Commit Graph

4238 Commits

Author SHA1 Message Date
Debanjum
bb6a6cbe19 Restart Khoj docker services unless stopped. Remove default network
- Seems less aggressive to use unless-stopped versus always.
- Default network is used anyway, so doesn't seem necessary to specify
2025-01-09 22:08:48 +07:00
Debanjum
01d27f5220 Do not show user logout button on web app side pane in anoymous mode
Refer
https://github.com/khoj-ai/khoj/issues/1050#issuecomment-2579119234
2025-01-09 21:18:50 +07:00
Debanjum
a739936563 Mark Github integration as unmaintained in documentation
Also mention what the Reflective Questions table is about

Resolves #1059
2025-01-09 20:42:38 +07:00
Debanjum
1eaa54b0ae Make PAT token requirement optional for Github indexing for now
The github integration has not been tested and may still be broken.
This change at least makes it easier to add repositories without
needing a PAT token if/when it does work.
2025-01-09 20:42:38 +07:00
sabaimran
c448c49811 Clean-up some code on homepage and disable initial card animations because of jitter 2025-01-08 22:07:23 -08:00
Ikko Eltociear Ashimine
9ce9f02886 Fix typo in Admin Doc (#1034)
appropiate -> appropriate
2025-01-08 21:27:34 -08:00
sabaimran
ad5f0c7a02 Merge pull request #1029 from DPS0340/master
Improve docker-compose.yml

- Do not expose dependencies on host internet
- Put all services on the same network
2025-01-08 10:42:47 -08:00
sabaimran
875cdde9b9 Release Khoj version 1.33.2 2025-01-07 15:32:18 -08:00
sabaimran
5c5c4a6bbc Add help text for Enterprises in the README 2025-01-07 14:55:57 -08:00
sabaimran
8d028e10c6 Fix populating login url in sign in email 2025-01-07 14:53:00 -08:00
omahs
36bdaedd2d Fix typos in Khoj Docs (#1033) 2025-01-07 15:55:57 +07:00
sabaimran
25c1c1c591 Release Khoj version 1.33.1 2025-01-06 09:08:01 -08:00
sabaimran
689d9d8b3a Update formatting in admin.py and utils.py 2025-01-06 09:07:28 -08:00
thinker007
aa442c28eb Handle reporting chat estimated cost when some fields unavailable (#1026)
Fix AttributeError: 'NoneType' object has no attribute 'model_extra'

* cost = chunk.usage.model_extra.get("estimated_cost", 0) if hasattr(chunk, "usage") and chunk.usage else 0  # Estimated costs returned by DeepInfra API
2025-01-06 09:03:49 -08:00
sabaimran
4aed6f7e08 Add a link around the header khojlogotype to go home 2025-01-06 08:55:00 -08:00
Debanjum
266d274e21 Make automation should_notify check robust to non json mode chat models
Use clean_json to handle automation should_notify check For gemini and
other chat models where enforcing json mode is problematic, not supported
2025-01-06 20:16:32 +07:00
Debanjum
9a5e3583cf Remove bullet styling only from sidebar items on web app
Previous fix had removed bullet styling from all components in web
app. This made chat messages on the web app lose bullet styling too.
2025-01-06 20:15:42 +07:00
Debanjum
dc0bc5bcca Evaluate information retrieval quality using eval script
- Encode article urls in filename indexed in Khoj KB
  Makes it easier for humans to compare, trace retrieval performance
  by looking at logs than using content hash (which was previously
  explored)
2025-01-06 13:19:52 +07:00
Debanjum
daeba66c0d Optionally pass references used by agent for response to eval scorers
This will allow the eval framework to evaluate retrieval quality too
2025-01-06 13:19:52 +07:00
Debanjum
8231f4bb6e Return accuracy as decision to generalize across IR & standard scorers 2025-01-06 13:19:52 +07:00
Jiho Lee
c1c086e431 fix: Use localhost on SEARXNG_BASE_URL
Co-authored-by: sabaimran <65192171+sabaimran@users.noreply.github.com>
2025-01-06 13:32:06 +09:00
sabaimran
eb9aadf72a Add an Obsidian README documentation for development 2025-01-05 19:06:27 -08:00
sabaimran
e89e49818b Merge pull request #1028 from ReallyVirtual/patch-1
Update image_generation.md
2025-01-05 13:53:00 -08:00
sabaimran
616cc189d1 Remove bullet points from li styling explicitly 2025-01-05 13:52:08 -08:00
sabaimran
a5705a5aa6 After agent prompt safe check is parsed as json, load it into a json object 2025-01-05 13:40:20 -08:00
Jiho Lee
a5c7315874 feat: Improve docker-compose.yml
- Remove host port mappings on dependencies
- Add 'restart: always'
- Add default network for lookup by docker dns
2025-01-05 15:17:50 +09:00
Sohaib Athar
95a2387e9b Update image_generation.md
Fixed typo
2025-01-05 07:25:41 +05:00
sabaimran
756f4a2a66 Update regeneration logic to run for all entries now that we have a single search model ID 2025-01-03 15:16:10 -08:00
Debanjum
33f85b4a55 Fix title of Query Filters documentation 2025-01-01 18:56:16 +08:00
sabaimran
0ef787b57c Set waitBeforeSeconds to 0 2024-12-30 15:08:34 -08:00
Debanjum
90b4e03454 Pull out query filters as top level documentation page
- Note perf eval from 2022
- Update links to query-filters in docs
- Fix links
- Update image model docs
2024-12-30 14:35:05 -08:00
sabaimran
8f69eb949b Release Khoj version 1.33.0 2024-12-29 10:39:26 -08:00
Debanjum
1a43ca75f3 Update to latest jinja python package dependency 2024-12-27 01:44:41 -08:00
Debanjum
ca197ba9ba Improve Automation Flexibility and Automation Email Format (#1016)
- Format AI response to send in automation email
- Let Khoj infer chat query based on user automation query
- Decide if automation emails should be sent based on response
  - Fix the `to_notify_or_not` decider AI
  - Ask reason before decision to improve to_notify decider AI
- Show error message on web app when fail to create/update automation
2024-12-27 01:36:38 -08:00
Debanjum
90685ccbb0 Show error message if update automation via web app fails 2024-12-26 22:21:56 -08:00
Debanjum
c4bb92076e Convert async create automation api endpoints to sync 2024-12-26 21:59:55 -08:00
Debanjum
9674de400c Format AI response to send in automation email
Previously we sent the AI response directly. This change post
processes the AI response with the awareness that it is to be sent to
the user as an email to improve rendering and quality of the emails.
2024-12-26 21:04:50 -08:00
Debanjum
6d219dcc1d Switch to let Khoj infer chat query based on user automation query
This tries to decouple the automation query from the chat query. So
the chat model doesn't have to know it is running in an automation
context or figure how to notify user or send automation response. It
just has to respond to the AI generated `query_to_run' corresponding
to the `scheduling_request` automation by the user.

