# PR Summary
This small PR resolves the deprecation warnings on `datetime` in
Python3.12+. You can find them in the [CI
logs](https://github.com/khoj-ai/khoj/actions/runs/14538833837/job/40792624987#step:9:134):
```python
/__w/khoj/khoj/src/khoj/processor/content/images/image_to_entries.py:61: DeprecationWarning: datetime.datetime.utcnow() is deprecated and scheduled for removal in a future version. Use timezone-aware objects to represent datetimes in UTC: datetime.datetime.now(datetime.UTC).
timestamp_now = datetime.utcnow().timestamp()
```
- Update API to allow free tier users to switch between free models
- Update web app to allow model switching on agent creation, settings
chat page (via right side pane), even for free tier users.
Previously the model switching APIs and UX fields on web app were
completely disabled for free tier users
Rely on deepthought flag to control reasoning effort of low/high for
the grok model
This is different from the openai reasoning models which support
low/medium/high and for which we use low/medium effort based on the
deepthought flag
Note: grok is accessible over an openai compatible API
Disregard chart types as not using rich chart rendering
and they are duplicate of chart images that are rendered
Disregard text output associated with generated image files
Server:
- Rate limit based on unverified email before creating user
- Check email address for deliverability before creating user
- Track rate limit for unverified email in new non-user keyed table
Web app:
- Show error in login popup to user on failure/throttling
- Simplify login popup logic by moving magic link handling logic
into EmailSigninContext instead of passing require props via parent
Previously the sidebar could recurse on opening chat page (from home?)
due to child modelSelector component updating parent chatSidebar prop
which was passed back down to it in a loop.
The chatSidebar decides if agent has been modified in a single
useEffect and enables the Save button accordingly.
- Track agent modification wrt agent info received from server in
chatSidebar instead.
- Reduce modelSelector's mandate to just notify
when the user changes the model.
- Fix to infer, show & update agent state from chat sidebar on web app
This logic is fragile and convoluted because:
- the default agent chat model is dynamically determined.
- need to disambiguate tools not set vs none set vs all set by user
The default agent's tool selection is stored as undefined to show
not set scenario, which allows for all tools to be dynamically
used by agent.
But the user can also set no tools or all tools for their agents.
All 3 scenarios are handled differently.
- Track tools to be displayed vs tools to be stored
This is triggered by mismatch between "dark" class present on server
sent layout but not in client sent layout on initial render.
That mismatch exists because the server applies dark-mode styling
early to avoid FOUC flickering of UX.
Related 43e032e
Remove html, body elements from child page layouts. Let only the root
layout handle it.
Next.js router structure mounts child layouts inside parent layouts,
as defined by their directory hierarchy. So the html, body component
should only be defined in the parent layout.
This avoids the child layout mounting its html, body component within
the actual root layout's existing html, body component.
Previously the chat model associated with the default agent was always
the first chat model populated on the server. This doesn't match
behavior of the rest of the system, where the server chat settings is
preferred over the user chat settings over the first chat model.
This change brings the default agent's chat model in line with the
preference order used in the reset of the system.
Previous change to fallback to default agent was not functional. It
would error out if the conversation agent wasn't set when trying to
get conversation.agent.slug for calling aget_agent_by_slug func
Issue introduced in commit 5a3c7b1.
Usage of KHOJ_DOMAIN
---
KHOJ_DOMAIN is tri-state for local, official and other production deployments:
- If KHOJ_DOMAIN is unset (for local):
- sets CSRF cookie to localhost
- adds khoj.dev variants to ALLOWED_HOSTS, CSRF_TRUSTED_ORIGINS
- adds app.khoj.dev variants to CORS origins
- If KHOJ_DOMAIN is set to empty (for official):
- sets CSRF to khoj.dev
- adds khoj.dev variants to ALLOWED_HOSTS, CSRF_TRUSTED_ORIGINS
- adds app.khoj.dev variants to CORS origins
- If KHOJ_DOMAIN is set (for other prod deployments):
- sets CSRF cookie to KHOJ_DOMAIN
- adds KHOJ_DOMAIN variants to ALLOWED_HOSTS, CSRF_TRUSTED_ORIGINS
- adds KHOJ_DOMAIN variants to CORS origins
Related #1137, #1152Resolves#1123
Unsure why this error triggers on every request to the Django admin
panel these days but all the requests are completing fine and the
client is clearly not aborting the request when the RequestAborted
exception is raised.
Suppress these errors for now via middleware to prevent them from
unnecessarily cluttering up the server logs and confusing folks.
Related #1152
Previously Gemini 2 flash and flash lite were using context window of
10K by default as no defaults were added for it.
Increase default context for small commercial models to 120K from 60K
as cheaper and faster than their pro models equivalents at 60K context.
We'd moved research planner to only use tools in enum of schema. This
enum tool enforcement prevented model from terminating research by
setting tool field to empty.
Fix the issue by adding text tool to research tools enum and tell
model to use that to terminate research and start response instead.
Make research planner consistently select tool before query. As the
model should tune it's query for the selected tool. It got space to
think about tool to use in the scratchpad already.
- Control auto read webpage via eval workflow. Prefix env var with KHOJ_
Default to false as it is the default that is going to be used in prod
going forward.
- Set openai api key via input param in manual eval workflow runs
- Simplify evaluating other chat models available over openai
compatible api via eval workflow.
- Mask input api key as secret in workflow.
- Discard unnecessary null setting of env vars.
- Control randomization of samples in eval workflow.
If randomization is turned off, it'll take the first SAMPLE_SIZE
items from the eval dataset instead of a random collection of
SAMPLE_SIZE items.
Only setup speech to text and text to image models served via openai
compatible APIs when explicitly specified during initialization.
This avoids setup of whisper and dalle when an openai compatible API
is being setup instead of the openai API itself.
- Specify min, max number of list items expected in AI response via JSON schema enforcement. Used by Gemini models
- Warn and drop invalid/empty messages when format messages for Gemini models
- Make Gemini response adhere to the order of the schema property definitions
- Improve agent creation safety checker by using response schema, better prompt
Without explicitly using the property ordering field, gemini returns
responses in alphabetically sorted property order.
We want the model to respect the schema property definition order.
This ensures control during development to maintain response quality.
For example in CoT make it fill scratchpad before answers.
Require at least 1 item in lists. Otherwise gemini flash will
sometimes return an empty list. For chat actors where max items is
known, set that as well.
OpenAI API does not support specifying min, max items in response
schema lists, so drop those properties when response schema is
passed. Add other enforcements to response schema to comply with
response schema format expected by OpenAI API.
