in sequence (or holding ctrl and selecting individual pages), and apply the same rotation as well.

    1. Here, instead of just one page, we've selected all the landscape oriented pages.
    2. Right-click any of the selected pages' thumbnail.
    3. Select Rotate Left to rotate them all 90 degrees to the left/counter-clockwise.

    Flagging Pages

    You may need to flag a page for one reason or another. Likely, this will have to do with something specific to your business case. You may need to flag an image that needs elevated review, for example.

    Commonly, you may need to flag a page that needs to be re-scanned. A Batch Process can be designed in such a way that that pages without flags would go on to the next logical step (probably Recognize to OCR the pages). However, any page with a flag would be submitted to a secondary Scan step so any issues scanning the page can be resolved by replacing the image with a new scan.

    1. For example, this page was scanned with a folded corner. So, there's a large swath of white pixels covering the document's text.
      • This page should be rescanned with the corner unfolded, so we get the full page's image.
    2. To flag a page, right-click its thumbnail.
    3. Select Flag Item.
      • Alternatively, select the thumbnail and press Ctrl + L on your keyboard.

    2. This will bring up the Flag Item window.
    3. You may manually type in any text for the flag in the Flag Reason textbox.
    4. Or, if your Grooper designer has pre-populated a list of flags, you can access them using the hamburger icon.
    5. This will give you a dropdown menu of any flags that have been added.
    6. Press the Execute button when finished to flag the page.

    2. After the page is flagged, you'll see a red flag icon on the upper-corner of the thumbnail.

    Applying an IP Profile

    Your Grooper designer may ask you to apply what's called an IP Profile to pages in a Thumbnail Viewer. An IP Profile is a collection of image processing commands that are applied to clean up a page's image to prep it for further processing.

    For example, your images may have already been processed by an IP Profile by this point. However, this kind of IP Profile should be designed to account for most problems for most images. It should be fairly generic, generally cleaning up the pages in the Batch. You might have outlier pages that need more fine tuned processing. In that case, you may need to identify those images, and manually apply a more specific profile.

    If this is the case, you'll apply the IP Profile by right-clicking a page's thumbnail and selecting the one you want.

    1. Right-click the page's thumbnail you want to process.
    2. Select Image
    3. Select Apply Image Cleanup

    2. This will bring up the Apply Image Cleanup window.
    3. Press the hamburger icon at the end of the IP Profile property to select an IP Profile.
      • This will bring up a dropdown list containing any IP Profiles your Grooper designer has added for you to select.
      • We have three in this list "Aggressive Deskew", "Big Blob Removal", and "Crop and Deskew".
        • The "Aggressive Deskew" profile is designed to handle images that weren't deskewed by the previous Image Processing step in our Batch Process
        • The "Big Blob Removal" profile is designed to remove large "blobs" on the page, large marks on the page like our smiley face.
        • The "Crop and Deskew" profile is similar to the "Aggressive Deskew" profile, but it also allows cropping out unnecessary borders.
    4. Choose the IP Profile you want to apply.
      • We've selected "Big Blob Removal" to remove the smiley face doodle.
    5. Select Execute to apply the selected profile.
    ⚠ Please note image processing takes time.

    The more an IP Profile mutates an image, the longer it's going to take. If you execute the profile and it doesn't immediately look like anything happened, the image may still be processing behind the scenes and/or the server hasn't handed the processed image to you yet.

    You may need to wait several seconds to see the changes reflected in the Document Viewer.

    2. After the IP Profile is applied, you'll see any changes made to the image in the Document Viewer.
      • Poof! No more smiley face.
      • Please note applying an IP Profile takes time to process. It may take a few seconds before the changes are reflected in the Document Viewer.

    Undoing Changes

    If you make changes to an image and decide you need to undo those changes, you can undo them with the Undo Image Cleanup command.

    You can undo the following:

    • Rotation applied by the Rotate Left and Rotate Right commands.
    • Changes made by an IP Profile using the Apply Image Cleanup command.
    • Pixel inversion applied by the Image > Invert command.

    You cannot undo the following:

    • Page deletion.
    1. To undo changes, right-click the page's thumbnail in the Batch.
    2. Select Image.
    3. Select Undo Image Cleanup.
      • You may also select a thumbnail and use the Ctrl + Z keyboard shortcut.

    2. In the case of this image, the changes made by the Apply Image Cleanup command are reversed.
      • Hello again smiley face!

