Spaces:
Running
Running
Commit History
-
fix: direct DB trace persistence with live polling and mobile dropdown full-screen viewport positioning
004bd38
-
feat: Context-Aware Glossary, 7 Chunking Strategies, Dual-Tier Memory System, and Structure-Aware Overhaul
e9da452
-
fix(benchmark): normalize retrieval_method literals and mount models endpoint
edc23a5 -
feat: multi-select query enhancements, conversation-scoped docs, rich markdown tables, and UI overhaul
- Implemented full multi-select query enhancement pipeline (condensation, coreference, rewrite, decomposition, metadata filtering, routing, step-back, keyword expansion, multi-query, HyDE, RAG-Fusion). - Added conversation-scoped document uploading, indexing, and selection with session-id tracking in DB. - Implemented rich GFM Markdown renderer with frosted-glass responsive tables, code copy buttons, and preprocessed block parsing. - Added ChatGPT-style message controls (inline query editing, query/response copy, answer regeneration). - Synchronized LLM models across DocPilot and GaugePilot with fail-safe fallback to GROQ_MODEL. - Refined mode-adaptive theming and dominant PilotMaster branding across auth and workspace pages.fd60d97
-
feat: implement multi-document retrieval and pluggable chunking architecture
d8c1ecb
-
insights and analysis
51d01c7
-
readme update and gaugepilot custom config fix
75100aa
-
multi query enhancement
db792bd
-
tracepilot pipeline config initialization and gaugepilot expsetup frontend fix
69fb503
-
benchmark initial backend and frontend
263c06c
-
tracepilot trace data and reset traces button behavior decoupled between experimental and production modes, added clear all button to clear conversations/chats from docpilot with decoupled behavior
168c81e
-
docpilot chat data decoupled between experimental and production mode, selector behavior fixed
16fd8b4 -
assigning re-ranker its own dropdown
477649c -
AWS deployment using EC2 and dropdown selector initial mapping
2d02cbf
-
experimentation and production mode frontend
f61fc2e
-
initial benchmark and improve query rewriting
5c9ec12
-
query rewriting initial
8bfabbf
-
Added multi model selector capability in chat
8c68a0d
-
feat(metadata): add section-title extraction and trace visibility
51d3cde
-
Add sigmoid-normalized reranker confidence & margin, compute retrieval consensus
6d3a979
-
debug
96f6ab5 -
Patch the trace serialization / pipeline assembly layer
45a3389
-
hybridize
d510f69 -
readme update
a8e5c8a
-
Increase context injection size
bebb100 -
Improve RAG chunking and retrieval observability
b9a80c6 -
debug
c88eaf7 -
Initial PilotMaster deployment
e7e5bf0