Vind - Information Discovery for Java

Changelog

This document shows the changes that are introduced into Vind with the releases.

Pre 1.0.0

For information about older releases check the history on GitHub.

1.0.0

• Full package renaming from its original Searchlib structure.
• Included suggestion-handler module and configured Vind Solr backend to use it by default.

1.0.1

• Bugfix: Suggestion handler and Solr backend runtimelib configuration issue in tests fixed.

1.0.2

• Collection manager improvements: Improved logging.
• Reduced logging level to avoid not helpful noise in the logs.
• Reduced solr calls in Vind updates of documents with nested documents.
• Added the possibility of faceting on nested documents fields.
• BugFix: Error in solr schema.xml.multi stored filter field was not defined as stored.
• BugFix: solved issue with long queries by using POST method in Solr requests instead of GET.

1.0.3

• BugFix: nested document facet queries updated to use edismax and added missing parenthesis to nested document search query.
• BugFix: non escaped characters in nested document Json facet query.
• BugFix: changed return type of searchServer getById method for annotated java pojos to BeanGetResult.
• BugFix: avoid document duplication in index (Solr issue with coexisting documents and documents with nested docs).

1.1.0

• Full renaming from reporting to monitoring module.
• First stable version of monitoring model.
• Added monitoring log writer.
• Added monitoring ElasticSearch writer.
• Added first monitoring analyzer: report, able to write Json or HTML reports.
• Added to SearchServer a getRawQuery method to get the backend specific query.
• Added the possibility ti configure by config properties the backend connection timeout and read timeout.
• BugFix: Added real deep copy of fulltext search object.
• BugFix: Added missing search facet limit to term facet Json implementation.

1.1.1

• BugFix: added mixIn to properly serialize DateMathExpression objects and its sub objects.
• API change: deprecated method getName() from Facet API. Use getFacetName() instead when looking for the Facet name or getFieldName().
• Simplified InteractionEntry data model.
• Added monitoring to Index, Real time Get, Update and Delete actions from monitoringServer.
• Added silent option to MonitoringServer allowing to ignore unnecessary exceptions thrown during monitoring, which do not affect search execution. By default the MonitoringServer silent is set to false, and this can be changed with the method MonitoringServer.setSilent(boolean).

1.1.2

• BugFix: Reduced logging level from error to debug of Filter.getScope(FieldDescriptor descriptor) when falling back to default scope due to a null FieldDescriptor.
• Added helpful debug logging in MonitoringServer.

1.1.3

• BugFix: CollectionManagementService now closes the SolrCloudClient connection
• BugFix: Solved issue with javadoc version
• BugFix: Homogenize monitoring property types to return always a number(some interval facets where giving a start date or a number depending on the original field descriptor type).

1.1.4

• SuggestionHandler code cleaning and improvements in logging and Exception handling.
• Bugfix: Fixed issue when adding and filters.

1.2.0

• General improvements on reporting analyzer module and monitoring.
• Updated suggestion handler result to NamedList in order to be more homogeneous with Solr result types.
• Added support to Vind configuration via environmental variables.
• Bugfix: Added to NOT filter Solr serialization a base ‘*:*’ to ensure the subtraction has a positive base operator.
• Bugfix: solved issue with wrong parsing of children doc filters for suggestions on fields members of both, parent and nested doc.
• Bugfix: Removed hard coded logical operator on suggestion handler, allowing to override it by parameter or configuration.

1.2.1

• Changed configuration properties preference order (from most relevant to less): Environment properties > provided config file > default config file.
• BugFix: suggestion handler maps every different type of single quote to the same (the same goes for double quotes and other special characters).
• BugFix: Solved issue with connection timeout when creating reports.

1.2.2

• Bugfix #46: fixed issue in children filter serialization normalizing the filters to logical DNF and grouping them afterwards by hierarchical scope.
• Improvement #51: removed ElasticSearch Jest client dependency from monitoring-api module by extracting it to a new monitoring-utils module.
• Improvement #50: added helper methods to create before, after and between filters from java util Date.
• Feature #45: added a new filter TermsQueryFilter which enables the possibility to do huge comma separated term filters by the means of the TermsQuery parser in Solr.
• Bugfix #43: fixed preprocessing from analysis report to add final flag to fulltext queries which the last entry of a session and have no following closing action (like selection of a document or restarting of the search).
• Feature #53: Added field analysis handler.

