Release Notes for Mindbreeze InSpire

Version 21.2

Innovations and new features

AI-based content labeling with Insight Services

Text classification with Mindbreeze InSpire has never been easier. Tag a portion of your documents with predefined labels. With the help of Mindbreeze Insight Services and Machine Learning, Mindbreeze InSpire is able to expand your knowledge and store it for future use cases. Based on this knowledge, all other documents can subsequently be classified fully automatically.

Labeling can be done easily and directly via the Insight app without a predefined data set. So you can start without having a predefined dataset.

The main steps to perform this use case are:

1. Preparation of the training dataset
   1. Define the possible labels for the documents.
   2. Manual labeling via the Insight App for documents and creation of the dataset to train the classification model.
2. Train a classification model
3. Labeling of documents from an index using the classification model in the Semantic Pipeline (Item Transformation).

Link to the documentation

Seamless use of Insight Apps in Microsoft SharePoint Online

Through some technical enhancements, users now have the ability to seamlessly leverage our variety of available Insight Apps within Microsoft SharePoint Online.

A simple single sign-on (SSO) configuration now enables users logged into SharePoint Online to also be automatically logged into the embedded Mindbreeze InSpire Insight app (via OAuth 2 and JWT).

Link to the documentation

Usability improvement in the administration of the search experience and best bets

Thanks to the optimized table editor, large amounts of best bets, synonyms or other data can now be edited effortlessly in the Search Experience tables in Mindbreeze Management Center. Using the new filter function, the relevant rows can be displayed and edited directly. All rows that are not required can be hidden so that you do not lose focus on the essentials. In addition, you can now sort the tables by column. This makes it even easier to find relevant data quickly.

Link to the documentation

Personalized search results

Our new Mindbreeze InSpire 21.2 release again includes optimizations in the area of personalization. The Mindbreeze InSpire relevance model, for example, ensures that the preparation of results is optimally adapted to the different requirements of individual employees, positions and teams. By parameterizing the Insight Apps with context information such as the department or location, documents or results can be weighted differently depending on the current user and the assigned department. If the department is added as information to the Insight App, an application can be that users in Sales accordingly receive primarily results such as offers, presentations, market analyses, contracts, projects, while a user from Purchasing receives offers and those responsible from Accounting receive open orders, open invoices or reminders.

Link to the documentation

Optimized responsive layout

As part of the Mindbreeze InSpire 21.2 release, some optimizations were made to the layout of the Management Center, thus improving its use on small screens (smartphones, tablets) for end users. In addition to numerous detail improvements, the left menu bar is now automatically hidden to create more space for the really relevant content.

Custom Client Services (in the Management Center)

Since the last release (Mindbreeze InSpire 21.1 release), the Mindbreeze Management Center (MMC) supports all configured client services for Insight Apps. This includes the Insight App Designer, search results from the selected client service as well as the Search Experience. In addition, there is now also the possibility to specify custom client service URLs.

Link to the documentation

Advanced connectors

Microsoft Teams Connector (New)

The new Microsoft Teams connector is able to efficiently index messages and threads from all Teams channels. Attachments (managed in Microsoft SharePoint Online) can be conveniently referenced in this way. The connector supports efficient incremental processing of messages so that the search is always up-to-date.

Link to the documentation

Microsoft Dynamics CRM Online OAuth 2.0 Client Credentials

In the Mindbreeze InSpire 21.2 release, the Microsoft Dynamics CRM connector supports authentication via an Azure app (OAuth 2.0 client credentials). This authentication method is also suitable if you have a federated IDP (e.g. Okta) in use through which users normally log in.

Link to the documentation

Opentext Documentum connector: Deleting documents without audit trail

Normally, the Opentext Documentum connector detects deleted documents based on the audit trail. If this audit trail is not available, this enhanced version of the Opentext Documentum connector now also allows regular removal of deleted documents via a regular database query.

Link to the documentation

Microsoft SharePoint Online connector: indexing sites as documents

The new version of the SharePoint Online connector now also supports the indexing of site objects. This means that in addition to all other documents on the site, the site metadata itself is now available.

Link to the documentation

Atlassian Confluence connector: support for labels

With the new Atlassian Confluence connector, labels for Spaces, Pages, as well as Attachments can be indexed as metadata.

Link to the documentation

Technical extensions

Improvements Personalized relevance

In larger installations, there are often users whose individual data should not be used to create personalized relevance models (for example, test users). This release makes it possible to explicitly exclude such users from the relevance calculation. In this way, personalized relevance better reflects actual user behavior.