    Confirming Pages - Completion Criteria

    Depending on how the Thumbnail Viewer is set up, one (or both) of the following completion criteria may be necessary to complete your review:

    • You may be required to "confirm" all pages.
      • This is a way of ensuring a reviewer has visually inspected each page in the Batch.
    • You may be required to remove any flags on a page.
      • This is a way of ensuring any flag thrown by a previous step is inspected and the page is reviewed.

    This Batch's process was configured so that the Review step could not be completed until all pages are confirmed.

    1. The yellow exclamation mark tells us there are pages that have not been confirmed.
    2. The Complete Task button will remain greyed out and unclickable until all pages are confirmed.

    To confirm a page:

    1. Select the page's thumbnail.
    2. Ensure there is nothing wrong with the page in the Document Viewer (according to the business needs of your organization).
    3. Right-click the thumbnail.
    4. Click Confirm Page.
      • Or, simply select the thumbnail and press the Enter key.

    2. Upon confirming the page, Grooper will place a green checkmark in the center of the thumbnail.
    3. It will automatically navigate you to the next page.

    2. Once all pages are confirmed, Grooper will allow you to complete your review.
    3. The Complete Task button is now lit up and clickable.

    FYI Please note if you confirm a page, you will also remove its flag.

    If you the user are flagging pages in a Thumbnail Viewer as part of your organization's internal processes, you will not be required to confirm pages (or remove flags) in order to complete your review.

    Click here to return to the top

    Shortcuts

    Shortcut Keystrokes Description
    General Commands
    Confirm Page Enter Confirms the page is reviewed
    Flag Item Ctrl + L Flags the page. The user will enter their own flag message or select one from a pre-generated list.
    Go To Next Unconfirmed Ctrl + Enter Navigates to the next page that has not been confirmed yet.
    Delete Del Deletes the page from the Batch.
    Image Modification Commands
    Rotate Left Ctrl + Left Rotates the page 90 degrees to the left (counter-clockwise)
    Rotate Right Ctrl + Right Rotates the page 90 degrees to the right (counter-clockwise)
    Apply Image Cleanup Alt + I Allows the user to select an IP Profile to modify the selected page's image.
    Invert Ctrl + Shift + V Inverts the image's color scheme. This can be useful when viewing scans of film negatives (such as microfilm or microfiche).
    Undo Image Cleanup Ctrl + Z Undoes any image modifications. Replaces the page's image with its stored undo image.
    Sorting Commands
    Sort By Aspect Ratio Ctrl + Shift + A Sorts the pages by its picture aspect ratio. This would group all landscape oriented images together and all portrait oriented images together.
    Sort By Batch Order Ctrl + Shift + B Sorts the pages by their order in the Batch.
    Sort By Flag Status Ctrl + Shift + F Sorts the pages based on if they have a flag thrown on the page, or not. This will group all pages with flags together. If no pages are flagged, this will do nothing.
    Sort By Height Ctrl + Shift + H Sorts the pages based on their physical height, in inches.
    Sort By Width Ctrl + Shift + W Sorts the pages based on their physical width, in inches.
    Sort By Intensity Ctrl + Shift + I Sorts the pages based on their intensity (or brightness). Pages will be ordered from brightest to darkest. Or put differently, least amount of text (and other marks) to most amount of text (and other marks).
    Display Format Commands
    Display as Binary Ctrl + Shift + B Changes the page's image format to black and white.
    Display as Grayscale Ctrl + Shift + G Changes the page's image format to grayscale.
    Display as Grayscale Ctrl + Shift + G Changes the page's image format to color.
    Reset Ctrl + Shift + R Clears all ScanOnce settings on the page.

    Folder Viewer

    The Folder Viewer gives reviewers a fairly generic view into a Batch's folder structure. Most often, this is added as a secondary Review View to allow users a different view into a Batch.

    For example, a Data Viewer is well suited to view and edit documents' extracted data. However, you don't have the same view into a Batch's folder and page structure. This can make it difficult to navigate if you need something more like a hierarchical folder view.

    This Review step has two Review Views, a Data Viewer to review and edit the document's data, as well as a Folder Viewer.

    1. Currently, the Data Viewer is selected.
    2. Select the Folder Viewer tab to navigate to the Folder Viewer.

    The Folder Viewer's interface is fairly basic.