1.2.3

• Updated Solr dependencies to 5.5.5.
• Improvement #60: Added health check functionality.
• Feature #57: Added the possibility to define more than one and or or children searches.
• Bugfix #62: Fixed issue which enforced collections created with Vind release <= 1.2.0 to be completely deleted and re-indexed in order to don’t throw an exception on suggestions.

1.2.4

• Fix #64: removed suggestion string field copy rule to suggestion analyzed field from schema.

1.2.5

• Fix #66: Fixed normalization of AND filters containing just OR children filters.

1.3.0

• Improvement #68: Documented and applied new release policy.
• Changed suggestion handler internal operator from AND to OR as default.

2.0.0

• Upgrade to Solr7.5.
• Ensure that number of the page is 1 or bigger.
• Feature #73: Support index Within.
• Improvement #67: Provide Docker-Image for Solr Backend.

2.1.0

• Feature #87:Added the possibility to set autoAddReplicas to true for the collection manager tool
• Bugfix: avoid null pointer exception whe updating a non existing doc.
• Bugfix: #86: moved missing suggestion handler fix from Vind 1.x.x to 2.x.x

2.1.1

• Hotfix: fixed issue in suggestion handler when q.op was set to OR.

2.1.2

• Hotfix: Fixed issue introduced by solr inconsistent response types.

2.1.3

• Hotfix: Releated to previous fix, remove cast to integer by NumberUtils toInt method.

2.1.4

• Hotfix: Reduced log level from info to debug on document update to avoid spaming the logs.

2.1.5

• Hotfix: Ping fails against non-cloud solr-backend (#90)

2.2.0

• Improvement: Simplified suggestion regex query.
• Improvement: Added word boundary as prefix for words in suggestion handler instead of blank space.
• Improvement: Instead of reading local solr schema as file to do schema check, now reading it as input stream.
• Improvement: Removed invalid characters from file name and path on collection download

2.2.1

• Fix: alignment of language specific fields tokenizer with generic text definition

2.3.0

• Improvement: Delete by query no longer queries for the ids of the documents as it was not performing for long sets of documents matching the delete filters.

2.3.1

• Fix: Issue on nested doc deletion when deleting by query.

2.4.0

• Improvement: Builders for StatsFacets now have for each foo()-method also a foo(boolean fooEnabled) method.

2.5.0

• Fix: Possible IllegalArgumentException in DocumentFactory (#106)
• Fix: Added base Match all clause to NOT filter serialization (#110)

3.0.0

• Feature: Implemented Elasticsearch backend (7.6.1)
• Feature: Smart parse of fulltext search input
• Feature: Reverse search (Only Elasticsearch)
• Feature: Facet sorting
• Improvement: Support basic and ApiKey backend authorization
• Feature: MasterSlaveBackend for parallel indexing

3.0.1

• Improvement: Add support to ‘.’ on field names for elasticsearch backend.
• Improvement: Extended smart parser to support range filters on numeric and date fields.(#158)

3.0.2

• Improvement: Added strict parse mode.(#158)
• Bugfix: Fixed Smart parser issues regarding multi clauses and external parentheses.

3.0.3

• Bugfix: Refactor smart parser grammar to fix issues on boolean operations within filter values.

3.0.4

• Bugfix: Fix quoted literal issues.
• Bugfix: Fixed priority on unary operator grammar.

3.0.5

• Improvement: change default search string on elastic backend from * to *:* in oder to avoid empty results when no fulltext values are indexed.(#163)
• BugFix: Added track_total_hits default to true and configurable in elastic searches to get real number of results over 10000.(#162)
• BugFix: Escape single quotes on painles script literal to avoid errors updating.(#165)

3.0.6

• Improvement: produce significant error messages on search server instantiation.(#168)
• BugFix: Term facets results are properly build now.

3.0.7

• BugFix: Fixed nested filer scopes on boolean operation filter (AND, OR, NOT).

3.0.8

• BugFix: Facet limit <0 is translated to aggregation size Integer.MAX_VALUE on elastic backend.