Link to the documentation

Improving relevance in boosting combinations

By default, if multiple boosting rules apply, the sum of the possible boosting factors is now calculated instead of using the maximum value. Also, dampening rules when the boost factor is less than 1.0 are taken into account. The result is an improved relevance of hits due to more differentiated application of multiple rules.

Link to the documentation

Optimizations when exporting search results

In the "Export search results" function, the handling of special characters in actions such as "Open" has been optimized. Among other things, this could be implemented through changes in the Microsoft File Category descriptor. If you have made manual adjustments in the Microsoft File Connector, we recommend that you make sure that the new changes in the Microsoft File Category descriptor are active after the update. Otherwise, we recommend that you undo your manual adjustments.

Link to the documentation

Important notes on the update

If all of the following points apply to your environment:

• Distributed Operation (Producer/Consumer)
• Update from a version older than 21.2 HF2 to version 21.2 HF2 or newer.
• In the case of sequential updates of the nodes (instead of simultaneous updates), older and newer versions are operated simultaneously for a short time.

please update to version 21.3 or newer instead and follow these instructions.

Security relevant changes

21.2.7.1157

• Removed log4J2 from product (CVE-2021-45105)

21.2.6.1151

• Log4J Update in QueryPerformanceTester plugin (CVE-2021-45046)

21.2.5.1145

• Log4J update in QueryPerformanceTester plugin (CVE-2021-44228)
• Set log4j2.formatMsgNoLookups=true (CVE-2021-44228)

21.2.3.1112

• CentOS Updates: kernel, CPU microcode (CVE-2021-3715, CVE-2020-0543, CVE-2020-0548, CVE-2020-0549, CVE-2020-8695, CVE-2020-8696, CVE-2020-8698, CVE-2020-24511, CVE-2020-24512)

21.2.1.1027

• Update Apache Tika: 1.26 (CVE-2021-27807, CVE-2021-27906, CVE-2021-28657, CVE-2021-22696)
• Update commons-compress: 1.21 (CVE-2021-35515, CVE-2021-35516, CVE-2021-35517, CVE-2021-36090)
• Update Apache Batik: 1.14, xmlgraphics-commons: 2.6 (CVE-2020-11987,CVE-2020-11988)
• Update Apache Tomcat: 8.5.65 (CVE-2020-9484, CVE-2021-25122, CVE-2021-25329)
• Kernel Update 3.10.0-1160.36.2 (CVE-2021-33909, Sequoia, CVE-2019-20934, CVE-2021-3347, CVE-2020-8648, CVE-2020-27170)
• Update nginx (CVE-2021-23017)
• CentOS 7 Updates: screen (CVE-2021-26937), bind (CVE-2021-25215), kernel (CVE-2021-27363,CVE-2021-27364,CVE-2021-27365), nettle (CVE-2021-20305)
• Update Keycloak + Keycloak HA: 14.0.0 (CVE-2020-10770, CVE-2020-14302, CVE-2020-14359, CVE-2020-1725, CVE-2020-27838, CVE-2021-20202, CVE-2021-20222)

Additional Changes

21.2.6.1151

• Fix for: Index Crash with ACL Precomputation Cache and Empty Index

21.2.4.1137

• Increased resource allocation for the Keycloak and Keycloak-HA services
• Cache for Precomputed ACLs
• Fix for: Index may crash when querying for computed properties with invalid reverse References
• Fix for: Under heavy index/update load an index shutdown may mark a bucket as final prematurely

21.2.3.1112

• Coyo SSO (ADFS, OpenIDC) Workplace Integration
• Fix for: TermHitAccessorCache: Incorrect cache behaviour with regards to bitsets
• Advanced Search: Improvements when grouping or editing query info
• Fix for: Summarized by: Preview Action always opens main result
• PSTFilter: reduced memory usage
• Fix for: PSTFilter: Folders with more than 2^16 Items are crawled correctly
• Keycloak HA: Implementation with jdbc_ping instead of multicast
• Fix for: Incorrect instrumentation of the SubQuery BitSetEvaluationStrategy
• Extended Query Expression Label Transformer: Support for value transformation of terms and unparsed queries
• OpenJDK Update to 1.8.0_302
• Fix for: NOT restriction in advanced search does not work with SubQuery expressions
• Fix for: CORS requests are not working inside additional contexts
• Fix for: client.js Chrome cannot read the localStorage property when 3rd-Party cookies are disabled
• Fix for: Persisted Resources: Deadlock when updating the schema in a multi-node setup
• Fix for: Manager UI: Empty custom properties may prevent the start of services

21.2.2.1053

• SubQuery BitSetEvaluationStrategy is configurable to HITINFO_APPENDER if PrecomputeACL is active
• Yammer ACL references can be disabled
• Fix for: POI Filter sets wrong categoryclass
• Number of query threads and precompute threads is changeable at runtime

