Skip to content

Initial Release

Compare
Choose a tag to compare
@vkehfdl1 vkehfdl1 released this 03 Feb 21:19

This is our first MVP release. Super alpha version, so possible bugs, errors, and so on. Be careful!

What's Changed

  • Make readme for this repo by @vkehfdl1 in #3
  • Make brief structure for RAGround by @vkehfdl1 in #4
  • Create dependabot.yml by @vkehfdl1 in #5
  • Update README.md by @bwook00 in #8
  • Initial module, node, evaluation structure by @vkehfdl1 in #9
  • �Advanced RAG linear automation framework structure by @vkehfdl1 in #22
  • Make sphinx docs by @vkehfdl1 in #24
  • Make github action for sphinx docs by @vkehfdl1 in #26
  • add autorag to requirements.txt by @vkehfdl1 in #28
  • try to upgrade pip by @vkehfdl1 in #29
  • Upgrade sphinx action by @vkehfdl1 in #30
  • Hot fix/#27 by @vkehfdl1 in #31
  • make new test.yml for unit testing by @vkehfdl1 in #35
  • add summary feature by @vkehfdl1 in #37
  • Add three generation metrics, BLEU, ROUGE, METEOR by @vkehfdl1 in #40
  • Add generation evaluation decorator by @vkehfdl1 in #41
  • add global logger at init.py with RichHandler by @vkehfdl1 in #45
  • Change summary csv file to parquet file by @vkehfdl1 in #47
  • Add cli command execution for Evaluate initialize and start_trial by @vkehfdl1 in #50
  • Create vectordb retrieval by @bwook00 in #49
  • Add deploy module and Runner for running optimal pipeline directly from trial folder or yaml file. by @vkehfdl1 in #54
  • fix crucial bug that can't make combinations of multiple modules by @vkehfdl1 in #57
  • Add run api server at Runner by @vkehfdl1 in #59
  • Implement vectordb ingest by @bwook00 in #58
  • Add API cli feature by @vkehfdl1 in #61
  • Implement HyDE module by @bwook00 in #20
  • Implement Query Decompose module by @bwook00 in #18
  • Add Generator module 'llama_index_llm' and its decorator and run function. by @vkehfdl1 in #64
  • Add fstring module and make make_combinations and explode util functions. by @vkehfdl1 in #66
  • Add Prompt Maker run function. by @vkehfdl1 in #68
  • resolve error at making summary_df with filtered results by @vkehfdl1 in #72
  • Add Query_expansion node and its decorator and run function. by @bwook00 in #67
  • Add evaluation retrieval contents metric and its evaluation functions by @vkehfdl1 in #74
  • Add Passage Compressor Node by @vkehfdl1 in #78
  • Implement monot5 reranker module by @bwook00 in #17
  • change 'datetime' metadata at corpus data to 'last_modified_datetime' by @vkehfdl1 in #81
  • hotfix error at filter_by_threshold and avoid_empty_result by @vkehfdl1 in #87
  • Create passage_reranker decorator and run.py by @bwook00 in #88
  • Implement TART reranker module by @bwook00 in #19
  • Implement UPR module by @bwook00 in #89
  • Change summary.parquet file to summary.csv file, and use load_summary_file when you want to load it by @vkehfdl1 in #94
  • change module result filename to numbers by @vkehfdl1 in #96
  • create full trail test and full.yaml file by @bwook00 in #95
  • Make Hybrid Retrieval RRF module and edit run.py for it by @vkehfdl1 in #98
  • Make Hybrid CC module by @vkehfdl1 in #99
  • add embedding_models and generator_models by @bwook00 in #105
  • Fix errors at Runner for deploy feature by @vkehfdl1 in #103
  • Enable using hybrid retrieval at deploy. by @vkehfdl1 in #107
  • Edit readme for launching. by @vkehfdl1 in #109

New Contributors

Full Changelog: https://github.com/Marker-Inc-Korea/AutoRAG/commits/v0.0.2