Metadata Enrichment

Configuration

General Information

In order for the enrichment processes to be applied, the new metadata must first be added to the aggregated metadata keys in the index settings. For example:

Entity Recognition

Introduction

This chapter deals with the concept, setup, and troubleshooting methods for configuring entity recognition.

Entity recognition configuration

In this chapter, the concept of entity recognition is explained using a simple example.
To set up entity recognition, follow these steps:

• Connect to the Management Center.
• Navigate to the Configuration menu and then to the Index tab.
• Activate the Advanced Settings and open the index settings of the index you want to configure with Entity Recognition.
• Search for the setting “Entity recognition parameters“ in the Management Center
• Define your Entity Recognition rules in the “Pattern Rules” field to match your metadata.
• The following rule formats are supported: https://github.com/google/re2/wiki/Syntax

In our concrete example:

rule=/\// digits /\//. 
digits=/\d+/.

Explanation:

The first rule defines that all numbers between two slashes should match (regex):
Example: test/1234test1234/test/543/test (543 is extracted)

Now add a new “Metadata Definition“ to apply the rules for metadata.

Example:

In this example, Mindbreeze searches for numbers between two slashes in the string of the existing metadata with the name “fullstring”. If there are numbers between two slashes, Mindbreeze takes the part of the matches configured in the sub-rule “digits” and writes it as a string in the new metadata “myextractedVal”.

Example: 
fullstring: xyz/1234/herbert543/345test
Match of the rule “rule”: /1234/
Value of the rule “digits”: 1234
Value of the metadata myextractedVal==1234

Notes on configuration in the Management Center

When configuring as a metadata in Mindbreeze InSpire, the following fields must be filled in:

• "If Rule Matches": Name of the rule.
• “Name": Name of the metadata
• "Value": Value of the rule. Example: "{{month}}" - can also be normal text or composite. Ex: "Date {{day}}.{{month}}.{{year}}".
• "Format": format of the rule. "string", "date", "number".
• "Format Options": formatting options - especially for date  as for simpledateformat
• "In Existing Metadata": scope where the rule will be applied e.g.: content, title, datasource/mes:key, <own metadata>, ...)
• Scope: With the Scope setting it is possible to select an area or several areas with one entity recognition rule, in which the rules for extraction are to be applied. For this purpose, the name of the rule for selecting the area(s) is entered in the scope field. In contrast to value extraction, you have to enter the name without {{}}.

Entity recognition (example: file system)

This chapter uses a simple example to explain entity recognition and its setup with Mindbreeze.

Configuration of entity recognition for a file system:

Pattern Rules:

host=/[^\\]+/.

share=/[^\\]+/.

directory=/[^\\]+/.

UNCPath="\\\\" host "\\" share "\\" directory "\\".

Metadata-Definition 1:

If Rule Matches: UNCPath
Name: FileShare
Value: {{share}}
In Existing Metadata: datasource/mes:key

Metadata-Definition 2:

If Rule Matches: UNCPath
Name: ProjectPath
Value: {{directory}}
In Existing Metadata: datasource/mes:key

Aggregated Metadata Keys (; separated): FileShare; ProjectPath

Date formats for entity recognition are based on the ICU patterns (e.g. locale … de_AT)

Configuration for entity recognition for file system paths (variant 2) – with exceptions:

If the rules are ambiguous, alternative rules, ordering by naming and correct ordering of the metadata extraction definitions can be used to enable more complex usecases. The path as metadata is lower-case and thus better for CSV mapping.

Even a complex use case, in which the rules are ambiguous can be achieved using alternative rules and sequencing by name, as well as the correct sequencing of the multiple metadata extraction. The path, a metadata, is lower-case and thus better for CSV mapping.

Note: An OR (|) operator of sub-rules does not work!

Simple solution without exception:

Pattern rules:

LWPath=/\\\\[^\\]+\\[^\\]+\\[^\\]+\\[^\\]+/.
FilePath=/[^\\]+/.
FullPath=LWPath "\\" FilePath.

Solution with an exception (data\it):

Pattern rules:

ASpecialPath="data\\it".
OtherPath=/[^\\]+/.
BaseShare=/\\\\[^\\]+\\[^\\]+\\[^\\]+/.
LWPathA= BaseShare "\\" ASpecialPath.
LWPathOther= BaseShare "\\" OtherPath.
FilePathA=/[^\\].*/.
FilePathOther=/[^\\].*/.
FullPathA=LWPathA "\\" FilePathA.
FullPathOther=LWPathOther "\\" FilePathOther.

The following screenshot demonstrates the configuration of the rules.

CSV-transform: the extracted value (file share) is case sensitive, so the cases must match. That way the path can be used as source metadata.

fileshare;letter
\\fileserver.mycompany.com\qa\fstest\projekte;U:
\\fileserver.mycompany.com\qa\fstest\vorlagen;T:
\\fileserver.mycompany.com\qa\fstest\allgemein;G:
\\fileserver.mycompany.com\qa\fstest\spezial;M:
\\fileserver.mycompany.com\qa\fstest\data\it;H:
\\fileserver.mycompany.com\qa\fstest\data;H:
\\fileserver.mycompany.com\qa\fstest\data-services;H:
\\fileserver.mycompany.com\qa\fstest\allgemein-retail;G:

Note: A match with mes:key only works in CSV transformation (as well as in ER rules) using: In Property = datasource/mes:key.

Attention: /documents-Servlet does not provide values that are created via index re-invert!

Troubleshooting entity recognition

This chapter deals with troubleshooting the entity recognition rules.

Important information

1. In Mindbreeze InSpire, regular expressions are surrounded by a “/”.
2. Each rule entry must be separated by a period.
3. Rule names may not contain “_”
4. Rules are “greedy”, meaning they match as much as possible be careful with “.*” or “.+” configurations).
5. Rules are processed alphabetically (case-sensitive!). First in line are uppercase letters from A to Z, then lowercase letters from a to z.
6. If a rule matches an entity, no second rule can match. Assumption: If the words “managing board” are used both for the committee and in the keyword, only the metadata with the rule "committee" will include the words “managing board”.
7. Entity recognition rules can only be created per index, that is, across all data sources within the index.

Index

Check the index status at http://localhost:8443/index/<Indexport>/statistics

Privileged servlets:

• Connect to the Management Center
• Navigate to the Configuration menu, then go to the index tab
• Activate the advanced settings and open the index setting of the index, which you want to use for testing Entity Recognition
• Deactivate the “Disable Unrestricted Privileged Servlets” checkbox
• Then save the settings and restart the services
• After the services are restarted:
  • Open https://yourappliance:8443/index/Indexport (in our example: 23101/processitems) https://yourappliance:8443/index/23101/processitems
  • On this page, you can test the rules (pattern rules) with a specific query (e.g. ALL)
  • After entering the patterns, click on process. If the syntax of the rules is valid, you’ll have more options to test after pressing the button.
  • Select the rule that you want to match and configure the values ​​of the rule(s).
  • Then click on process to start testing the rule(s):
  • 
    
    

Deactivating the greedy strategy of the entity recognition rules

Entity Recognition rules are usually greedy. In the following example, the selected rows are matched:

Rule

R1=/ (?s)(test)(?P<line>.+)\s+(.*Page) /.

Match:

If greedy is deactivated, not everything is matched. Instead, only those blocks that start with test and end with Page are matched:

Rule:

(?U)(?s)(test)(?P<line>.+)\s+(.*Page)(?U)

Match:

Common error sources

An error with the following error message occurred while parsing the ER rules:

“MesQuery::Text::RE2Tokenizer ERROR: Matched empty (epsilon) token, pattern is”

… for instance, a „\“ at the end of a regex is not supported (LWPath=/\\\\[^\\]+\\/. … and an error occurs. Better: LWPath=/\\\\[^\\]+/ “\\“.).

There can also be possible problems with “.*“ in rules.

Entity recognition rules are analyzed in alphabetical order and the first complete match wins.

Regex rules for German words do not match all characters (umlauts, etc.) with \w. Instead, you can use \pL to match all unicode characters.

Typical use cases

Personal information

Social security number

RegEx

\d{4}(\s|\.|\-)\d{6}

Example

1237 010180
1237.010180
1237-010180

Telephone number

RegEx

(\+)([\s.\(\)]*\d{1}){8,13}(-)?(\d{1,5})

Example

+43 732 606162-0
+43 732 606162-609
+49(732)606162-609

Number (with delimiters)

RegEx

z1=/\d/.z2=/\d/. (…)Dlmtr=/[\s\-_.:]?/.
z1 Dlmtr z2 Dlmtr z3 Dlmtr z4 Dlmtr z5 Dlmtr z6.

Example

12-34567
12 34 56-7
1-2 3456.7

Amount

RegEx

((\d{1,3}(\.(\d){3})*)|\d*)(,\d{1,2})

Example

0.84
100,000.49
100,000.00
1,000,000,000,000.00

Date

Handbook for date formats: http://userguide.icu-project.org/formatparse/datetime

• dd(.|-|/)MM(.|-|/)yyyy
  • RegEx
    ((0[1-9])|[1-9]|([1-3][0-9]))(\.|\/|-)((0[1-9])|[1-9]|10|11|12)(\.|\/|-)((19|20)\d{2})
  • Example
    11.03.2014
    11.3.2014
    3.3.2014
    03.2.2010
    11/03/2014
    11/3/2014
    3/3/2014
    03/2/2010
    11-03-2014
    11-3-2014
    3-3-2014
    03-2-2010
• dd. MMM yyyy
  • RegEx
    ((0[1-9])|[1-9]|([1-3][0-9]))\..(|January|February|March|April|May|June|July|August|September|October|November|December).((19|20)\d{2})
  • Example
    3 January 2014
    4 February 2012
    30 November 2013
• MMM yyyy
  • RegEx
    (January|February|March|April|May|June|July|August|September|October|November|December).((19|20)\d{2})
  • Example
    February 2014
    September 2014
• MM(.|-|/)yyyy
  • RegEx
    (January|February|March|April|May|June|July|August|September|October|November|December).((19|20)\d{2})|((0[1-9])|[1-9]|10|11|12)(\.|\/|-)((19|20)\d{2})
  • Example
    03-2014
    03.2014
    03/2014
• yyyy(.|-|/)mm(.|-|/)dd
  • RegEx
    ((19|20)\d{2})(\.|\/|-)((0[1-9])|[1-9]|10|11|12)(\.|\/|-)((([1-3][0-9]|0[1-9])|[1-9]))
  • Example
    2014-03-21

• Date-Regex total
  ((0[1-9])|[1-9]|([1-3][0-9]))(\.|\/|-)((0[1-9])|[1-9]|10|11|12)(\.|\/|-)((19|20)\d{2})|((0[1-9])|[1-9]|([1-3][0-9]))\..(January|February|March|April|May|June|July|August|September|October|November|December).((19|20)\d{2})|(January|February|March|April|May|June|July|August|September|October|November|December).((19|20)\d{2})|((0[1-9])|[1-9]|10|11|12)(\.|\/|-)((19|20)\d{2})|((19|20)\d{2})(\.|\/|-)((0[1-9])|[1-9]|10|11|12)(\.|\/|-)((([1-3][0-9]|0[1-9])|[1-9]))
• Date-Regex total II
  ((((0?[1-9]|[12]\d|3[01])[\.\-\/](0?[13578]|1[02])[\.\-\/]((1[6-9]|[2-9]\d)?\d{2}))|((0?[1-9]|[12]\d|30)[\.\-\/](0?[13456789]|1[012])[\.\-\/]((1[6-9]|[2-9]\d)?\d{2}))|((0?[1-9]|1\d|2[0-8])[\.\-\/]0?2[\.\-\/]((1[6-9]|[2-9]\d)?\d{2}))|(29[\.\-\/]0?2[\.\-\/]((1[6-9]|[2-9]\d)?(0[48]|[2468][048]|[13579][26])|((16|[2468][048]|[3579][26])00)|00)))|(((0[1-9]|[12]\d|3[01])(0[13578]|1[02])((1[6-9]|[2-9]\d)?\d{2}))|((0[1-9]|[12]\d|30)(0[13456789]|1[012])((1[6-9]|[2-9]\d)?\d{2}))|((0[1-9]|1\d|2[0-8])02((1[6-9]|[2-9]\d)?\d{2}))|(2902((1[6-9]|[2-9]\d)?(0[48]|[2468][048]|[13579][26])|((16|[2468][048]|[3579][26])00)|00))))
• Example
  31.12.2005
  12.12.12
  1.2.2003
  1.3.98
  04-05-2004

Time

RegEx

(([0-1]?[0-9])|([2][0-3])):([0-5]?[0-9])(:([0-5]?[0-9]))?

