Anserini Release Notes (v0.12.0)

April 29, 2021 · View on GitHub

Release date: April 29, 2021

  • Added/updated/fixed regressions for MS MARCO doc ranking, TREC 2019 DL, and TREC 2020 DL.
  • Added regressions for TREC 2020 background linking.
  • Added support for C4 Corpus.
  • Added ability to index and search pre-tokenized documents.
  • Improved end-to-end test harness.
  • Cleaned up LTR code (renamed features), improved documentation.
  • Implemented getDocumentTokens in IndexReaderUtils
  • Removed code related to knowledge graphs.
  • Refactored TopicReader, improved building of TOPIC_FILE_TO_TYPE mapping.
  • Fixed bug in parsing of multi-line TREC topics (impact on regressions for Disks 1 & 2).
  • Fixed bug where DocumentCollection was not following symlinks properly.

Contributors (This Release)

Sorted by number of commits:

All Contributors

All contributors with more than one commit, sorted by number of commits, according to GitHub: