Automated image projects

Download Quick Guide

Creating an image project

What to know

• Blackout can redact and highlight images in Relativity in bulk utilizing terms, phrases, and regular expressions.
• Image projects are a flexible way to target specific documents and locate content that must be marked up for the case.
• There is a wide range of workflows that can be utilized to accomplish many different goals. However, for this guide, we will focus on the basic tooling.
• When setting the word/phrase rules for the project, either in the project itself or through a CSV template, wildcards and dtSearch are not supported.
• Foreign language support is not enabled by default. To enable this setting, review the steps here.

How to do it

1. Log into Relativity as a system administrator or a user who has access to the Blackout tab.
2. Navigate to Blackout | Projects
3. Click Create New Project
4. Select Image
5. Fill out the project form using the details below.
6. Click Create
7. A form will be displayed that is used to define how the project will function when running.

The following fields make up the form:

Field	Description
Project Name	An identifier for the project, making it easy to refer to or for others to find.
Language Code	Code for specifying which languages should be used for OCR. Not required to be changed away from 'eng' unless other language packs are installed.
Saved Search	The document source. All actions will be performed against these documents.
Markup Set	The markup set that the markups will be associated with.
Pre OCR and prepare documents for Find & Redact	Generate OCR for a set of images without running rules. This is required for Find & Redact to prepare documents larger than 2 pages. This can also be used for larger image projects to pre-OCR images before running rule-based projects.
Rules	The rules are what drives the project. How they are configured will determine how the content will be located on each image. There are two different kinds of rule groups: Redaction and Highlight. Every time the fields are filled out for a rule group, another section will appear below it, allowing the creation of multiple rule groups. Each of these can be independently configured, targeting different sets of content.
Redact	Redaction rules will result in removing the content from the document and replacing it with a black, white, text, cross, or inverse redaction.
Highlight	Highlight rules will result in changing the foreground color of the matched content with the selected color.
Markup Reason	A simple message that can be associated with the markups made by the rule group. It can be reviewed using the Blackout reports. This field is not required to run the project.
Markup Scope	Defines the behavior of the project when placing a redaction or highlight. Word: Will place the markup of the entire word where the match is located Line: Will place the markup on the entire line of text Paragraph: Will place the markup between the most recent previous carriage return and the next carriage return Page: Will place a markup on the entire page Character: Will place a markup on selected characters where a match is located
Gear Button	Clicking the gear button will expand another option called Expand Markup to Full Width of Page. Checking this box will take any markup that is made for the rules and extend it from the left margin of the image to the right margin of the image.
Markup SubType	Specified whether to place a Black, White, or Text markup. If text is selected, custom text can be input for the markup. Otherwise, the default text markup language will be used. If the highlight is selected, this option will be replaced with choices for yellow, blue, green, orange, pink, and purple.
Word/Phrase	'Word/Phrase' are the words, phrases, and text that will be marked up across the document set. Multiple words or phrases can be added to a single rule group. These rules are case insensitive.
Regex	Regular expressions can be used to identify important patterns like email addresses, social security numbers, credit card numbers, and any other content that may appear regularly throughout the document set that constitutes sensitive information. Regular expressions require a name and an expression. After saving the project, these regular expressions will be available to be selected by name on other projects within the same case. Rules do not need to be added to create a project. Rules may be uploaded afterward, or the project can be edited and rules can be added later. Blackout uses the PCRE flavor of regular expression Runs with the global flag (/g), multi-line flag (/m), and case insensitive flag (/i) enabled by default. The case-insensitive flag can be modified inline by adding (?-i) to turn off and (?i) to enable.
dtSearch	dtSearch includes special characters and other operators that you can use to define search criteria. Noise words are not supported in Blackout Projects. Use the W/N operator for proximity matches in these instances. By default, all terms in a dtSearch rule will be redacted. The terms should be surrounded with (?<redact>{term}) to redact part of a dtSearch rule. To redact more than one term, multiple named groups must be defined.

Uploading rules from CSV

After successfully creating the image project, rules can be added via uploading a template CSV.

What to know

• Blackout can support up to 100,000 rules being created via CSV.
• Extra rules mean each doc takes a little longer to process as the image content needs to be checked for each rule.

The CSV template and explanation of the columns can be downloaded from the Milyli Support Center.

How to do it

1. From the project view screen, click the Upload Rules CSV button
2. Navigate to the completed CSV
3. Click the upload icon
4. If the CSV file is valid, the rules will be created and the button will display how many rules have been uploaded.

Running an image project

What to know

• When a project is in a valid state to run, with a name, saved search, markup set, and rules, the status message of the project view screen will display that is ready to run.

How to do it

1. After creating the project, click the Run button.
2. Once the project begins running, the project will be queued and then start being processed by the Blackout agent.

While the project is running

While the project is running, the items below will be displayed.

Message	Description
Documents prepared	The number of documents that have been distributed to the system as tasks for markup review
Documents completed	The number of documents that have completed markup review, which can result in markups being placed if there are rules matches
Progress bar	Displays the number of actions completed for preparation and review. The light blue color represents document preparation and the dark blue represents markup review.
Time elapsed	The total time the project has processed
Current activity	The current activities that the Blackout Agent is performing which includes preparation, reviewing, and marking up documents. The most recent action is shown at the top along with when it began. Clicking the down arrow will display the currently active tasks and when the task began.
Stop	Begins the stop operation, which will cancel all unfinished work for the project. Work completed to that point will remain and can be reverted.

3. After the project completes, a results page displays information about the completed work and provides a launchpad for further quality review.

When the project is completed

When the project compeletes, the items below will be displayed.

Message	Description
Documents marked up	The total number of documents that have had markups placed on them by Blackout using the rules from this run. Clicking this button will navigate to a saved search of all of the documents that have been marked up.
Pages marked up	The total number of pages for each marked up documents that have had markups placed on them. Clicking this button will navigate to a saved search of all of the documents that have been marked up.
Markups placed	The total number of markups that have been placed, both redaction and highlight, by Blackout using the rules from this run. Clicking this button will navigate to a saved search of all of the documents that have been marked up.
Documents to review	While Blackout does not require extracted text to review an image and place markups on, whenever extracted text is available, it will perform an automated quality review on the placed markups. This happens by looking at the number of matches that the project terms have in the extracted text and comparing them to the number of placed markups. If a rule has placed less markups than there are matches in the extracted text, the rule is flagged and the document is marked for review. Clicking this button will navigate to a saved search of the documents requiring review.
Documents with warnings	The number of documents that encounter a non-project breaking warning. Clicking this button will download a CSV of the document warnings that include information about what happened and any error messages. Providing this CSV to Milyli can greatly assist in diagnosing any issues that may have been encountered. Typical reasons for warnings include missing images or failed OCR.
History	Projects are incremental. After the project has run, clicking the Run button again will queue the project up again and process any new documents in the document source. As a result, the history section will show the work that the project has performed and the results of said work.

Reverting an image project

What to know

• Occasionally the markups created by a image project may need to be reverted.
• This need may arise when rules need to be modified or a case settles.

How to do it

1. From the project view screen, click the Revert button
2. A dialogue will appear; confirm reverting the project
3. The project will be queued
4. The Blackout agent will pick it up and begin reverting the documents marked up by the project
5. A progress bar will display how many documents have been reverted