Commit graph

4041 commits

Author SHA1 Message Date
sabaimran
e919d28f1c Add support for custom search model-specific thresholds 2024-08-24 19:28:26 -07:00
sabaimran
fa4d808a5f Encode uri components when sending automations data to the server 2024-08-24 18:45:50 -07:00
sabaimran
387b7c7887 Release Khoj version 1.21.3 2024-08-23 11:15:15 -07:00
sabaimran
7b8b3a66ae Revert django version to previous patch 2024-08-23 11:12:41 -07:00
Debanjum Singh Solanky
5927ca8032 Properly close chat stream iterator even if response generation fails
Previously chat stream iterator wasn't closed when response streaming
for offline chat model threw an exception.

This would require restarting the application. Now application doesn't
hang even if current response generation fails with exception
2024-08-23 02:06:26 -07:00
Debanjum Singh Solanky
bdb81260ac Update docs to mention using Llama 3.1 and 20K max prompt size for it
Update stale credits to better reflect bigger open source dependencies
2024-08-22 20:27:58 -07:00
Debanjum Singh Solanky
238bc11a50 Fix, improve openai chat actor, director tests & online search prompt 2024-08-22 19:09:33 -07:00
Debanjum Singh Solanky
9986c183ea Default to gpt-4o-mini instead of gpt-3.5-turbo in tests, func args
GPT-4o-mini is cheaper, smarter and can hold more context than
GPT-3.5-turbo. In production, we also default to gpt-4o-mini, so makes
sense to upgrade defaults and tests to work with it
2024-08-22 19:04:49 -07:00
Debanjum Singh Solanky
8a4c20d59a Enforce json response by offline models when requested by chat actors
- Background
  Llama.cpp allows enforcing response as json object similar to OpenAI
  API. Pass expected response format to offline chat models as well.

- Overview
  Enforce json output to improve intermediate step performance by
  offline chat models. This is especially helpful when working with
  smaller models like Phi-3.5-mini and Gemma-2 2B, that do not
  consistently respond with structured output, even when requested

- Details
  Enforce json response by extract questions, infer output offline
  chat actors
  - Convert prompts to output json objects when offline chat models
    extract document search questions or infer output mode
  - Make llama.cpp enforce response as json object

- Result
  - Improve all intermediate steps by offline chat actors via json
    response enforcement
  - Avoid the manual, ad-hoc and flaky output schema enforcement and
    simplify the code
2024-08-22 18:07:44 -07:00
Debanjum Singh Solanky
ab7fb5117c Release Khoj version 1.21.2 2024-08-20 12:38:54 -07:00
Debanjum Singh Solanky
de24ffcf0d Upgrade Axios, a desktop app dependency, to version 1.7.4 2024-08-20 12:32:36 -07:00
Debanjum Singh Solanky
a60baa55fb Upgrade Django, a Khoj server dependency, to version 5.0.8 2024-08-20 12:32:00 -07:00
sabaimran
1ac8de6c3a Release Khoj version 1.21.1 2024-08-20 11:55:34 -07:00
Debanjum Singh Solanky
5d59acd1f4 Stop pushing deprecated khoj-assistant package to pypi
- Also skip uploading package version to it already exists on pypi
  This happens when a release is new khoj tagged release is created
2024-08-20 11:43:02 -07:00
sabaimran
f6ce2fd432 Handle end of chunk logic in openai stream processor 2024-08-20 10:50:09 -07:00
sabaimran
029775420c Release Khoj version 1.21.0 2024-08-20 10:01:56 -07:00
sabaimran
4808ce778a
Merge pull request #892 from khoj-ai/upgrade-offline-chat-models-support
Upgrade offline chat model support. Default to Llama 3.1
2024-08-20 11:51:20 -05:00
Debanjum Singh Solanky
58c8068079 Upgrade default offline chat model to llama 3.1 2024-08-20 09:28:56 -07:00
sabaimran
2d9dd81e76 Re-add authenticated decorator to api_chat.py /chat endpoint 2024-08-19 05:37:18 -05:00
sabaimran
2c5350329a Remove the hashes from titles in found relevant notes 2024-08-18 22:31:15 -05:00
Debanjum Singh Solanky
acdc3f9470 Unwrap any json in md code block, when parsing chat actor responses
This is a more robust way to extract json output requested from
gemma-2 (2B, 9B) models which tend to return json in md codeblocks.

