For a list of breaking changes, check here.
- Upgrade Lucene to 9.8.0 (#220)
- Upgrade Lucene to 9.7.0
- Support for the
Word2VecSynonymFilter - Update
com.github.clj-easy/graal-build-timeto 1.0.5 - Load Analyzers via SPI
- Upgrade GraalVM to 23.0.0 GraalVM distribution Java 17
- Raudikko 0.1.4
- Upgrade Lucene to 9.6.0
- Upgrade GraalVM to 22.3.1
- Support building native images with GraalVM based on Java 19 on macOS
- Upgrade to Lucene 9.5.0
- Upgrade GraalVM to 22.3.0 (still Java 17)
- Stack traces are not printed in tests
- Upgrade to Lucene 9.4.0
- Upgrade GraalVM to 22.2.0
- Option for Lucene Monitor queries index disk storage under option
:queries-index-dir - Update
lucene-text-analysisto 1.0.17 - Handle
NO_COLORenvironment variable - Support for multiple
--analyzers-file
- Remove usage of the
clojure.core.reducersto simplify code.
- Upgrade to Lucene 9.2.0
- Lucene custom analyzer as an external library
- Lucene custom text analysis as an external library
- Lucene query parsing as an external library
- Deploying to the DockerHub
--presearcherflag to specify Lucene Monitor Presearcher- more efficient
MonitorQuerySerializer
- Native build for the Apple Silicon
- Upgrade to Lucene 9.1.0
- Graalvm Native image "quick build mode" for dev
- Update to GraalVM 22.1.0
- Text analysis as a separate module
- Eager processing of Lucene Monitor matches which makes ~20% throughput improvement.
- Change the default presearcher to NO_FILTERING, less RAM
- Streamed matching with a flag
--streamed
- Fixed flushing behaviour when consuming from STDIN in the ruby-percolator example.
LMGREP_FEATURE_EPSILON_GC: use Epsilon GC in the binary.
- Fixed the Windows release file name construction in Github Actions
- Remove clojure.tools.logging
- ~3x throughput improvements for matching!
--only-analyzewith--no-preserve-orderprevent OOM--only-analyzeoption--queue-sizeto specify the Java executor service queue size- Flag
--show-analysis-componentsto print out a list of available text analysis components --only-analyzewith--preserve-orderbased on Java ExecutorService
- Modular architecture
- Build uberjar with Clojure tools.build
- Release with the official GraalVM Github Actions
- Update to Java 17
- Update to GraalVM 22
- Update Lucene to 9.0.0
- Workaround for the Linux static image compilation
- Flag
--query-update-buffer-sizeto control how many queries are indexed at once
- Global query parser settings are applied for queries in a file
- Raudikko token filter for Finnish Language as an option
- Flag
--config-dirto specify directory where text analysis resources are stored. synonymtoken filter- Flag
--analyzers-filefor custom analyzers definitions.
- Refactored custom TokenFilterFactories to act like services
- Flag to specify the input matching concurrency
--concurrency - Flag
--[no-]preserve-orderthat attempts to increase throughput while not preserving the order, applicable to--only-analyze - Flag
--reader-buffer-sizein bytes - Flag
--writer-buffer-sizein bytes - Flag
--graphto modify--only-analyzethat output a dot program for visualization of a token stream - Flag to limit the depth of file system traversal
--max-depth - Flag to skip hidden files
--[no-]hidden - Add support for the
org.apache.lucene.analysis.miscellaneous.ConcatenateGraphFilterFactorytoken filter - Add support for the
org.apache.lucene.analysis.en.LovinsSnowballStemTokenFilterFactorytoken filter - Add support for the
org.apache.lucene.analysis.miscellaneous.DropIfFlaggedFilterFactorytoken filter - Add support for the
org.apache.lucene.analysis.pattern.PatternTypingFiltertoken filter - Support configuration of various query parsers with a flag
--query-parser-conf
- Updated to Lucene 8.10.1
- Updated to GraalVM 21.3.0
- Updated to quarkus-lucene 0.3
- Faster file system search with GLOB
- Support for multiple Lucene query parsers with flag
--query-parser
- MacOS and Linux installation via
brew - Base static linux build on musl. The binary should now work in most linuxes out of the box.
- Upgrade to graalvm-ce:21.1.0
- Query parsing errors propagates exceptions
--explainflag for the--only-analyzeto return details of tokenssynonymgraphtoken filter now can load synonyms from fileshyphenationcompoundwordtoken filter now can load data from files- Windows installer based on scoop
- Handle autocompleted files in Windows
- Hyperlinking on STDIN throws a NPE #92
- Snowball token filter classes are included into native binary
- Exceptions in clojure.core.async pipeline are handled properly
- Tests for native-image
- Expose a declarative text analysis pipeline specification in JSON (see #81)
- Provide a list of predefined analyzers
- Support customizable char Filters
- Support customizable tokenizers
- Support customizable token filters
- Use Java PrintWriter to write to STDOUT (see #79)
- Flag to only analyze text and spit to stdout
--only-analyze - Lucene Monitor is executed in another thread
- Optimized loading of large query dictionaries
- Scored Highlights with a flag
--with-scored-highlights - Flag to force printing an empty line on no match
--with-empty-lines - Flag to output matching details,
--with-details
- Handle the zero queries input
- Allows specifying multiple queries with the
--queryflags. --queries-fileflag to provide a JSON file with Lucene queries.
- Use the
io.quarkiverse.lucene/quarkus-luceneextension instead of building from source - Exit status should be 0 when -h is specified
- Wait for the input from STDIN
- Display file path as hyperlinks with the
--hyperlinkflag
- Errors when matching directory
- Accepts multiple files (#30)
- A proper static binary builder Docker container
- WordDelimiterGraphFilter added as an optional filter
- Stable Clojure release 1.10.2
- Remove redundant options from native-image building script
- Flag to search in whole files
- Compile Linux executable with
--staticflag - Dockerfile to build static Linux executables
- Option to skip binary files on Linux and MacOS
- Updated the native-image docker to 21.0.0
- Support for output formats JSON and EDN
- Support for output template when format is string
- Support for custom highlighting tags
- Validation of the CLI options
- Building an uberjar
- Scoring (disables highlighting and tags)
- Exclude files with GLOB
- Support for Lucene query syntax
- chore: version number of previous release
- Fixed GLOB patterns so that the full file path should match
- Removed
--slopand--in-orderCLI params as no longer needed.
- Initial release
- Quarkus Lucene extension
- Matching every line of provided text
- text analysis
- Coloring output
- GLOB support
- STDIN support
- CLI params for text analysis
- queries file entries no longer support these keys:
[:case-sensitive? :ascii-fold? :stem? :tokenizer :stemmer :word-delimiter-graph-filter]
- removed
synonymtoken filter, because deprecated and required a patch.
- Removed token filters "lovinssnowballstem" "concatenategraph" because of compilation issues
- Removed
--slopand--in-orderCLI params as no longer needed.