For example, a `scheduling_request' of `notify me when X happens'
results in the automation calling the chat api with a `query_to_run`
like `tell me about X` and deciding if to notify based on information
gathered about X from the scheduled run. If these two are not
decoupled, the chat model may respond with how it can notify about X
instead of just asking about it.

Swap query_to_run with scheduling_request on the automation web page
2024-12-26 21:04:50 -08:00
Debanjum
3600a9a4f3 Ask reason before decision to improve to_notify decider automation AI
Previously it just gave a decision. This was hard to debug during
prompt tuning. Asking for reason before decision improves models
decision quality.
2024-12-26 21:04:50 -08:00
Debanjum
dcc5073d16 Fix the to_notify decider automation chat actor. Add detailed logging 2024-12-26 21:04:50 -08:00
sabaimran
03b4667acb Merge pull request #1017 from khoj-ai/features/update-home-page
- Rather than chunky generic cards, make the suggested actions more action oriented, around the problem a user might want to solve. Give them follow-up options. Design still in progress.
2024-12-24 12:12:37 -08:00
sabaimran
5985ef4c7c Further improve prompts 2024-12-24 12:11:27 -08:00
sabaimran
012e0ef882 Add tooltip for file input ref. 2024-12-24 11:20:39 -08:00
sabaimran
a58b3b4a37 Remove some of the step one suggestions 2024-12-24 10:52:42 -08:00
sabaimran
cf78f426d3 Merge branch 'master' of github.com:khoj-ai/khoj into features/update-home-page 2024-12-24 09:52:06 -08:00
sabaimran
d27d291584 Merge pull request #1015 from khoj-ai/features/clean-up-authenticated-data
- De facto, was being assumed everywhere if authenticatedData is null, that it's not logged in. This isn't true because the data can still be loading. Update the hook to send additional states.
- Bonus: Delete model picker code and a slew of unused imports.
2024-12-24 09:51:39 -08:00
sabaimran
cd4cf4f9f6 Merge pull request #1014 from khoj-ai/features/improve-agent-management
- Add support for seeing all steps of the agent modification flow via tabs at the top of the modal
- Separate knowledge base & tool selection into two separate parts
2024-12-24 09:50:38 -08:00
sabaimran
3e6ba45cbe Merge pull request #1013 from khoj-ai/features/use-sidebar
Use the [shadcn sidebar](https://ui.shadcn.com/docs/components/sidebar#sidebarmenusub) across Khoj and standardize how the side panel experience works across the app. This helps us generalize the code better, while re-using the same components. 

Deprecates current usage of the chat history side panel, replacing it with the new `appSidebar.tsx` component.

We'll eventually move out the `Manage Files` section and move it into a separate panel dedicated to chat-level controls.
2024-12-24 09:50:09 -08:00
sabaimran
95fdbe13ae move all conversations button to bottom of side panel 2024-12-23 23:13:19 -08:00
sabaimran
d0256f267e Fix submit state for form with buttons sticky to button 2024-12-23 23:10:19 -08:00