- history of changes: see https://github.com/delving/hub3/compare/v0.1.11...master
- Support for test-containers for ikuzo service and storage tests [GH-27]
- history of changes: see https://github.com/delving/hub3/compare/v0.1.10...v0.1.11
- Allow for custom []ProxyRoute to be configured when configuring a DataNode proxy [GH-26]
- Delete DataSet does not work when when no v1 index is created [GH-25]
- history of changes: see https://github.com/delving/hub3/compare/v0.1.9...v0.1.10
- Support for running hub3 in DataNode mode [GH-24]
- Support for generic bulk service orgID-aware posthooks
- Implementation of orgID-aware posthook for Ginger platform
- escape html entities in v1 indexing output
- history of changes: see https://github.com/delving/hub3/compare/v0.1.8...v0.1.9
- Tree: make sure fields in Tree API remain sorted by their sortorder key.
- EAD: removed deprecated pre-task EAD processing functions
- EAD: add support for meta.tags in EAD upload form
- License: add license headers to all Go and Protobuf source files
- Search: add backtick '`' to characters to be ignored or removed during tokenization
- Memory: full-text-index phrase search now docID aware
- EAD: when EAD processing tasks stop with an error the worker tasks should not exit
-
EAD: allow treePage calls with search hits
-
Revision: add support for Git RPC calls
-
Search: caching proxy for ElasticSearch
-
Search: add support for index and alias management
-
Logging: custom logger based on rs/zerolog
-
Index: add index service with support for NATS based queued indexing and official ElasticSearch library
-
Server: add TLS support via the configuration
-
Server: dependency injection via new config package based on viper configuration
-
EAD: task-based processing of uploaded EADs
-
Protobuf: update to new golang Protobuf API, see https://blog.golang.org/protobuf-apiv2
-
Bulk: Bulk service migrate to Index Service
-
EAD: structured logging for all task states
-
history of changes: see https://github.com/delving/hub3/compare/v0.1.7...v0.1.8
-
bugfix: add c level xml tag for nested clevels when unmarshalling EADs
-
bugfix: revert back to ElasticSearch QueryStringQuery because unexpected handling of boolean queries by SimpleQueryStringQuery.
-
ead: allow for deeplinking into EAD tree view
-
ead: support next and previous scrollIDs in the ScrollPager
-
history of changes: see https://github.com/delving/hub3/compare/v0.1.6...master
-
Protobuf definition for primary domain model for metadata
-
Tokenized highlighting is now tag aware.
-
EAD: when paging query returns zero redirect to first page.
-
EAD: Add eadid, filedesc, 'archdesc>did' to the Description DataItems.
-
EAD: Added generator for EAD numbered clevel support.
-
EAD: Enable support for numbered clevels again.
-
EAD: CLI option to update EAD description indices.
-
boyscout: small fixes.
-
history of changes: see https://github.com/delving/hub3/compare/v0.1.5...v0.1.6
- Set ResponseSize to 1 to properly calculate the scrollID and next cursor position when searching through the tree.
- history of changes: see https://github.com/delving/hub3/compare/v0.1.4...v0.1.5
- Remove EAD disk-store on introspect reset
- Use more robust elastic.SimpleQueryStringQuery for full-text searches
- Used config.EAD.SearchFields for building EAD full-text queries
- QueryParser bugfix for wrongly assigned phrase queries.
- search.QueryTerm to elastic.Query converter
- history of changes: see https://github.com/delving/hub3/compare/v0.1.3...v0.1.4
-
history of changes: see https://github.com/delving/hub3/compare/v0.1.2...v0.1.3
-
Remove EAD cache directory on dataset delete.
-
New search.Tokenizer that integrates with search.Vector and is tag-aware.
-
search.TokenStream supports vector-based highlighting
-
TextIndex and TextQuery use search.Vectors and search.Matches
-
fragments.Resources NewFields() support vector-based highlighting
-
DescriptionIndex and Description now use 'gob' serialisation format for faster retrieval.
-
DescriptionIndex is persisted to disk on upload
-
all search options that interact with ead.Description use persisted DescriptionIndex
-
moved toplevel 'pkg' content to 'hub3' package
-
New fast search.Autocomplete that takes search.TokenStream as input.
-
New search.SpellCheck for spelling suggestions for historic texts
-
Automatic extraction to triples of EAD "", "", and "". These can be used for search or aggregations.
-
ElasticSearch mapping update. 'html_strip' char-filter is part of the default analyzer.
- minor fixes
- update to elasticsearch mapping fields
- fix for access error on EAD genreform.
- history of changes: see https://github.com/delving/hub3/compare/v0.1.1...v0.1.2
- history of changes: see https://github.com/delving/hub3/compare/v0.1.0...v0.1.1
- refactor hub3 package structure inside the hub3 subpackage.
- hub3 runnable is now inside 'hub3ctl'
- removal of top-level 'middleware' and 'logger' packages in favor of ikuzo packages
- updates to 'ead' package, see EAD section
- integration of ikuzo proof-of-concept new package organisation in 'ikuzo' sub-package
- integration of RAML-based API console. This is all the scaffolding. The API documentation is still in progress.
- implementation of custom Lucene-like query-parser in 'search.QueryParser'
- implementation of memory-based single document Full-Text indexing in 'memory.TextIndex'
- implementation of memory-based incremental highlighter 'memory.TextQuery'. Unicode search and highlighting is fully supported.
- extended search.Analyser to also transform phrases
- added wrappers to 'logger' package to support integration with stdlib logging implementations such as 'olivere/elasticsearch'.
- initial experimental version of search.QueryTerm builder for ElasticSearch queries. This will replace the elastic.SimpleQueryStringQuery in future versions.
- new EAD parser based on the full public set of EADs of the Dutch National Archive. Future parsers can be generated from source EADs.
- implementation of RDF triple generator for EAD clevel and did fields. For search these fields are stored in ElasticSearch in 'tree.rawContent'.
- disabled the support for numbered clevels in the EAD. This functionality can be activated when the generator is completed.
- added support for custom fields in the EAD tree API. Which fields to show can be configured through ead.TreeFields in the TOML configuration.
- tree API handler now supports 'withFields=true' parameters to enable the display of the tree-fields.
- replaced highlighting of EAD description API with memory.TextQuery.
- initial Go Modules complaint version