Other models should remain unaffected by this change.

Also removed request to not wrap json in codeblocks from prompts. As
code is doing the unwrapping automatically now, when present
2024-08-16 14:16:29 -05:00
Debanjum Singh Solanky
ca45fce8ac Break long links in train of thought to stay within chat page width 2024-08-16 14:16:29 -05:00
sabaimran
c0316a6b5d
Enable free tier users to have unlimited chats with the default chat model (#886)
- Allow free tier users to have unlimited chats with default chat model. It'll only be rate-limited and at the same rate as subscribed users
- In the server chat settings, replace the concept of default/summarizer models with default/advanced chat models. Use the advanced models as a default for subscribed users.
- For each `ChatModelOption' configuration, allow the admin to specify a separate value of `max_tokens' for subscribed users. This allows server admins to configure different max token limits for unsubscribed and subscribed users
- Show error message in web app when hit rate limit or other server errors
2024-08-16 12:14:44 -07:00
Debanjum
8dad9362e7
Improve search model config display for admin (#887) from aam-at/feature/improve_search_model_config_admin
Currently, the search model config display for admins only shows the id of the search model config, which is not very informative. 

The changes enhances the admin console by displaying the name of the search model config (name), as well as the bi-encoder model (bi_encoder) and cross-encoder model (cross_encoder) along the id.
2024-08-16 07:33:55 -07:00
Debanjum
2b1482d2b4
Fix indexing content from Emacs #883 from aam-at/bugfix/fix_emacs_if
Previously `force' was passed as a query param to the single indexing API. After the recent API updates, it is meant to select the API method to use (PATCH vs PATCH). Converting `force' argument to a bool fixes implementing this new behavior
2024-08-16 07:32:46 -07:00
Debanjum
0b568e204e
Add model_config for cross-encoder model (#885) from aam-at/feature/crossencoder_model_config
Add `model_config' for the cross-encoder model, so the server admin can use models which require the `trust_remote_code' argument to run locally
2024-08-16 07:32:19 -07:00
Debanjum
39e566ba91
Improve Document, Online Search to Answer Vague or Meta Questions (#870)
- Major
  - Improve doc search actor performance on vague, random or meta questions
  - Pass user's name to document and online search actors prompts

- Minor
  - Fix and improve openai chat actor tests
  - Remove unused max tokns arg to extract qs func of doc search actor
2024-08-16 06:46:13 -07:00
Debanjum Singh Solanky
27ad9b1302 Remove unused max tokns arg to extract qs func of doc search actor 2024-08-13 12:53:39 +05:30
Debanjum Singh Solanky
f75606d7f5 Improve doc search actor performance on vague, random or meta questions
- Issue
  Previously the doc search actor wouldn't extract good search queries
  to run on user's documents for broad, vague questions.

- Fix
  The updated extract questions prompt shows and tells the doc search
  actor on how to deal with such questions

  The doc search actor's temperature was also increased to support more
  creative/random questions. The previous temp of 0 was meant to
  encourage structured json output. But now with json mode, a low temp is
  not necessary to get json output
2024-08-13 12:53:39 +05:30
Debanjum Singh Solanky
3675938df6 Support passing temperature to offline chat model chat actors
- Use temperature of 0 by default for extract questions offline chat
  actor
- Use temperature of 0.2 for send_message_to_model_offline (this is
  the default temperature set by llama.cpp)
2024-08-13 12:53:00 +05:30
Shantanu Sakpal
b5bcce7f85
Cycle through chat history in chat input on Obsidian (#861)
* Add ability to cycle through the chat history in the chat input on Obsidian (similar to terminal history navigation)
* Add mod key shortcut to cycle through chat history in chat input
* Add shortcut help text in chat input placeholder

---------

Co-authored-by: Debanjum Singh Solanky <debanjum@gmail.com>
2024-08-12 23:55:25 -07:00
srikary12
05c0aa3882
Support exclusion file filters (#826)
### Overview
Support exclude file filter in user search queries