    1. You can navigate the Batch's folder structure using the review panel on the left.
      • This is identical to the way you navigate through folders and pages in the Classification Viewer. The only difference between these two viewers is you don't have access to the classification statistics at the bottom of the review panel in the Folder Viewer.
    2. The folder or page you have selected is displayed in the Document Viewer panel on the right.

    Using Review Views for Separation

    Another reason you may find a Folder Viewer useful is to aid in document separation. "Separation", from Grooper's perspective, is the act of organizing pages into folders. Typically, Grooper will separate during the Separate step of a Batch process (FYI: There are many methods to automate document separation. Your organization may even use a "real time" method that allows documents to be separated at scan-time).

    Examining the Batch Before Separation

    Next, we're going to use a Folder Viewer to gain some insights into how it can be used to fix document separation issues in Grooper. We will review a Batch using a Folder Viewer before the Separate activity runs (just to get a "before" look in to the Batch) and we will review the documents after to see how we can use the Folder Viewer to manually fix the separation issues.

    1. We are going to take a look at the Folder Viewer for this HR Packets Batch.
    2. This Review step will allow us to see the Batch before separation.

    In this Batch, we have a fairly common situation. We've imported a handful of PDFs that are packets of multiple documents. Each PDF has four individual documents contained within that we will need to separate into document folders.

    In this screenshot, we've executed the "Before Separation" Review step to see what the documents look like before the Separate step.

    1. The PDF files have been imported, forming Level 1 of our folder structure.
    2. One page has been split out for each page in the PDF file. We will manipulate these pages to form documents.
    3. The Separate activity will define points where folders are inserted and pages are placed.
      • We should end up with four documents, separated at Page 1, Page 3, Page 4 and Page 6.

    Examining the Batch After Separation


    Let's check out how our Batch looks after document separation.

    1. As you can see here, we completed our review, and Grooper has performed the Separate step.
    2. Now it is waiting for us to review the batch after separation.

    2. After the Separate step runs, pages are placed into folders (according to the separation logic your Grooper designer has configured).
    3. All the newly separated documents now exist at Level 2 in the Batch's folder structure.


    However, our Separate step did not execute perfectly. There are some problems we're going to need to resolve.

    2. Folder (2) is an example of an overly separated document.
      • This is the second page of a W-2. It should not be its own document.
    3. This should be a part of the document folder before it, Folder (1).

    2. Folder (4) is an example of an under-separated document.
    3. The Separate step should have identified Page 3 as a new document, splitting it and the following page into a new folder.


    Next, we're going to look at how to resolve these issues. The process of fixing separation issues is actually very simple. We just need to insert folders and move pages around to make sure they're in the right spots, so that one folder contains all the pages to form a single document. In the case of Folder (2), we need to move its pages to the folder before it. In the case of Folder (4), we'll need to split the document in two, creating a new folder and moving Pages 3 and 4 into it.

    First, we're going to look at our overly separated document. Generally, speaking, we need to move the pages in Folder (2), to Folder (1). Put another way, we need to append those pages to the folder before it. There's at least two ways you could do this. One will require more effort (the hard way) than the other (the easy way).

    Appending Folders - The Hard Way

    Let's start with the hard way. We're going to simply cut a page from a folder and paste it to another.

    • The only thing that makes this "hard" is there's a different way of doing things that requires less clicking around in the UI. There's nothing wrong with this approach. It will certainly get the job done.

    You can move pages around in a Batch by cutting and pasting them into different folders (and moving them up and down in the folder as needed).

    1. Right-click a page you want to move.
      • We want to move the single page in Folder (2) into Folder (1)
    2. Select Cut.
      • Or just select the page you want to cut and use the Ctrl + X shortcut.

    1. Then, right-click the folder you want to move the page to.
    2. Select Paste.
      • Or just select the folder you want and use the Ctrl + V shortcut.

    1. The page we pasted is now Page 2 of Folder (1), as we wanted.
    2. Now, we just have an empty folder we need to get rid of.
    3. Either right-click it and select Delete or use the Delete key on your keyboard.

    Appending Folders - The Easy Way

    Now, there shouldn't be anything mind-blowing and cutting and pasting. You probably do it every day, whether on your work computer or personal computer or even smartphone.

    There is however, a simpler shortcut, the Append To Previous command.

