Dash-Web.git/src/client/views/nodes/chatbot/vectorstore, branch masterA copy of AVD's Dash-Web software, hosted on my own hardware for local use.
https://git.mfoi.dev/Dash-Web.git/atom?h=master2025-07-07T18:39:06Zchanges (seeing if they work)2025-07-07T18:39:06ZA.J. ShulmanShulman.aj@gmail.com2025-07-07T18:39:06Zurn:sha1:9092494778abd55b6aa299fe06b4f70e7c7a767ffeat: add codebase exploration tools for agent assistance2025-05-21T17:11:52ZA.J. ShulmanShulman.aj@gmail.com2025-05-21T17:11:52Zurn:sha1:c3dba47bcda10bbcd72010c177afa8fd301e87e1
Add three new agent tools to improve navigation and understanding of the codebase:
FileContentTool: retrieves complete content of specified files (max 3)
FileNamesTool: lists all available files in the codebase
CodebaseSummarySearchTool: performs semantic search across file summaries
Added codebase search and retrieval to Vectorstore2025-05-21T16:38:55ZA.J. ShulmanShulman.aj@gmail.com2025-05-21T16:38:55Zurn:sha1:0e98320d3b237f1927b9f1367494dccd7f66eda9
Summary indexing: Added functionality to embed and index file summaries from file_summaries.json in Pinecone
Vector search: Implemented semantic search to find the top 5 most relevant files for a query
Content retrieval: Added method to fetch full file content from file_content.json
API endpoints:
/getFileSummaries - Retrieves all file summaries
/getFileContent - Fetches file content by path
/getRawFileContent - Returns content as plain text to avoid JSON parsing errors
Error handling: Added comprehensive error handling and debugging throughout
Initialization: Implemented proper async initialization sequence with verification
Performance: Added streaming for large files to improve memory efficiency
Testing: Added automated test queries to validate functionality
workign better2025-05-12T19:46:30ZA.J. ShulmanShulman.aj@gmail.com2025-05-12T19:46:30Zurn:sha1:b3aa238043d01cbc58293b45867706fa9b36cefechanegd everything to be more consistent2025-05-11T21:18:18ZA.J. ShulmanShulman.aj@gmail.com2025-05-11T21:18:18Zurn:sha1:e141307dbd9b951f76c908610e7b89e296ad92b8
- made both web related tools use doc manager and chunk Ids
Made it so chunk Ids are seperately managed and made sure the doc id is sonsistent and not created in python spawn2025-05-11T17:42:00ZA.J. ShulmanShulman.aj@gmail.com2025-05-11T17:42:00Zurn:sha1:a5d7f5c38192b91b7df3bd6ecace5ba7365449a6Merge branch 'master' into agent-web-working2025-05-11T14:46:15ZA.J. ShulmanShulman.aj@gmail.com2025-05-11T14:46:15Zurn:sha1:b87b2105e966928518c96c7702b68c12344ffdd7improved consistency across doc types and parsing2025-04-27T18:57:39ZA.J. ShulmanShulman.aj@gmail.com2025-04-27T18:57:39Zurn:sha1:393b7f8286422c933102449eba1ba82874a48896a bit more consistent2025-04-27T17:37:37ZA.J. ShulmanShulman.aj@gmail.com2025-04-27T17:37:37Zurn:sha1:67a7996278ce176e227393fa410e7afc80228a83new attempt with new citation unification2025-04-27T17:14:49ZA.J. ShulmanShulman.aj@gmail.com2025-04-27T17:14:49Zurn:sha1:3ef3d40506348d9fd537cc8f4aea975b9770689f