1
0
Fork 0
mirror of https://github.com/khoj-ai/khoj.git synced 2024-12-26 14:18:08 +00:00
Commit graph

429 commits

Author SHA1 Message Date
Debanjum Singh Solanky
886d49e3a4 Merge branch 'master' into migrate-to-llama-cpp-for-offline-chat 2024-03-31 00:59:20 +05:30
Debanjum Singh Solanky
7923903d21 Improve date filter regexes to extract structured, natural, partial dates
- Much faster than using dateparser
  - It took 2x-4x for improved regex to extracts 1-15% more dates
  - Whereas It took 33x to 100x for dateparser to extract 65% - 400% more dates
  - Improve date extractor tests to test deduping dates, natural,
    structured date extraction from content

- Extract some natural, partial dates and more structured dates
  Using regex is much faster than using dateparser. It's a little
  crude but should pay off in performance.

  Supports dates of form:
  - (Day-of-Month) Month|AbbreviatedMonth Year|2DigitYear
  - Month|AbbreviatedMonth (Day-of-Month) Year|2DigitYear
2024-03-30 00:07:19 +05:30
Debanjum Singh Solanky
104eeea274 Extract natural language and locale specific dates in content
Previously we just extracted dates in YYYY-MM-DD format from content
for date filterings during search.

Use dateparser to extract dates across locales and natural language

This should improve notes returned as context when chat searches
knowledge base with date filters

Fallback to regex for date parsing from content if dateparser fails

- Limit natural date extractor capabilities to improve performance
  - Assume language is english
    Language detection otherwise takes a REALLY long time
  - Do not extract unix timestamps, timezone
    - This isn't required, as just using date and approximating dates as UTC
2024-03-30 00:06:56 +05:30
Debanjum Singh Solanky
8ca39a436c Use llama.cpp for offline chat models
- Benefits of moving to llama-cpp-python from gpt4all:
  - Support for all GGUF format chat models
  - Support for AMD, Nvidia, Mac, Vulcan GPU machines (instead of just Vulcan, Mac)
  - Supports models with more capabilities like tools, schema
    enforcement, speculative ddecoding, image gen etc.
- Upgrade default chat model, prompt size, tokenizer for new supported
  chat models

- Load offline chat model when present on disk without requiring internet
  - Load model onto GPU if not disabled and device has GPU
  - Load model onto CPU if loading model onto GPU fails
  - Create helper function to check and load model from disk, when model
    glob is present on disk.

    `Llama.from_pretrained' needs internet to get repo info from
    HuggingFace. This isn't required, if the model is already downloaded

    Didn't find any existing HF or llama.cpp method that looked for model
    glob on disk without internet
2024-03-26 22:33:01 +05:30
sabaimran
fdf78525b4
Part 2: Add web UI updates for basic agent interactions ()
* Initial pass at backend changes to support agents
- Add a db model for Agents, attaching them to conversations
- When an agent is added to a conversation, override the system prompt to tweak the instructions
- Agents can be configured with prompt modification, model specification, a profile picture, and other things
- Admin-configured models will not be editable by individual users
- Add unit tests to verify agent behavior. Unit tests demonstrate imperfect adherence to prompt specifications

* Customize default behaviors for conversations without agents or with default agents

* Add a new web client route for viewing all agents

* Use agent_id for getting correct agent

* Add web UI views for agents
- Add a page to view all agents
- Add slugs to manage agents
- Add a view to view single agent
- Display active agent when in chat window
- Fix post-login redirect issue

* Fix agent view

* Spruce up the 404 page and improve the overall layout for agents pages

* Create chat actor for directly reading webpages based on user message

- Add prompt for the read webpages chat actor to extract, infer
  webpage links
- Make chat actor infer or extract webpage to read directly from user
  message
- Rename previous read_webpage function to more narrow
  read_webpage_at_url function

* Rename agents_page -> agent_page

* Fix unit test for adding the filename to the compiled markdown entry

* Fix layout of agent, agents pages

* Merge migrations

* Let the name, slug of the default agent be Khoj, khoj

* Fix chat-related unit tests

* Add webpage chat command for read web pages requested by user

Update auto chat command inference prompt to show example of when to
use webpage chat command (i.e when url is directly provided in link)

* Support webpage command in chat API

- Fallback to use webpage when SERPER not setup and online command was
  attempted
- Do not stop responding if can't retrieve online results. Try to
  respond without the online context

* Test select webpage as data source and extract web urls chat actors

* Tweak prompts to extract information from webpages, online results

- Show more of the truncated messages for debugging context
- Update Khoj personality prompt to encourage it to remember it's capabilities

* Rename extract_content online results field to webpages

* Parallelize simple webpage read and extractor

Similar to what is being done with search_online with olostep

* Pass multiple webpages with their urls in online results context

Previously even if MAX_WEBPAGES_TO_READ was > 1, only 1 extracted
content would ever be passed.

