17 Commits

Author	SHA1	Message	Date
Paul Koudelka	c8c1e38475	made tokenizer concurrent and threadsafe	2026-08-10 21:36:01 +02:00
Paul Koudelka	6985678d2f	changed DateTime to DateTimeOffset	2026-08-10 20:06:51 +02:00
Paul Koudelka	9ab196c13c	ensured prompts fit embedding model	2026-08-10 19:42:39 +02:00
Paul Koudelka	3918b48409	fixed tokenization issue for long files	2026-08-10 18:21:31 +02:00
PaulKoudelka	2a3a82820c	forced overlap on chunks	2026-08-04 15:08:47 +02:00
PaulKoudelka	010f35b08a	added retrieval with vector and bm25 search	2026-08-03 20:56:19 +02:00
PaulKoudelka	c78a8f05ef	added metadata for bm25 and also confidence levels	2026-08-03 19:45:47 +02:00
PaulKoudelka	b3ba945b98	made the embedding procedure more robust	2026-08-03 17:53:31 +02:00
PaulKoudelka	89f7156fdf	improved chunking pipeline and handling of expert settings	2026-07-29 18:47:59 +02:00
PaulKoudelka	fd2d99872a	added warning and improved error handling	2026-07-28 20:22:37 +02:00
PaulKoudelka	759c7f9d89	improved stability of embedding chunks	2026-07-28 17:36:06 +02:00
PaulKoudelka	1bf9328fb9	made the chunking process more robust	2026-07-28 16:37:07 +02:00
PaulKoudelka	95ca74809b	made file watchers more robust against multiple triggers	2026-07-28 15:52:59 +02:00
PaulKoudelka	c59e33e1b1	added sql integration instead of json files	2026-07-28 15:25:10 +02:00
Paul Koudelka	6989cbe3b2	changed qdrant to qdrant edge	2026-06-10 17:07:34 +02:00
Paul Koudelka	ea87d79ca9	current main is now successfully merged into branch	2026-05-27 20:02:43 +02:00
Paul Koudelka	637bf9701b	added manual refresh and cleaned up code	2026-05-13 18:13:34 +02:00