33 Commits

Author SHA1 Message Date
Adityavardhan Agrawal
adc0b2dbb8
Add batch ingestion (#55) 2025-03-18 23:27:53 -04:00
Adityavardhan Agrawal
4ae132ff46
Implement knowledge graphs, and graph enhanced querying (#48) 2025-03-17 17:36:43 -04:00
Arnav Agrawal
989d25d8c7 colpali hotfix 2025-03-16 17:08:38 -04:00
Adityavardhan Agrawal
32a5d787fe
Add update methods with add update strategy (#53) 2025-03-13 11:26:01 -04:00
LukeZekes
39d7bc7bfd
Include chunk score in query response and don't attempt to initialize colpali store if it is disabled (#52) 2025-03-11 15:53:42 -04:00
Adityavardhan Agrawal
38683df0f3
Add completion sources and batch retrieval for docs and chunks (#51) 2025-03-09 18:42:04 -04:00
LukeZekes
e56691a1c5
add filename option for text documents (#47) 2025-03-05 10:56:02 -05:00
Arnav Agrawal
8428616dd6 add support for .docx files 2025-02-28 15:06:59 -05:00
Arnav Agrawal
07eec6b9e3 add image processing to ollama 2025-02-26 22:36:25 -05:00
Arnav Agrawal
821e9d7e20
Add support for ColPali (#43)
* debug mps not supported

* further debug (i think i lost some braincells)

* fix mps bug and resolve dependency issues

* remove libmagic dependence

* add colpali embedding model

* multi-vector store works - verified with testing

* add integration testing

* support text embedding in colpali

* complete colplai integration and testing

* formatting + some PR comments

* remove experimental files

* resolve PR comments
2025-02-26 20:17:12 -05:00
Adityavardhan Agrawal
c8ed46b12b
Separate parsing and chunking into different function for easy rules processing (#41) 2025-02-15 13:02:15 -05:00
Adityavardhan Agrawal
a46fa064c7
Add natural language rules based ingestion (#34) 2025-02-07 21:08:40 -05:00
Arnav Agrawal
d124e6aa0d
Add support for cache-augmented-generation (#30) 2025-01-28 23:49:28 -05:00
Adityavardhan Agrawal
0a933e5fd6
Add docker support (#24) 2025-01-09 05:17:25 -05:00
Arnav Agrawal
20faae8903
Add reranking (#14) 2025-01-02 03:42:47 -05:00
Arnav Agrawal
48e6aeb8b7
use local unstructured by default (#12) 2025-01-01 09:18:23 -05:00
Arnav Agrawal
abccf99974
add contextual embedding with claude prompt caching (#11)
* add context augmentation while chunking

* add contextual embeddings

* default config should be combined

* fix comments on PR

* update example environment

* update config and api to support env-variable optionality
2024-12-31 06:58:34 -05:00
Arnav Agrawal
3ad55129b7
Rename imports (#9) 2024-12-30 11:58:53 -05:00
Arnav Agrawal
0e4a43645a reformat files 2024-12-29 12:48:41 +05:30
Arnav Agrawal
80db083471 added frame/transcript augmentation for video retrieval 2024-12-29 12:45:12 +05:30
Arnav Agrawal
7830b42c6b fix typing errors 2024-12-29 11:10:51 +05:30
Arnav Agrawal
196655fea3 pipethrough video timestamps on query 2024-12-28 19:41:05 +05:30
Arnav Agrawal
418054e9a3 update configuration style to support easy model editing 2024-12-27 11:19:07 +05:30
Arnav Agrawal
13ab54fbf8
add a video parser + formatting changes (#4) 2024-12-26 11:34:24 -05:00
Adityavardhan Agrawal
03345dcc07
Add completions API (#3) 2024-12-26 08:52:25 -05:00
Arnav Agrawal
4f2f221d40 bug fixes and end-to-end testing 2024-12-17 21:40:38 -05:00
Adityavardhan Agrawal
df8d7fcdd0
refactor some stuff (#2)
* refactor some stuff, remove bare try catches
2024-12-15 14:31:25 -05:00
Adityavardhan Agrawal
251e38828a clean up 2024-12-04 20:26:14 -05:00
Adityavardhan Agrawal
1f68fb99d3 sdk and querying in api works 2024-12-03 21:46:25 -05:00
Arnav Agrawal
f1f52d9b67 pass api tests 2024-12-02 20:03:35 -05:00
Arnav Agrawal
000887a4dc pass all tests apart from querying 2024-11-28 19:09:40 -05:00
Adityavardhan Agrawal
983a4ee854 separate text and doc ingestion pathways 2024-11-24 14:29:25 -05:00
Adityavardhan Agrawal
d70f53cf86 system changes 2024-11-22 20:58:17 -05:00