3.0.9

• BugFix: Fix sorting when sort fields do not (yet) exist (#174)
• BugFix: Do not throw Exception when deleting a non existing document (#176)
• BugFix: Allow brackets in fulltext search for Smart Parser (#177)

3.0.10

• BugFix: Fix indexing of non-stored complex fields (#175)

3.0.11

• BugFix: Fix atomic updates for complex fields (#183)
• Improvement: Add special sorting by score (#184)
• BugFix: Support open parentheses within fulltext smart parser section (#186)

3.0.12

• BugFix: Minimum should match elastic unexpected behaviour. (#190)
• Improvement: Ignore ()[]{}’“ characters in smart parser fulltext search section. (#191)
• BugFix: Atomic update and complex fields name issue. (#189)
• Improvement: Improved update scripts to use parameters for values and fields.

3.0.13

• BugFix: Update script issue with dates.

3.0.14

• BugFix: Complex field field name creation issue.

3.0.15

• BugFix: Avoid null pointer exception on populaltion of complex fields.

3.0.16

• BugFix: problem on foorprint creation,missing complex field

3.0.17

• BugFix: Fixed issue on complex field footprint creation.
• Improvement: some refactoring

3.0.18

• Improvement: Increase version conflict retries for elastic backend and add a config parameter.
• BugFix: add a null check to avoid null pointer except on complex fields footprint

3.0.19

• BugFix: fixed boolean condition in foot print creation

3.0.20

• BugFix: Avoid null pointer exception when not storage type defined for complex fields

3.0.21

• BugFix: Avoid type check validation for multi field when value is null.

3.0.22

• BugFix: Avoid null pointer exception when there is no suggestion results on spellchecked queries.

3.0.23

• BugFix: Added check for non collection values when building the default sorting values.

3.0.24

• BugFix: fix binary field footprint.

3.0.25

• Improvement: Applied max retries config value to delete by query.

3.0.26

• Improvement: Added search configuration to delete on version conflict or not.

3.0.27

• Improvement: Implemented support fot [* TO *] range.
• bugfix: Made field descriptor aware of context to avoid unknown field error in elastic.

3.0.28

• Bugfix: Escaped elastic special characters from search when the syntax is not right.

3.0.29

• Bugfix: fulltext search elastic input treated to avoid errors with whitespaces.
• Bugfix: Avoid NPE when doing contextualize search on fields which do not have the context.

3.0.30

• BugFix: roll back fulltext search text treatement.
• BugFix: between filter uses datemath to elastic string.
• BugFix: get FieldName now returns Optional to avoid NPE and elastic errors with null fields.

3.0.31

• BugFix: Fix context field issues.
• Improvement: Added better elasticsearch footprint handling based on mapping fields.
• BugFix: fixed issue with spellcheck results.

3.0.32

• Feature: Implemented cursor search for elasticsearch backend.

3.1.0

• Improvement: Added cursor reference to individual document for cursored based search(#199).

3.1.1

• Bugfix: avoid index out of bounds on cursor results.

3.1.2

• Bugfix: Fixed issue on query pareser with ‘.’ caracter at the end of fulltext search

#3.1.3

• Bugfix: Use Scope define at facet level for aggregations.
• Bugfix: Implement stats facets on text fields and enable cardinality facet.
• Feature: Added pagination to pivot facet.

#3.1.4

• Bugfix: Update check on term aggregation usecase from fixed facet to provided usecase.
• Improvement: Added size to the paging on pivot facets.

#3.1.5

• Bugfix: cardinality calculation on pivot paging is fixed to use original query filters and text.

#3.1.6

• Improvement: Support sorting by id.

#3.1.7

• Bugfix: escape colon when the previous word is not a field or wildcard.

#3.1.8

• Bugfix: Added Unicode hex 0060 and 00B4 mapping to simple quote ‘.
• Bugfix: Run validate query just against the field ‘fulltext’ to avoid hitting the field expansion limit per query.

#3.1.9

• Bugfix: Fixed issue with special regex lucene character on suggestion input

#3.1.10

• Feature: disable lucene syntax flag (#201)
• Feature: Sorted Suggestions (#205)
• Improvement: Suggestions can be limited by a full text search (#204)