v0.0.4

What's Changed

• change VERSION filepath at docs/conf.py by @vkehfdl1 in #122
• Rough structure for documentation by @vkehfdl1 in #125
• enable target_module_params at hybrid retrieval in run.py by @vkehfdl1 in #126
• add sample data eli5, msmarco, triviaqa by @bwook00 in #111
• [HotFix] do not join inner at hybrid_cc for consistent 'top_k' result. It can … by @vkehfdl1 in #129
• Extract strategy for using test data evaluation by @vkehfdl1 in #135
• change file name trivia qa_train by @bwook00 in #134
• dump version 0.0.4 by @vkehfdl1 in #136
• Getting Started Documentation by @vkehfdl1 in #137

Full Changelog: v0.0.3...v0.0.4

v0.0.3

What's Changed

• dump version 0.0.2 by @vkehfdl1 in #110
• Fix VERSION file and change logging level to INFO by @vkehfdl1 in #115
• add process_batch() function and add it to all llm related by @vkehfdl1 in #116
• Enable async operation at hyde and query decompose by @vkehfdl1 in #118
• dump version 0.0.3 by @vkehfdl1 in #119

Full Changelog: v0.0.2...v0.0.3

Initial Release

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

• @vkehfdl1 made their first contribution in #3
• @bwook00 made their first contribution in #8

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