Hugging Face's logo

Spaces:
DataEyond
/
Agentic-Service-Data-Eyond
Paused

App Files Community
16
Agentic-Service-Data-Eyond / src

Commit History

  1. fix sorted ranking so model uses overall sorted retrieved chunks

    Rifqi Hafizuddin
    3e7924d

  2. make executors self-contained, remove redundant pre-filter

    sofhiaazzhr
    73b7fe3

  3. [NOTICKET][doc] correct metadata key path in _format_context

    sofhiaazzhr
    16ab916

  4. [NOTICKET][doc] remove column filter and fallback cap for full-schema approach

    sofhiaazzhr
    959b1b0

  5. [NOTICKET][doc] add sheet-level leg and RRF voting for tabular retrieval

    sofhiaazzhr
    5f86993

  1. fix: minor returned type if sql writes limit yang melebihi batas

    Rifqi Hafizuddin
    b4df8b1

  2. [NOTICKET][db] add sheet-level retrieval and focus LLM schema context to retrieved columns

    sofhiaazzhr
    a205d0c

  3. [NOTICKET] pass orchestrator search_query to sql executor for multi-turn context⁠

    Rifqi Hafizuddin +1: Claude Sonnet 4.6
    23eeb2d

  4. [NOTICKET][doc] validate embedding vector for NaN/Infinity in manhattan retriever

    sofhiaazzhr
    2167a5b

  1. fix fallback to fresh retrieval on corrupted Redis cache

    sofhiaazzhr
    9cb950f

  2. [NOTICKET][doc] add guard if filename None

    sofhiaazzhr
    b7fbaeb

  3. [NOTICKET] now retrieve db tables first, then get column from the obtained tables. reduce k to 5

    Rifqi Hafizuddin
    bb29492

  4. [NOTICKET][doc] fix aggregate count operation when value_col is not specified

    sofhiaazzhr
    00aa61d

  1. fix: 5 bug fixes on tabular executor⁠

    sofhiaazzhr +1: Claude Sonnet 4.6
    a49dc1b

  2. feat: add sheet-level chunk on CSV/XLSX ingestion⁠

    sofhiaazzhr +1: Claude Sonnet 4.6
    8daf9b5

  3. fix: increase K in chat endpoint to 10

    Rifqi Hafizuddin
    b59ef76

  4. fix: query executor now use user question as prompt (sebelumnya pakai hasil orchestrator)

    Rifqi Hafizuddin
    0935ede

  5. [NOTICKET] minor refactoring

    Rifqi Hafizuddin
    e4f62b8

  6. [NOTICKET] rrf merge now at router level

    Rifqi Hafizuddin
    de32ab0

  7. fix: fix dedup logic

    Rifqi Hafizuddin
    c9d3b33

  8. [KM-533] now also retrieves table level chunk

    Rifqi Hafizuddin
    4150ba7

  9. [KM-533] add table level schema, differentiate with chunk level. expand retrieval result with FK exploration

    Rifqi Hafizuddin
    fc1239a

  1. [KM-455][document] decided methods retrieval for document

    sofhiaazzhr
    cf77d20

  2. [KM-515][document] Make Query for Tabular Type (XLSX & CSV)

    sofhiaazzhr
    3604994

  3. [KM-507] now only uses hybrid (cosine and bm25)

    Rifqi Hafizuddin
    40925b4

  4. [NOTICKET] fix query now use orchestrator msg, rework db pipeline replace ingestion logic

    Rifqi Hafizuddin
    be9bbd9

  1. [KM-516][KM-517] add new feature; ai can now see table & column names that have fk relationship with retrieved result

    Rifqi Hafizuddin
    f86da27

  1. [KM-520] Integrate db query executor pipeline with existing rag retrieve pipeline

    Rifqi Hafizuddin
    a25febe

  2. [NOTICKET] fix-revert string change

    Rifqi Hafizuddin
    110ee34

  3. [NOTICKET] db_executor: CTE DML check now walks entire AST root, schema: cast instead of string interpolation

    Rifqi Hafizuddin
    bd2b1d9

  4. [NOTICKET] fix delete, now can filter by user

    Rifqi Hafizuddin
    f273db0

  5. [KM-512] connect query executor to user question. add logging for db_executor

    Rifqi Hafizuddin
    abc494f

  6. [KM-513][document] add convert to parquet if type file is XLSX and CSV

    sofhiaazzhr
    770f26b

  7. [NOTICKET] rename file name, updated after uv sync

    Rifqi Hafizuddin
    948d6dd

  8. [KM-512] add Pydantic model the LLM fills via function calling in sql_query, and add same signature for db and tabular

    Rifqi Hafizuddin
    220f59e

  9. [NOTICKET] minor fix in chat.py, add package for query, change schema used to hybrid (cosine+bm25)

    Rifqi Hafizuddin
    15cd3a7

  1. [KM-512] create folder for querying from bd/tabular docs

    Rifqi Hafizuddin
    2c8a3e8

  2. [KM-507] add different methods, now using dense cosine

    Rifqi Hafizuddin
    145bca3

  3. [NOTICKET] add db_client for querying

    Rifqi Hafizuddin
    e49db60

  4. [KM-507] add changes to methods

    Rifqi Hafizuddin
    8218650

  5. [KM-507] add multiple retrieval method to compare (dense, mmr, bm25, hybrid)

    Rifqi Hafizuddin
    ac6b78d

  1. [NOTICKET] fixed multiple sources

    sofhiaazzhr
    e9f2a26

  2. [NOTICKET] fix single source to multiple sources

    sofhiaazzhr
    589ca32

  1. [KM-438][KM-439] framework for knowledge retriever

    Rifqi Hafizuddin
    d1e1264

  2. [NOTICKET][document]: make a clean output to status error unsupported file type

    sofhiaazzhr
    2814813

  3. [NO TICKET][document]: delete vector embedding on table langchain_pg_embedding if user delete document on knowledge

    sofhiaazzhr
    ac3d8c1

  4. [NO TICKET][document]: add updated_at on metadata

    sofhiaazzhr
    d2f7a48

  5. [NOTICKET] add updated_at field for metadata & delete old embedding before appending

    Rifqi Hafizuddin
    cb5ab32

  6. [NOTICKET] add total token logging

    Rifqi Hafizuddin
    b9703fc

  1. [KM-496] [DED][AI] Fix Response data structure for field sources (db flow)

    ishaq101
    5c4067e