Weves
82b9cb4cc1
Add check to ensure auth is enabled for every endpoint unless explicitly whitelisted
2024-04-19 01:26:24 -07:00
Yuhong Sun
89ff07a96b
Slack improvement
2024-04-17 21:38:00 -07:00
Yuhong Sun
1b41ec2b50
Remove Search Only Model ( #1331 )
2024-04-14 18:49:07 -07:00
Yuhong Sun
a17060af5a
Provide Additional Context for Chunk Options in APIs ( #1330 )
2024-04-14 18:32:22 -07:00
Weves
31bfbe5d16
Fix chat sharing
2024-04-09 11:57:09 -07:00
Weves
eb367de44d
Small token budget tweaks
2024-04-04 20:58:45 -07:00
Chris Weaver and Nick Donohue
447791b455
Token budgets ( #1302 )
...
---------
Co-authored-by: Nick Donohue <ndonohue@gmail.com >
2024-04-04 20:43:24 -07:00
Weves
7ba7224929
Allow seeding of chat sessions via POST
2024-04-04 12:59:39 -07:00
Weves
e82061a5ec
Add support for specifying title via search params
2024-04-01 11:25:57 -07:00
Yuhong Sun
5b8cdd4eee
Gpt-3.5-0125 Option ( #1282 )
2024-03-31 21:58:00 -07:00
Weves
a4869b727d
Add ability to control available pages
2024-03-31 21:49:34 -07:00
Yuhong Sun
15f7b42e2b
More Options for Search APIs ( #1280 )
2024-03-31 21:45:46 -07:00
Weves
32f55ddb8f
URL-based chat seeding
2024-03-31 18:10:49 -07:00
Weves
49acde0a8f
URL-based chat sharing
2024-03-29 00:51:17 -07:00
Weves
5a967322fd
Add ability to specify custom embedding models
2024-03-27 00:02:51 -07:00
Weves
f135ba9c0c
Rework LLM answering flow
2024-03-25 13:34:03 -07:00
Weves
1ba74ee4df
Refactor search pipeline
2024-03-25 13:34:03 -07:00
Yuhong Sun
7a861ecec4
Session Dependency for Chat Streaming ( #1256 )
2024-03-24 19:40:06 -07:00
Johannes Vass and Johannes Vass
3107edc921
Do not obtain DB session via Depends() ( #1238 )
...
Endpoints that use Depends(get_session) with a StreamingResponse have
the problem that Depends() releases the session again after the endpoint
function returns. At that point, the streaming response is not
finished yet but still holds a reference to the session and uses it.
However, there is no cleanup of the session after the answer stream
finishes which leads to the connections accumulating in state "idle in
transaction".
This was due to a breaking change in FastAPI 0.106.0
https://fastapi.tiangolo.com/release-notes/#01060
Co-authored-by: Johannes Vass <johannes.vass@cloudflight.io >
2024-03-24 19:31:07 -07:00
Weves
89e72783a7
Add some private Persona / Document Set stuff
2024-03-22 21:44:31 -07:00
Weves
ec48142a2d
Move some of the user re-work stuff to MIT repo
2024-03-22 16:29:24 -07:00
Yuhong Sun
c28a95e367
Port File Store from Volume to PG ( #1241 )
2024-03-21 20:10:08 -07:00
Yuhong Sun
3a6d32da7c
Port KV Store to Postgres ( #1227 )
2024-03-19 16:21:22 -07:00
Weves
7869f23e12
Improve slack flow
2024-03-04 19:22:46 -08:00
Weves
a8cc3d5a07
Add ability to add starter messages
2024-03-03 14:23:34 -08:00
Weves
0366f3313a
Fix gen ai model correctness check frequency
2024-03-01 21:59:18 -08:00
Yuhong Sun
0f23effe7e
Change Endpoint Assumption ( #1127 )
2024-02-23 20:19:23 -08:00
Yuhong Sun
7748f4df94
Auto-Detect if Better Default LLM available for OpenAI ( #1106 )
2024-02-21 16:10:22 -08:00
Johannes Vass
918bc385a2
Remove documents from index which are not returned by connector
2024-02-21 16:09:04 -08:00
Yuhong Sun
c1d1651b43
Option to stop sync-ing primary index when building secondary one ( #1096 )
2024-02-18 22:53:26 -08:00
Weves
6059339e61
Improve initial flow
2024-02-18 18:40:44 -08:00
Yuhong Sun
d2ce3033a2
Enable Chat Without Connectors ( #1090 )
2024-02-17 21:12:44 -08:00
Yuhong Sun
514e7f6e41
Kill Index Attempts for previous model ( #1088 )
2024-02-16 18:35:01 -08:00
Yuhong Sun
1c4f7fe7ef
Pass Tags to LLM ( #1071 )
2024-02-11 15:58:42 -08:00
Weves
236fa947ee
Add full exception trace to UI
2024-02-10 20:52:10 -08:00
Yuhong Sun
2362c2bdcc
Reindex All Backend ( #1049 )
2024-02-06 23:07:24 -08:00
Yuhong Sun
6768c24723
Default LLM Update ( #1042 )
2024-02-05 01:25:51 -08:00
Weves
e54ce779fd
Enable selection of long documents
2024-02-03 17:55:24 -08:00
Yuhong Sun
0060a1dd58
Immediate Mark Resolved SlackBot Option and Respond to Bots Option ( #1031 )
2024-02-01 22:18:15 -08:00
Weves
8b9e6a91a4
Fix change model popup
2024-01-30 10:15:58 -08:00
Yuhong Sun
4b45164496
Background Index Attempt Creation ( #1010 )
2024-01-28 23:14:20 -08:00
Moshe Zada
60a16fa46d
Add space before new line in order to fix typo ( #1013 )
2024-01-28 13:00:53 -08:00
Weves
824677ca75
Add option to add citations to Personas + allow for more chunks if an LLM model override is specified
2024-01-27 10:16:17 -08:00
Yuhong Sun
cf4ede2130
Embedding Models Table ( #1006 )
2024-01-26 18:40:53 -08:00
Yuhong Sun
d0fa02c8dc
Multiple Indices in Vespa ( #1000 )
2024-01-25 13:56:29 -08:00
Yuhong Sun
92628357df
Prevent Scheduling Multiple Queued Indexings ( #997 )
2024-01-24 16:31:29 -08:00
Itay and Yuhong Sun
692fdb4597
Gmail Connector ( #946 )
...
---------
Co-authored-by: Yuhong Sun <yuhongsun96@gmail.com >
2024-01-22 16:25:10 -08:00
Weves
fa879f7d7f
Add new APIs specifically for GPTs
2024-01-15 22:35:58 -08:00
Yuhong Sun
4cd9122ba5
Slack Followup Option ( #948 )
2024-01-15 14:26:20 -08:00
Yuhong Sun
7029bdb291
Remove unused doc endpoint ( #942 )
2024-01-13 15:22:23 -08:00
Weves