    1. Document (2) has the same problem Document (1) did.
    2. Folder (2) should be appended to Folder (1).

    The Append To Previous command does the exact same thing we did in the previous steps, with the click of a single button.

    1. Right-click the folder you want to append to the folder before it.
    2. Select Foldering.
    3. Select Append To Previous.

    2. All pages in the folder are moved to the end of the previous folder and the empty folder is removed automatically.
      • In our case, the single page in Folder (2) is moved to the end of Folder (1).
    FYI There is also a Prepend to Next command.

    Instead of moving pages to the end of a folder before in a Batch, it would move all pages to the beginning of the next folder.

    We also have a document at was not separated enough. The last document folder contains four pages. But really, this should be split into two folders, containing two pages each. Long story short, the Separate activity failed to insert a new folder at the right point.

    Splitting Folders - The Hard Way

    We'll start with the hard way again. In this case, we want to split the pages in a single folder into two folders. To do this, we'll manually insert a folder, then place the pages forming the document into the new folder.

    • Again, there's nothing really that "hard" about this. It's just going to take more clicks than the "easy" way. That said, it still gets the job done.

    The first thing we need to do is add a new folder. The only (mildly) tricky thing is to make sure you're selecting the right spot to ensure its inserted where you want.

    1. I want the folder added to Document (1).
      • Remember, documents, from Grooper's perspective, are always folders (just a special kind of . Just like in any folder system, you can add a folder to a folder.
    2. It should be at the same folder level of the rest of its folders.
      • The new folder should be a "peer" or a "sibling" with all of these folders.

    Whether its the root Batch folder, a simple folder, or document folder at any level in the Batch, inserting a folder is the same.

    1. Right-click the folder you want to add the folder to.
    2. Select Contents.
    3. Select New Folder.

    2. This will add the empty folder at the bottom of the selected folder.
      • FYI: If you need, you can move folders up and down, just like you can pages, with the Move Up and Move Down commands.


    Now we just need to take the pages out of the wrong folder, and put them in this new folder, correcting the separation issue.

    In our case, the last two pages of Folder (3) need to be moved into Folder (4). This will split out the pages that don't belong to Folder (3) and place them into the new folder.

    1. We will multi-select the last two pages of Folder (3).
    2. Right-click any of the selected pages.
    3. Select Cut.
      • Or simply right click the desired pages and use the Ctrl + X shortcut.

    1. Right-click the folder in which you want to place the pages.
    2. Select Paste.
      • Or use the Ctrl + V shortcut.

    2. The selected pages are moved into the new folder.
    3. Now we properly have two documents separated out instead of just one.

    Splitting Folders - The Easy Way

    Again, there's nothing particularly difficult about this. You're just making folders and putting pages in them to make a new document.

    However, there is a simpler shortcut, the Split Folder command.

    1. Document (2) has the same problem Document (1) did.
    2. Folder (3) should be split into two documents, with the second document starting at Page 3.

    The Split Folder command does the exact same thing we did in the previous steps, with the click of a single button.

    1. Find the page in the folder where a new document should have started. Right-click the page.
      • In our case this is Page 3.
    2. Select Foldering.
    3. Select Split Folder.
    ⚠ DO NOT attempt to use the Split Folder command on loose pages at the root folder of the Batch.

    This would attempt to split the root folder in two, which is akin to splitting the Batch itself in two. This cannot be done with the Split Folder command. If you attempt to do so, Grooper should give you an error message.

    2. The selected page and all subsequent pages are placed into a new folder.
      • In our case, Pages 3 and 4 were placed in a new folder, Folder (4).

    Using Classification Viewer to Resolve Separation Problems

    As mentioned before, the Classification Viewer and the Folder Viewer are extremely similar.

    1. They both use an expandable folder hierarchy to navigate between folders and pages in the Batch.
    2. The only real difference is the Classification Viewer has access to the Document Types panel, which shows the classification statistics generated from a Classify step in a Batch Process.

    You also have access to the same kind of foldering commands in the Classification Viewer that you do in the Folder Viewer.

    1. For example, you can append pages to the previous folder, using the Append To Previous button.

    These commands manipulate the folders and pages in the Batch in the exact same manner. They even use the same keyboard shortcuts!

    2. Continuing using the previous example, pressing the Append To Previous button appended the selected folder's pages to the folder before it, just as we saw previously in the "Append Folders" tab of this tutorial.


    For this reason, if document separation is part of your organization's Grooper process, you may find it most beneficial to review document separation at the same time as document classification using the Classification Viewer. Separation and classification go hand in hand in a lot of ways. It makes a lot of sense to do separation and classification review at the same time, and you have all the tools necessary to do so within the Classification Viewer itself.


    Merging Documents

    One last common command is the Merge Selected command. This allows users to select one or more pages (or even one or more folders) and create a new folder at the same folder level, place the selected pages (or folders) in the new folder, and assign the folder a Document Type. Depending on the circumstance, this can be the quickest way to fix a separation issue and classify a document all at the same time.

    The process looks like this:

    Document (4) has some major separation problems. None of its pages were separated into folders.

    • We could insert one folder for every document, move the pages into the new folders, and assign them a Document Type, but the Merge Selected command is going to end up doing this with a lot less clicks.

    We should have ended up with three document folders:

    1. One containing these pages.
    2. Another containing these pages.
    3. The third containing these pages.

    We'll use the Merge Selected command to merge selected pages into a new folder.

    1. We've multi-selected the first two pages that should be in their own folder.
    2. Right-click any of the pages you want to merge.
    3. Select Foldering.
    4. Select Merge Selected.

    This will bring up the Merge Selected window, allowing you to choose a Document Type.

    1. Press the hamburger icon to bring up a dropdown list of the Content Model's different Document Types.
      • Or, use the search bar to search for one.
    2. Select the Document Type to be assigned to the new folder.
    3. Press the Apply button.

    Upon pressing the Apply button, three things happen.

    1. A new folder is inserted (in the same folder as the selected pages).
    2. The selected pages are placed in the folder.
    3. The folder is assigned the Document Type you chose.

    2. And, in our case, we'd simply repeat the process for the remaining two documents.

    Click me to return to the top

    Shortcuts

    The Folder Viewer keyboard shortcuts are identical to the Classification Viewer shortcuts.

    Scan Viewer

    For some organizations, your first step in document processing will be scanning paper pages into Grooper using a document scanner physically connected to your workstation. In that case, you're going to use the Scan Viewer. This interface allows users to scan paper documents into a Batch as the first step in a Batch Process. A lightweight application called Grooper Desktop installed on your workstation will listen for when the scanner runs and sends images to the Grooper web server as pages are scanned.

    About Grooper Desktop

    If you're using the Grooper Web Client to scan paper documents into a Batch, you'll need to have the Grooper Desktop application installed on your workstation. Grooper Desktop will run as a service on your machine and integrate your scanner with the Scan Viewer.

    When Grooper Desktop is installed and running, you can open it in the System Tray of your Windows Taskbar.

    Click the "Grooper G" icon to open Grooper Desktop.

    After opening Grooper Desktop there are four important properties that need to be configured in order to scan documents.

    1. The Web Scanning property must be Enabled in order to scan paper documents at all.
    2. The Capture Device property defines what kind of scanner you're connecting to.
      • Both ISIS and TWAIN scanners are supported, as well as ColorTrac large format scanners.
      • Less commonly, you can use the Import Device option to import files from a digital source, such as a Windows file folder.
        • FYI: It is atypical to use a Scan Viewer to import digital files into a Batch. It is more to use something called an Import Watcher service rather than the Scan Viewer. This automates the import process for you. That way you, the user, doesn't have to manually import files every time you want to create a new Batch.
    3. The Device Name property allows you to select the scanner model you've connected to your machine.
    4. The Device Settings property allows you to configure any settings available to that model, such as double-sided (or duplex) scanning.
      • Technically, this property is optional. You can use your scanner's default settings, if you so choose. However, most installations will configure your scanner settings to one degree or another.
    FYI If you have a dedicated scanner hooked up to your machine, it's most likely your Grooper designer will have already set this up and you'll never need to worry about configuring Grooper Desktop. However, if you are connecting multiple different scanner models (or connecting a new scanner to your workstation), you may need to be casually aware of these settings.

    After enabling web scanning, selecting your scanner model and configuring any device settings, you will need to start the Grooper Desktop service.

    Press the Start button to start the service.
    ⚠ Any time Grooper Desktop is opened, the scanning service is stopped. You should always press the Start button to run the Grooper Desktop scanning service, even if you open it and don't change any settings.

    Grooper Desktop runs as a service in the background of your machine. When you scan documents using the Scan Viewer, this service will start up your scanner when you hit the Scan button in Grooper and upload the scanned images to the Batch.