Example

11:00:23
12:30

E-mail

RegEx

([a-zA-Z0-9_\-\.]+)@([a-zA-Z0-9_\-\.]+)\.([a-zA-Z]{2,5})

Example

david.porter@inspire.mindbreeze.com
egov@mindbreeze.com

IBAN

RegEx

AT\d{18}

Example

AT002105017000123456

Split List by “,” or other symbols

In this example a list of entries separated by semicolon will be interpreted as well as List in Mindbreeze InSpire.

Input: List of word, word,…

value=/[^\s,][^,]*[^,\s]?/.
rule= /\s*/ value /\s*(,\s*|$)/.

Catalog enrichment

New metadata is added to a document. The new metadata is created using a “catalog” (e.g. CSV file).

The following options are outlined here:

• Index CSV transformation (functionality already available in the index)
• HierarchicalCSVEnricher (separate ItemTransformation plugin)
• FileMetadataEnricher (separate ItemTransformation plugin)

CSV Transformation

This section focuses on metadata enrichment using a CSV file. In doing this, it is possible to compare the value of a metadata with the value of a particular column in the CSV. If the value from the metadata matches the value from the column, you can write the value of another column from the same row to a new metadata and attach it to the result.

Setting up the CSV transformation

This chapter uses a concrete example to illustrate setting up CSV transformation. The following steps must be performed for the configuration:

Connect to the management center (default: https://IhreAppliance:8443).

Navigate to the Indices tab, enable the Advanced Settings, and then expand your Index.

Search for the “CSV Transformation” settings and set the function as shown in the example below.

 Example:

CSV File Path: Path on the server of the CSV file copied by you

If Expression Matches: The name of the column in the CSV, which must match the value of a metadata for the transformation to be performed

In Property: The existing metadata which should be compared with the value from the column in the CSV

Name: The name of the metadata which should contain the new enriched value

Value: The name of the column, the value of which is to be written to the new metadata "Name" in order to enrich the result

Copy any CSV file to the /data/ directory on your Mindbreeze InSpire.

How does it work?

The value of the existing metadata (“medication”) is compared with the value from the column “Medikament”. If a row was found in which these two values were equivalent, the value is extracted from the ATC_CODE column and attached to the metadata ATC_CODE.

Changes in the CSV file using a spreadsheet program

If you edit the CSV using a spreadsheet program such as Excel, you must ensure that the CSV is still in UTF8 format rather than UTF8-BOM format after processing.

You can check this with any text editor such as Notepad++ and, if necessary, convert it back to the UTF8 format.

Synthesized Metadata

Using Synthesized Metadata Definitions, rules can be created to enrich documents with metadata. These rules are dynamically evaluated at the time of the search.

Precomputed Synthesized Metadata

Using Precomputed Synthesized Metadata Definitions, rules can be created to enrich documents with metadata. These rules are evaluated during index inversion to synthesize metadata.

Please note that only a subset of Mindbreeze Property Expression Language is supported for Precomputed Synthesized Metadata Definitions. All expressions in Mindbreeze Property Expression Language - Search specific language elements are not supported, which also affects expressions that contain references, such as paths or inverse references. If you want to use expressions with references, you can create Synthesized Metadata Definitions instead.

Please note that predefined Metadata, like mes:key, Category, CategoryInstance and FQCategory are not available for Precomputed Synthesized Metadata Definitions. Instead, Synthesized Metadata can be used.

The following options are available:

Hierarchical CSV Enricher (ItemTransformer.HierarchicalCsvEnricher)

Introduction

This plugin annotates special words in PDF documents, which are specified in a CSV file. These annotated words (matches) are displayed in the PDF preview as clickable hyperlinks. Custom actions can be linked to these hyperlinks in a custom client. For example, a mouse-over on a particular word can display an explanation in the form of a tool tip. Additionally, the plugin can generate metadata for matches and annotate entities.