21.2.1.1027

• Fix for: Microsoft Teams: Incorrect yam_full_name in category descriptor
• Microsoft Teams: No indexing of system event messages
• Responsive layout in MMC for Tasks/Nodes/Snapshots
• ServiceNow: Caching Principal Resolution Service repeatedly failed API requests
• ServiceNow: HR Criteria without conditions are effective for all
• ServiceNow: Meaningful default value for config option: KB Article Size Optimization
• Teams - Delta Indexing and Principal Cache
• Teams - Access to Public Teams Channels
• Fix for: Teams Principal Cache: Everyone is not included in the principals
• Fix for: Yammer: Delta Crawl defect after API change
• Fix for: Filemanager is displayed for Mindbreeze Webmin Administrator
• Opentext Documentum: deletion of documents with existence query
• SharePoint Online: Pages are indexed as documents
• Fix for: SharePoint Online: Requests are retried for too long on HTTP 429.
• Fix for: SharePoint Online: SSL exceptions with high memory consumption
• Fix for: Insight App Designer: display problem when only the designer is used
• Fix for: Update Form: "Start" button is displayed during the whole update process
• Confluence: alternative REST URL is configurable
• Confluence: Support for Confluence labels in the REST API
• Personalized Relevance Transformer: Extension of the ListEvents API
• Personalized Relevance Transformer: Configured users can be excluded from relevance calculation
• Fix for: Web Connector: description of option "Incomplete Delta Crawl Runs" is misleading
• Fix for: Web Connector: JavaScript cannot parse HEAD request "content-type" header
• Fix for: Microsoft CRM Connector: OAuth 2 token responses != 200 lead to `NullPointerException`.
• Fix for: Microsoft File Connector: STATUS_ACCESS_DENIED (LsarOpenPolicy2) when indexing NetApp shares
• Fix for: Microsoft File Connector: Directory Level ACLs are in very rare cases not taken over by documents and indexed without ACLs
• Microsoft File Connector: SMB Multi-Protocol Negotiate can be disabled
• Fix for: Principal Cache: Retry on Incomplete does not work with Pusher implementations
• Fix for: LDAP Principals Resolution Cache: Failed to update cache when using Kerberos and an interval of 1 hour
• QueryParser draws search form (Queryinfo component)
• Hierarchical CSV Enricher provides links as html a tags configurable
• Insight App Designer loads too long when using only hostname instead of fqdn
• Fix for: Source filter restrictions are not checked exactly
• Fix for: aclStatistics index crash
• Fix for: MesUtil::bitwiseOrAtOffset bitset is increased for each segment
• Fix for: Mesmaster on Windows - ConfluenceAccess plugin upload failed
• Fix for: Index remains read-only after backup
• Fix for: InvertedReferenceIndex can hang on Syncdelta Announcements
• Fix for: Instrumentation export preview on client wrong: 0 results although client service > 0 results
• Fix for: Users and Alerts are displayed in the menu even if "Resource Persistence Settings" are disabled
• Fix for: trySetDeletedDocIDs can cause timeouts
• Persisted Resources can be used with encrypted owner information
• Fix for: `handleShortcuts` perordical actions: feedbacks (via app.telemetry) sometimes do not contain the relevant search request
• Fix for: Filtered search sorted by size does not return hits (size not calculated for filtered away items)
• Additives DocumentBoosting is active by default
• Renewal of codesigning certificate
• Fix for: Microsoft File Connector: deletion does not work if "too few" documents are indexed
• app.telemetry Predefined LogPool attributes (e.g. Retention Period) customizable via configuration in image
• Search and sort functions for handsontable
• Fix for: Excel export: date values with wrong time zone offset
• Fix for: Excel Export: wrong encoding of Open action
• Fix for: No results when folders have no size and are sorted by size
• Fix for: Mindbreeze Configuration: when exporting/importing settings from a data source, the names of used parameters are saved as text values
• Term2DocumentBoost: support of user context properties for variables
• Fix for: mesmaster sometimes does not start after product update
• Fix for: POIPlugin uses wrong encoding for RTF content of MSG files
• Charts in Designer: custom source name can be specified directly in source name
• Update app.telemetry / libtelemetry to version 21
• Increased swap area on appliances
• Embedded Mindbreeze client in Sharepoint Online uses existing session for login
• FPDF filter: "document:pages" number of pages as numeric value
• Extended configurability of inverter: max ngram length, limit document inversion after max term count reached
• Microsoft File Connector: excluded documenters are logged and instrumented
• When putting an index into read-only state, final buckets are verified to be completely inverted
• Crawler state contains the DocID
• Fix for: Missing XML tag in Web Connector documentation