    Desktop Scanning in Grooper - 2023

    Batch Management

    Depending on your role in our organization, you may be required to do some "Batch Management". This will require you to manipulate Batches in various ways.

    • Imagine the wrong Batch Process was assigned to a Batch. You would need to stop processing that Batch and assign it the right one.
    • The same would apply if your Grooper designer made changes to a Batch's Batch Process. You'd need to pause the Batch's processing, and update its processing instructions by telling it the Batch Process changed.

    Then, in both cases, you'd need to re-start processing the Batch. You can do all that from the "Batches Page" in the Grooper Web Client.

    In this section we will discuss how to do the following:

    • Pause processing for a Batch in production.
    • Resume processing for a Batch in production.
    • Update a Batch's Batch Process to either reflect changes made by your Grooper designer or update it to an entirely different Batch Process.
    • Reset tasks already completed in a Batch Process in order for them to be reprocessed.

    We will also point out the "Statistics", "Events", and "Details" panels to glean more information about the processing results for a selected Batch.

    Pausing and Resuming Batch Processing

    Pausing Batches

    There are a variety of reasons you may need to pause a Batch in production, but generally speaking, it's a way to momentarily halt any further tasks from processing. Batches can be paused using the "Batches Page".

    1. Notice all the Batches in our list have Ready or Working listed for their Status.
      • A Working status indicates a task is currently being processed. Either Grooper is processing an automated (aka unattended) task or a human user is currently working on a Review task.
      • A Ready status indicates a Review task is ready for a user to pick up and start processing.
    2. The Batch we have selected is currently on the "Data Review" Review step.


    Let's imagine you have a team of reviewers, and we want to hold off on reviewing this data for whatever reason. We don't want any user to process this Batch yet.

    We would need to pause the Batch.

    2. As you can see from the "Tasks Page", our "Data Review" task for that Batch shows up in the task list. Any user could grab this task and start working on it.
      • Pausing the Batch will place it on hold, removing it from any user's task list and preventing processing until it is "resumed".

    1. Batches are paused from the "Batches Page".
    2. To pause a Batch, right-click the Batch you want to pause.
    3. Then, select Pause.

    Grooper will pop up the Pause window to confirm you want to pause processing on the Batch.

    2. Select Pause to pause the Batch.

    1. After pausing, you will see the Batch's Status change to Paused.
    2. Furthermore, take note the "Data Review" step's color changed from gray (indicating it was Ready for processing) to black (indicating the step is awaiting processing).

    Pausing the Batch places its processing on hold, including any Review tasks.

    1. As seen in the "Tasks Page"...
    2. Because the Batch was paused, the "Data Review" step no longer appears as an available task for processing.


    The Batch will remain paused, until a user starts it up again, by "resuming" production on the Batch.

    Resuming Batches

    Once a Batch is paused, it must be "resumed" before any of its tasks are processed. When you're ready to resume processing a paused Batch, by-and-large you will repeat the same process you did to pause it. You'll right-click the paused Batch and press the Resume button.

    1. Batches are resumed from the "Batches Page".
    2. To resume a Batch, right-click the paused Batch you want to resume.
    3. Then, select Resume.

    Grooper will pop up the Resume window to confirm you want to resume processing on the Batch.

    2. Select Resume to resume the Batch.

    1. After resuming, you will see the Batch's Status change to Ready.
      • Or Working if the step is an unattended step in the Batch Process.
    2. Furthermore, take note the "Data Review" step's color changed from black to gray. Processing has resumed on the Batch, and the Review task is ready for processing.

    Click here to return to the top

    Updating Batch Processes and Resetting Steps

    The two largest reasons to pause a Batch are to:

    1. Update the Batch's Batch Process
      • This could be because the wrong Batch Process was assigned to it.
      • This could be because your Grooper designer made changes to the Batch Process and the Batch needs to be aware of the updated processing instructions.
    2. Reset steps already processed.
      • Steps need to be reset if work needs to be "re-done".
      • This often will happen out of necessity when a Grooper designer makes changes to a Batch Process or a Content Model. Imagine a Grooper designer made changes to the Document Types in a Content Model. You may need to reset the Classify step and re-process it so that those changes are reflected in the Batch.
      • This can also be a way to re-enter a Review step if further changes need to be made after the fact.