Installation

To use the plugin, the MetadataTransformationService plugin must be added to your Mindbreeze installation by loading the appropriate plugin.

Configuration

The plugin is generally started as an independent service (much like caching principal resolution services). References to the service are configured in the indices. The plugin itself is configured in the section “Services”:

• Enable the plugin for each desired index using the Management UI:
  • Go to the “Indices” tab and enable “Advanced Settings”
  • Scroll down to “Services“
  • Select the “ItemTransformationServicePlugin.HierarchicalCsvEnricher” plugin and click “Add”.

Basic settings

Configuring CSV transformation

Document exclusion rules

Usually, every document in the index that has configured the transformation service is transformed. With this rule, however, certain documents can be omitted on the basis of metadata and will remain unchanged.

Attributes

Metadata can be modified for matches based on rules from regular expressions. Add attributes with the plus icon.

Configuring the index

In order for the index to use the configured service, the service must be referenced in the index settings under "Item Transformation Services." To do this, click "Add" and select the service you have specified.

A property "launchedserviceid" is automatically created; there is no need to change it.

Use as a ContentFilter service

With the installation of the MetadataTransformationService plugin, a filter plugin "FilterPlugin.HierarchicalCsvEnricher" is also registered, which can be enabled for the extension “textcatalogenricher." Navigate to the "Filter" tab in the settings. The configuration is similar to the "ItemTransformer.HierarchicalCsvEnricher" plugin. The following metadata must be set in the filter request:

Enable the "Advanced Settings" and add settings for the “FilterPlugin.HierarchicalCsvEnricher – Plugin" under "Global Filter Plugin Properties."

File Metadata Enricher

This chapter deals with the use of the File Metadata Enricher. This plugin allows you to enrich indexed documents (e.g. PDF files) with external sources such as an XML file or a CSV file. This chapter differentiates between XML file metadata enrichment and catalog settings.

Activation

The file metadata enricher is available as an ItemTransformationService plugin and as a PostFilter plugin.

You can configure as the item transformation service for an index service:

The file metadata enricher for a filter can be configured as a post filter plugin. The transformation then happens after filtering.

XML file metadata enrichment

This mechanism is very similar to the mechanism of CSV transformation. In essence, this is about the possibility of comparing the value of a metadata with the value in an XML file. If, for example, there is a file with content (e.g. mindbreeze.pdf) in a data source and another file that contains the metadata separately (e.g. mindbreeze.xml), they can be merged into one result to link the content to the metadata. This mechanism is explained in more detail in the following example:

Configuration example

Explanation

File Path Source: Name of a metadata used as a source for the enrichment. For example, a metadata containing the path of the current result can be used. For instance, in the Microsoft File Connector, datasource/mes:key contains smb://myserver/testdaten/Content/.

File Path Pattern: Limits the enricher’s scope of application. All results where the mes:key values do not match the regex regex from File Path Pattern are ignored (The enricher is not applied).

File Path Replacement: The path that contains the metadata of a file is specified here. The files must be located locally on the appliance or at least mounted on it. It is possible to reference a matching value of the regex specified in File Path Pattern as a variable here. The matching groups (REGEX) can be referenced in ascending order with $1 (e.g. $1, $2, $3, ...). The group (.*) can therefore be referenced with $1. In our case, the name of the file is extracted from the string that matches the File Path Pattern.

Metadata Node XPath: Each XML node that is made by this XPATH is interpreted by the enricher as an object with metadata.

Metadata Key XPath: The string that is matched by this XPATH expression is used by the enricher as the name of the new metadata.

Metadata Value XPath: The string that is matched by this XPATH expression is used by the enricher as the value of the new metadata.

Date Format: If a format is specified in Java Simple Date Format, the enricher will try to interpret each string that is matched by Metadata Value XPath as a date in the specified format to provide the entire functionality of the Mindbreeze date format. If the string is not in the specified format, the enricher performs a fallback and interprets the matching string as a string.

Example

This chapter uses a concrete example of the enricher for illustration.