### Details
- All of the exclude file filter terms need to be satisfied
- Any one of the include file filter terms should be satisfied

### Example
- **Search Query**: *what happened yesterday? -file:"tasks.org" -file:"work.md" file:"diary.org" file:"journal.org*
- **Behavior**: Query will try find relevant notes in any of `journal.org` or `diary.org` and not in `tasks.org` and not in `work.md`

### Details
* Add support for exclusion file filters
* Translate file filter to valid Django DB entry filter regex
* Exclude all files when multiple exclude file filter in query

Previously we were applying an "Or" filter, which would exclude any
file mentioned in a query with multiple exclude file filter.

This is not what we naturally mean when we ask excluding a file in a query

* Rename, rearrange, deduplicate and add file filter tests

Closes #728
---------

Co-authored-by: Debanjum Singh Solanky <debanjum@gmail.com>
2024-08-12 05:41:54 -07:00
Alexander Matyasko
2d9bf14ecb Improve search model config display for admin 2024-08-11 19:13:25 +08:00
Debanjum Singh Solanky
7815e02dd4 Release Khoj version 1.20.4 2024-08-11 16:00:13 +05:30
Debanjum Singh Solanky
d951e36945 Update khoj.el package description, it had gone stale 2024-08-11 15:52:46 +05:30
Debanjum Singh Solanky
16b31c3e35 Refresh automation data shown by edit automation card after update
Previously required the automation page to be refreshed to see updates
to the automation in the edit automation card. This would be seen when
user tries to edit an automation multiple times (without a page refresh)
2024-08-11 15:52:46 +05:30
Debanjum Singh Solanky
f2f37ae444 Fix creating, editing automations that start weekly on Sunday 2024-08-11 15:52:46 +05:30
Debanjum Singh Solanky
ec9add9a51 Fix automation edit cards height. Scroll when card longer than screen 2024-08-11 15:52:46 +05:30
sabaimran
d99f03e4f3 If the list of choices in a chunk is empty, continue in openai response 2024-08-11 15:30:09 +05:30
Alexander Matyasko
f16b0f628b Fix true/false evaluation in Emacs to prevent unintended index re-indexing
Previously, the code incorrectly treated all non-nil values as true, leading to
the index being re-indexed with the force flag whenever the user selected to
update the index.
2024-08-11 17:24:11 +08:00
Alexander Matyasko
0e9e9648e6 Fix emacs if syntax 2024-08-11 17:24:11 +08:00
sabaimran
6f94a076f7 Add conversation_id parameter to the create_automation method 2024-08-11 10:45:13 +05:30
sabaimran
acb825f4f5 Bug fixes for automations
- Pass the new conversation id as kwarg for the scheduled_chat function
- For edit automations, re-use the original conversation id
- Parse images correctly for image automations
2024-08-11 10:41:43 +05:30
Debanjum Singh Solanky
5075d13902 Give visual feedback when interact with chat message feedback buttons
- Use color to provide visual feedback when hover, click on feedback
  buttons
- Use color to provide visual feedback when hover on speech, copy
  buttons click
- Add cooldown period before being able to send feedback on that message again.
  Avoids inadvertent multiple consecutive clicks on feedback buttons
2024-08-10 20:09:52 +05:30
Debanjum Singh Solanky
b3c6c8c84b Add OpenGraph metadata to web app pages for improve social share links 2024-08-10 18:14:05 +05:30
Debanjum Singh Solanky
fc411091c8 Add apple favicon, load favicons for each web app page from assets folder 2024-08-10 18:14:05 +05:30
Debanjum Singh Solanky
a7623e64fa Move Khoj webmanifest, assets to new web app public directory 2024-08-10 18:14:04 +05:30
sabaimran
af1d4b9ba4 Remove the premium requirement for speech for now 2024-08-10 14:10:12 +05:30
sabaimran
1d581464e6 Filter out any undefined agents when rendering the home page 2024-08-10 13:33:55 +05:30
sabaimran
acf1c14122 Release Khoj version 1.20.3 2024-08-09 18:11:11 +05:30