    Updating Batch Processes

    Imagine a situation where the wrong Batch Process was assigned when the Batch was created. You'd need to tell Grooper you want the Batch to use a different Batch Process.

    Imagine a situation where certain processing instructions in the Batch Process changed. Maybe a step was added. Maybe the configuration of a step was changed slightly. You'd need to inform Grooper of the changes that were made.

    In either case, you would inform Grooper of the Batch's new Batch Process or changes to its existing Batch Process by "updating" the Batch Process.

    The general steps to update a Batch Process are as follows.

    1. Using the "Batches Page", pause the Batch.
    2. Right-click the Batch and select Update Process
    3. Select a "target step" in a Batch Process.
    4. Determine which (if any) steps need to be "reset".
      • Depending on the situation, you may also need to reprocess steps that were previously processed with the updated instructions. We will discuss this in the "Resetting Steps" tab of this tutorial.
    5. Resume the Batch.

    Scenario #1: Updating a Process to Change Processes

    1. The Batch we have selected was assigned the wrong Batch Process.
    2. It was assigned the "URLA Redaction" process, but it should have been assigned the "Invoices Process" process. The documents imported into this Batch are invoices not Uniform Residential Loan Application documents.
    • So, we need to update the Batch's Batch Process to the right one.

    1. A Batch's Batch Process is updated from the "Batches Page".
    2. In order for a Batch Process to be updated, it must first be paused.
      • Right-click the Batch and select Pause to pause the Batch.

    Now that the Batch is paused, we can update its Batch Process.

    1. Right-click the paused Batch.
    2. Select Update Process.

    This will bring up the Update Process configuration window. From here you will select the Batch Process you want to update to.

    The Target Step property determines at which step of which Batch Process you wish to begin processing.

    1. Use the hamburger icon to drop down a list of all published Batch Processes.
    2. Find the Batch Process you want to select and use the arrow icon to expand its steps.
      • In our case, we're changing the Batch's process to "Invoice Process". So, we've found it in the list.
    3. Select the "target step", the step where you want to start processing.
      • In our case, since we're picking an entirely different Batch Process, we're going to go ahead and start all over at the first step, "Split Pages".
    4. When you've made your selection, press the Apply button to update the process.

    1. Since we selected a target step of a different Batch Process, you can see its process has changed to "Invoices Process".
    2. The steps we see in the "Progress" panel are now the steps of the "Invoice Process" Batch Process.
    3. However, the Batch is still paused.
    4. You will need to resume processing to start processing the Batch.

    Scenario #2: Updating the Current Batch Process

    In Grooper, a Batch's process is tied to each Batch when the Batch is created. If a Grooper designer wants to make a change to a Batch Process, they aren't going to alter every single Batch's process individually. There could be a hundred Batches using the same process. That would be a tedious headache.

    Instead, they will make changes to the "working" Batch Process and "publish" those changes. Then, any user can make Batches currently in production aware of those changes by updating its Batch Process with the Update Process button.

    Imagine a scenario where the Grooper designer goofed when creating our "Invoice Process" Batch Process. They added a Data Viewer to a Review step where they should have added a Classification Viewer. They would need to make changes to the "Invoice Process" Batch Process. Then, any Batches currently in production would need to have their processes updated from the "Batches Page".

    1. This should be our "Classification Review" step.
    2. This should be our "Data Review" step.

    The steps are just out of order.

    So, our Grooper designer does his work behind the scenes in Grooper Design Studio, and he tells you that Batch is ready to be updated. No problem. This is even easier than our last scenario.

    1. A Batch's Batch Process is updated from the "Batches Page".
    2. In order for a Batch Process to be updated, it must first be paused.
      • Right-click the Batch and select Pause to pause the Batch.

    Now that the Batch is paused, we can update its Batch Process.

    1. Right-click the paused Batch.
    2. Select Update Process.

    This will bring up the Update Process configuration window. In this case, because we're just updating the Batch's current process and we don't need to reset any steps, truly all we need to do is press the Apply button.

    • FYI: We will talk more about resetting steps in the next tab of this tutorial.

    Remember, the Target Step property determines at which step of which Batch Process you wish to begin processing.

