Skip to content

Latest commit

 

History

History
104 lines (91 loc) · 5.16 KB

release-notes-v0.7.0.md

File metadata and controls

104 lines (91 loc) · 5.16 KB
Raw

Anserini Release Notes (v0.7.0)

Release date: December 13, 2019

New Features

  • Upgraded to Lucene 8.3.0.
  • Added basic support for indexing and search in non-English languages.
  • Added regressions for NTCIR-8 (Chinese), CLEF 2006 (French), TREC 2002 (Arabic), and FIRE 2012 (Bengali, Hindi, English).
  • Added instructions and regressions for docTTTTTquery on MS MARCO Passage Retrieval task.
  • Added initial support for indexing CORE open access research papers.
  • Added similarity that accurately computes document lengths for BM25.
  • Added support for approximate nearest-neighbor search (see arXiv:1910.10208).
  • Added ability to read topics directly from the fatjar.
  • Improved support for Elasticsearch.
  • Improved support for Solr.

Pyserini

  • Removed Pyserini from repo, now standalone project.
  • Refactored SimpleSearcher in coordination with initial Pyserini release.
  • Added IndexReaderUtils, exposing various hooks for Pyserini.

Cleanup, Testing, Documentation

  • Added integration with Codecov.
  • Added more test cases to increase test coverage.
  • Added links to Anserini notebooks.
  • Added end-to-end integration tests for Solr and Elasticsearch.
  • Moved CACM collection into repo for cleaner, more self-contained testing.
  • Refactored logging and counters in indexer, improved documentation of options.
  • Cleaned up warnings during build.
  • Slimmed down fatjar size, removed unnecessary dependencies.

Contributors (This Release)

Sorted by number of commits:

All Contributors

Sorted by number of commits, according to GitHub: