Skip to content

Latest commit

 

History

History
344 lines (220 loc) · 8.8 KB

CHANGELOG.md

File metadata and controls

344 lines (220 loc) · 8.8 KB
Raw

Changelog

For a list of breaking changes, check here.

Unreleased

  • Upgrade Lucene to 9.8.0 (#220)

v2023.08.31

  • Upgrade Lucene to 9.7.0
  • Support for the Word2VecSynonymFilter
  • Update com.github.clj-easy/graal-build-time to 1.0.5
  • Load Analyzers via SPI

v2023.06.18

  • Upgrade GraalVM to 23.0.0 GraalVM distribution Java 17
  • Raudikko 0.1.4

v2023.05.13

Fixed/enhanced

  • Upgrade Lucene to 9.6.0
  • Upgrade GraalVM to 22.3.1
  • Support building native images with GraalVM based on Java 19 on macOS

v2023.01.30

Fixed/enhanced

  • Upgrade to Lucene 9.5.0
  • Upgrade GraalVM to 22.3.0 (still Java 17)
  • Stack traces are not printed in tests

v2022.10.21

Fixed/enhanced

  • 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-analysis to 1.0.17
  • Handle NO_COLOR environment variable
  • Support for multiple --analyzers-file

Changed

  • Remove usage of the clojure.core.reducers to simplify code.

v2022.07.27

Fixed/enhanced

  • Upgrade to Lucene 9.2.0

v2022.05.12

Fixed/enhanced

  • 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
  • --presearcher flag to specify Lucene Monitor Presearcher
  • more efficient MonitorQuerySerializer

v2022.05.01

Fixed/enhanced

  • 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

v2022.02.19

  • Streamed matching with a flag --streamed

v2022.02.18

  • Fixed flushing behaviour when consuming from STDIN in the ruby-percolator example.
  • LMGREP_FEATURE_EPSILON_GC: use Epsilon GC in the binary.

v2022.02.17

Fixed/enhanced

  • Fixed the Windows release file name construction in Github Actions
  • Remove clojure.tools.logging

v2022.02.14

Fixed/enhanced

  • ~3x throughput improvements for matching!
  • --only-analyze with --no-preserve-order prevent OOM
  • --only-analyze option --queue-size to specify the Java executor service queue size
  • Flag --show-analysis-components to print out a list of available text analysis components
  • --only-analyze with --preserve-order based on Java ExecutorService

v2022.02.10

Fixed/enhanced

  • Modular architecture
  • Build uberjar with Clojure tools.build
  • Release with the official GraalVM Github Actions
  • Update to Java 17
  • Update to GraalVM 22

v2021.12.09

Fixed/enhanced

  • Update Lucene to 9.0.0
  • Workaround for the Linux static image compilation

v2021.11.09

New

  • Flag --query-update-buffer-size to control how many queries are indexed at once

Fixed

  • Global query parser settings are applied for queries in a file

v2021.11.08

New

  • Raudikko token filter for Finnish Language as an option
  • Flag --config-dir to specify directory where text analysis resources are stored.
  • synonym token filter
  • Flag --analyzers-file for custom analyzers definitions.

Fixed/enhanced

  • Refactored custom TokenFilterFactories to act like services

v2021.11.01

New

  • Flag to specify the input matching concurrency --concurrency
  • Flag --[no-]preserve-order that attempts to increase throughput while not preserving the order, applicable to --only-analyze
  • Flag --reader-buffer-size in bytes
  • Flag --writer-buffer-size in bytes
  • Flag --graph to modify --only-analyze that 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.ConcatenateGraphFilterFactory token filter
  • Add support for the org.apache.lucene.analysis.en.LovinsSnowballStemTokenFilterFactory token filter
  • Add support for the org.apache.lucene.analysis.miscellaneous.DropIfFlaggedFilterFactory token filter
  • Add support for the org.apache.lucene.analysis.pattern.PatternTypingFilter token filter
  • Support configuration of various query parsers with a flag --query-parser-conf

Fixed/enhanced

  • Updated to Lucene 8.10.1
  • Updated to GraalVM 21.3.0
  • Updated to quarkus-lucene 0.3
  • Faster file system search with GLOB

v2021.05.23

New

  • Support for multiple Lucene query parsers with flag--query-parser

Fixed/enhanced

  • 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

v2021.05.02

Fixed/enhanced

  • --explain flag for the --only-analyze to return details of tokens
  • synonymgraph token filter now can load synonyms from files
  • hyphenationcompoundword token filter now can load data from files
  • Windows installer based on scoop
  • Handle autocompleted files in Windows

Fixed/enhanced

  • Hyperlinking on STDIN throws a NPE #92

v2021.04.26

Fixed/enhanced

  • Snowball token filter classes are included into native binary
  • Exceptions in clojure.core.async pipeline are handled properly
  • Tests for native-image

v2021.04.23

Fixed/enhanced

  • 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)

v2021.04.06

New

  • Flag to only analyze text and spit to stdout --only-analyze
  • Lucene Monitor is executed in another thread
  • Optimized loading of large query dictionaries

v2021.03.24

New

  • 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

v2021.03.18

Fixed/enhanced

  • Handle the zero queries input

v2021.03.17

New

  • Allows specifying multiple queries with the --query flags.
  • --queries-file flag to provide a JSON file with Lucene queries.

Fixed/enhanced

  • Use the io.quarkiverse.lucene/quarkus-lucene extension instead of building from source
  • Exit status should be 0 when -h is specified
  • Wait for the input from STDIN

v2021.02.28

New

  • Display file path as hyperlinks with the --hyperlink flag

Fixed/enhanced

  • Errors when matching directory

v2021.02.06

Changed

  • Accepts multiple files (#30)
  • A proper static binary builder Docker container
  • WordDelimiterGraphFilter added as an optional filter

v2021.02.02

Changed

  • Stable Clojure release 1.10.2
  • Remove redundant options from native-image building script
  • Flag to search in whole files

v2021.01.24

New

  • Compile Linux executable with --static flag
  • Dockerfile to build static Linux executables
  • Option to skip binary files on Linux and MacOS
  • Updated the native-image docker to 21.0.0

v2021.01.15

New

  • 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

v2021.01.13

New

  • Support for Lucene query syntax
  • chore: version number of previous release

Fixed/enhanced

  • Fixed GLOB patterns so that the full file path should match

Changed

  • Removed --slop and --in-order CLI params as no longer needed.

v2021.01.11

New

  • Initial release
  • Quarkus Lucene extension
  • Matching every line of provided text
  • text analysis
  • Coloring output
  • GLOB support
  • STDIN support
  • CLI params for text analysis

Breaking changes

Unreleased

  • queries file entries no longer support these keys: [:case-sensitive? :ascii-fold? :stem? :tokenizer :stemmer :word-delimiter-graph-filter]

v2021.05.02

  • removed synonym token filter, because deprecated and required a patch.

v2021.04.26

  • Removed token filters "lovinssnowballstem" "concatenategraph" because of compilation issues

v2021.01.13

  • Removed --slop and --in-order CLI params as no longer needed.