This avoids installing pgserver on linux arm64 docker builds, which it
doesn't currently support and isn't required to support as Khoj docker
images can use standard postgres server made available via our
docker-compose.yml
Use pgserver python package as an embedded postgres db,
installed directly as a khoj python package dependency.
This significantly simplifies self-hosting with just a `pip install khoj'.
No need to also install postgres separately.
Still use standard postgres server for multi-user, production use-cases.
- The 3.4.1 release of sentence tranformer fixes offline load latency
of sentence transformer models (and Khoj) by avoiding call to HF
- The 4.50.0 release of transformers is resulting in
jax error (unexpected keyword argument 'flatten_with_keys') on load.
Previously google auth library was explicitly installed only for the
cloud variant of Khoj to minimize packages installed for non
production use-cases.
But it was being implicitly installed as a dependency of an explicit
package in the default installation anyway.
Making the dependency on google auth package explicit simplifies
the conditional import of google auth in code while not incurring any
additional cost in terms of space or complexity.
Claude 3.7 Sonnet is Anthropic's first reasoning model. It provides a
single model/api capable of standard and extended thinking. Utilize
the extended thinking in Khoj's research mode.
Increase default max output tokens to 8K for Anthropic models.
- Specify E2B api key and template to use via env variables
- Try load, use e2b library when E2B api key set
- Fallback to try use terrarium sandbox otherwise
- Enable more python packages in e2b sandbox like rdkit via custom e2b template
- Use Async E2B Sandbox
- Parallelize file IO with sandbox
- Add documentation on how to enable E2B as code sandbox instead of Terrarium
Overview
- The default django admin panel UI looks pretty dated and didn't
have any Khoj specific branding
- Used the Unfold Django admin panel theme for a modern look
- Used the Khoj logo and name in Admin panel title, headings, favicons
Details:
All models shown on Admin panel need to inherit from unfold's
ModelAdmin to get styling applied. So
- Make all models on Admin panel inherit from unfold's ModelAdmin
- Subclassed UserAdmin to inherit from unfold's ModelAdmin
- Deregistered the unused Auth Group model from the Admin panel
We can add it back when its actually used. Avoid confusion for now
- Explicitly register DjangoJobExecution on admin panel and again make
it inherit from the unfold.admin.ModelAdmin
JSON5 spec is more flexible, try to load using a fast json5 parser if
the stricter json.loads from the standard library can't load the
raw complex json string into a python dictionary/list
Google's FRAMES benchmark evaluates multi-step retrieval and reasoning
capabilities of an agent.
The script uses Gemini as an LLM Judge to evaluate Khoj responses to
the FRAMES benchmark prompts against the ground truth provided by it.
RapidOCR depends on OpenCV which by default requires a bunch of GUI
paramters. This system package dependency set (like libgl1) is flaky
Making the RapidOCR dependency optional should allow khoj to be more
resilient to setup/dependency failures
Trade-off is that OCR for documents may not always be available and
it'll require looking at server logs to find out when this happens
* Add functions to chat with Google's gemini model series
* Gracefully close thread when there's an exception in the gemini llm thread
* Use enums for verifying the chat model option type
* Add a migration to add the gemini chat model type to the db model
* Fix chat model selection verification and math prompt tuning
* Fix extract questions method with gemini. Enforce json response in extract questions.
* Add standard stop sequence for Gemini chat response generation
---------
Co-authored-by: sabaimran <narmiabas@gmail.com>
Co-authored-by: Debanjum Singh Solanky <debanjum@gmail.com>
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
- Pass system message as the first user chat message as Gemma 2
doesn't support system messages
- Use gemma-2 chat format
- Pass chat model name to generic, extract questions chat actors
Used to figure out chat template to use for model
For generic chat actor argument was anyway available but not being
passed, which is confusing
- Deprecate khoj-assistant pypi package. Use more accurate and
succinct pypi project name, khoj
- Update references to sye khoj pypi package in docs and code instead
of the legacy khoj-assistant pypi package
- Update pypi workflow to publish to both khoj, khoj-assistant for now
- Update stale python 3.9 support mentioned in our pyproject. Can't
support python 3.9 as depend on latest django which support >=3.10
- Because we're using a FastAPI api framework with a Django ORM, we're running into some interesting conditions around connection pooling and clean-up. We're ending up with a large pile-up of open, stale connections to the DB recurringly when the server has been running for a while. To mitigate this problem, given starlette and django run in different python threads, add a middleware that will go and call the connection clean up method in each of the threads.
* Add support for chatting with Anthropic's suite of models
- Had to use a custom class because there was enough nuance with how the anthropic SDK works that it would be better to simply separate out the logic. The extract questions flow needed modification of the system prompt in order to work as intended with the haiku model
- Render crontime string in natural language in message & settings UI
- Show more fields in tasks web config UI
- Add link to the tasks settings page in task scheduled chat response
- Improve task variables names
Rename executing_query to query_to_run. scheduling_query to
scheduling_request