Configuration

XML file (1.xml)

<?xml version="1.0" encoding="utf-8"?>

<Document>

<UserID>4711_12</UserID>

<DocID>PDF_4711_12_CV_001.pdf</DocID>

<DocType>CV</DocType>

</Document>

Explanation

If it matches the regex from the file path pattern, the metadata datasource/mes:key is compared with all local or mounted file names from the file path replacement path for each result that was configured for the index on which the metadata enricher is configured. In so doing, the file name that was defined as a regex group in the file path pattern is used in the File Path Replacement at the reference point.

Example:

Source file: …/1.pdf File Path Replacement: …/1.xml

If the paths match, the XML node /<Document>/* is searched for in the .xml and all child nodes of the node are interpreted as relevant information. The name of the node is interpreted as the metadata name of the new metadata to be created. If the current child node contains a text(), this is set as the value for the newly created metadata, and the metadata is attached to the current result in the index.

Example:

In our case, the following indexed metadata would be attached to the already indexed file 1.pdf:

UserID: 4711_12
DocID: PDF_4711_12_CV_001.pdf
DocType: CV

Catalog Settings

This mechanism uses a CSV file for enrichment. As with CSV transformation, information from Mindbreeze is compared with the value of a column in the CSV. Unlike CSV transformation, the metadata cannot be selected for comparison because the plugin is actively searching for matches in the content of the file. Another important function of the plugin is the recognition of negations. If, for example, there is a match for renal failure, but renal failure is not mentioned in the text, renal failure is attached to the result as a negation in a separate metadata. Additionally, this feature allows automatic links to be attached behind the hits and visualized in the PDF preview. The detailed operation of this function is explained in the following section.

Configuration example

Explanation

Catalog File: This setting includes the path of the CSV file to be used for enrichment. The file must be located locally on the Mindbreeze InSpire Appliance or mounted on it.

Catalog Match On: This setting specifies which column of the CSV file is compared with the information from the content of the results in order to recognize a match.

Extract Metadata: In this field, the name of the metadata is specified in which the text of the Metadata Value column is inserted if there is a match. The metadata is attached to the result.

Extract Negated Metadata: With this setting, you specify the name of the metadata that contains the text of the Metadata Value column if it is matched in the case of a negation. This metadata is attached to the result. The match must be recognized as follows:
Negation Prefix Pattern + Catalog Match On (string for matching) + Negation Postfix Pattern

Extract Metadata Item: This metadata contains a structured form of the entire applied CSV. This metadata is not intended to be a filter and is only intended to support the development of Insight Apps.

Metadata Value: This field must contain the column name of the CSV defined in the Catalog File setting. If the value of the Catalog Match On column matches the string currently compared from the content of the result, the value of the column specified in the Metadata Value field is attached as a string to the metadata Extract Metadata.

Link HREF Pattern: In this setting, a link can be assembled using the extracted metadata. This link is then available in the PDF preview of the client. This link can be interpreted by the developer of the Insight App. The format of the link can be specified as follows:

https://entity.mindbreeze.com/meddra/?code={{pt_code}}

Instead of the placeholder {{pt_code}}, the value that was actually extracted is inserted in the column pt_code from the CSV at time of inversion.

Catalog ID Column: This setting determines which column of the CSV file is unique. This is used internally by Mindbreeze.

Sentence Split Pattern: This setting serves to divide the content of a sentence into sentence parts. The enricher is applied only in those parts of sentences that correspond to the regular expression given here.

Negation Prefix Pattern: This pattern specifies the prefix (usually text) to be used to recognize a negation. Here again, syntax means the syntax of the regular expressions.

Negation Postfix Pattern: This pattern specifies the Postfix (usually text) to be used to recognize a negation. Here again, the syntax means the syntax of the regular expressions.

Replacement Patterns: This field allows certain occurrences of words, sentences, or letters to be synonymous for the enricher. For example, this means it would be possible for ä to be ae and also for ae to be ä. The syntax for this is shown in the following example:

(?i:ä|ae)|>(ä|ae)
(?i:ae|ä)|>(ae|ä)

The symbol, |>, is used as separator. Each rule must be entered as a separate line in the configuration field.

Using the new metadata

The document Development of Insight Apps illustrates how the metadata can be used in a PDF preview when developing Insight Apps.