    1. Use the hamburger icon to drop down a list of all published Batch Processes.
    2. You will always default to the current step in your Batch's current Batch Process
      • Since we're updating the Batch's current process, there's no need to hunt for another one. We just need to pick the Target Step where we want to resume processing.
    3. Select the "target step", the step where you want to resume processing.
      • In our case, we're starting the "Classification Review" step next.
    4. When you've made your selection, press the Apply button to update the process.

    Upon applying changes, the Batch's process will be replaced with the updated version.

    1. Our designer fixed the problem and our next step is appropriately configured with a Classification Viewer.
    2. However, the Batch is still paused.
    3. You will need to resume processing to start processing the Batch.

    Scenario #3: Updating a Process and Resetting Steps

    Our two examples were fairly unique in that they did not require us to reset any steps. In many cases, when you're updating a Batch's process, you'll need to reprocess one or more steps.

    There is an entirely separate command you'll need to run on the Batches to do this, using the Reset button. We will show you how to do this in the next tab.

    Resetting Steps

    Often when you're updating a Batch's process, you'll need to reprocess some of the steps in it. This may be because your Grooper designer changed the configuration of one or more activities and the work that's already been done needs to be re-done with the updated changes.

    For example, we have a situation here where a Batch's "Export" step was misconfigured, resulting in failed document exports.

    1. Not only do we need to update the Batch's process once the Batch Process is reconfigured, but we will need to reprocess this "Export" step.

    Let's imagine our Grooper designer has fixed the issue with the Batch Process. Next, we'll need to update the Batch's process with the new Export configuration.

    1. We've paused the Batch.
    2. We'll use the Update Process button to update the Batch's process.

    1. Ultimately, we want to re-process the "Export" step. So, we've selected that as our Target Step
    2. We will press the Apply button to apply the changes in the Batch Process.

    1. Even thought the Batch Process has been updated, the "Export" step still needs to be reprocessed. We will do that by using the Reset button.
    2. In order to reset steps, the Batch must be paused.
    3. Right-click the Batch whose steps you want to reset.
    4. Select Reset.

    This will bring up the Reset configuration window.

    1. The Target Step should be set to the step in the Batch at which you want to resume processing after the Batch is un-paused.
    2. You can expand the Steps To Reset property by clicking the arrow next to it if you need to reset multiple steps.
    3. You will check the box next to each step you wish to rest.
      • In our case, we only need to reset the "Export" step.

    1. This will reset the selected step(s).
      • The step's color changed from red to black, indicating the step is awaiting processing.
    2. All we need to do now is resume processing on the Batch.

    2. With the Batch's process updated, the "Export" step reset, and processing resumed. The step now executes sucessfully.

    Click here to return to the top

    The Progress, Statistics, Events, and Details Panels

    At the bottom of the "Batches Page" you'll find four tabbed panels:

    • Progress
    • Statistics
    • Events
    • Details

    These panels will give you more information about the Batch and the individual steps processed.

    Progress

    We've seen the "Progress Panel" throughout this article. This is the user's primary "at-a-glance" view into the processing progress of each step in the selected Batch.

    Each block represents a step in the Batch Process assigned to the Batch.

    Different colors indicate one thing or another in terms of the processing status of the tasks in that step.

    • Blue = The task has completed.
    • Green = The task is currently being processed.
    • Red = The task failed to process due to some kind of an error.
    • Grey = The task is ready to process.
    • Black = The task is awaiting processing.

    Statistics

    The "Statistics Panel" gives you processing data for each step in the selected Batch's process. The various different Grooper "activity types" will give you different kinds of information. For example, the Recognize activity runs OCR for image-based content. It will, therefore, give you stats on how many OCR characters were found in total for that step.

    Review steps will contain processing information that may be useful when obtaining various kinds of metrics, such as the number of fields edited, the number of keystrokes entered, and total time it took to complete the Review task.

    Events

    The "Events Panel" displays a step-by-step list of all tasks submitted for processing. This is designed to give you an audit trail into the processing history for the Batch.

    This will include "Audit" type events when tasks are submitted and completed, when the Batch was created, if and when it was paused, when processing was then resumed, and more.

    This panel is often most helpful when processing errors occur, logged as "Error" events. This will give you additional information to help you troubleshoot the issue.

    Statistics

    The "Statistics Panel" gives you the most "top level" details about items in the Batch. This will give you total counts for pages, folders, Content Types, Attachment Types and various other information pertaining to them.

    Retrieved from "https://wiki.grooper.com/index.php?title=2023:Review_User_Guide&oldid=12502"