URL of the extracted webpage content wasn't passed to clients in
online results context. This limited them from being rendered

* Render webpage read in chat response references on Web, Desktop apps

* Time chat actor responses & chat api request start for perf analysis

* Increase the keep alive timeout in the main application for testing

* Do not pipe access/error logs to separate files. Flow to stdout/stderr

* [Temp] Reduce to 1 gunicorn worker

* Change prod docker image to use jammy, rather than nvidia base image

* Use Khoj icon when Khoj web is installed on iOS as a PWA

* Make slug required for agents

* Simplify calling logic and prevent agent access for unauthenticated users

* Standardize to use personality over tuning in agent nomenclature

* Make filtering logic more stringent for accessible agents and remove unused method:

* Format chat message query

---------

Co-authored-by: Debanjum Singh Solanky <debanjum@gmail.com>
2024-03-26 18:13:24 +05:30
Debanjum
586654e2af
Allow directly reading web pages, even when SERP not enabled ()
### Overview
Khoj can now read website directly without needing to go through the search step first

### Details
- Parallelize simple webpage read and extractor
- Rename extract_content online results field to web pages
- Tweak prompts to extract information from webpages, online results
- Test select webpage as data source and extract web urls chat actors

- Render webpage read in chat response references on Web, Desktop apps
- Pass multiple webpages with their urls in online results context

- Support webpage command in chat API
- Add webpage chat command for read web pages requested by user
- Create chat actor for directly reading webpages based on user message
2024-03-24 16:25:25 +05:30
Debanjum Singh Solanky
85c62efca1 Test select webpage as data source and extract web urls chat actors 2024-03-24 15:46:29 +05:30
sabaimran
8abc8ded82
Part 1: Server-side changes to support agents integrated with Conversations ()
* Initial pass at backend changes to support agents
- Add a db model for Agents, attaching them to conversations
- When an agent is added to a conversation, override the system prompt to tweak the instructions
- Agents can be configured with prompt modification, model specification, a profile picture, and other things
- Admin-configured models will not be editable by individual users
- Add unit tests to verify agent behavior. Unit tests demonstrate imperfect adherence to prompt specifications

* Customize default behaviors for conversations without agents or with default agents

* Use agent_id for getting correct agent

* Merge migrations

* Simplify some variable definitions, add additional security checks for agents

* Rename agent.tuning -> agent.personality
2024-03-23 22:09:38 +05:30
Debanjum Singh Solanky
ecddf98430 Handle truncation when single long non-system chat message
Previously was assuming the system prompt is being always passed as
the first message. So expected there to be at least 2 messages in logs.

This broke chat actors querying with single long non system message.

A more robust way to extract system prompt is via the message role
instead
2024-03-15 15:58:39 +05:30
Debanjum Singh Solanky
6118d1ff57 Create chat actor for directly reading webpages based on user message
- Add prompt for the read webpages chat actor to extract, infer
  webpage links
- Make chat actor infer or extract webpage to read directly from user
  message
- Rename previous read_webpage function to more narrow
  read_webpage_at_url function
2024-03-14 14:58:37 +05:30
Debanjum Singh Solanky
dd883dc53a Dedupe query in notes prompt. Improve OAI chat actor, director tests
- Remove stale tests
- Improve tests to pass across gpt-3.5 and gpt-4-turbo
- The haiku creation director was failing because of duplicate query in
  instantiated prompt
2024-03-14 01:22:33 +05:30
Debanjum Singh Solanky
70b04d16c0 Test data source, output mode selector, web search query chat actors 2024-03-14 01:22:33 +05:30
Debanjum Singh Solanky
88f096977b Read webpages directly when Olostep proxy not setup
This is useful for self-hosted, individual user, low traffic setups
where a proxy service is not required
2024-03-11 18:41:02 +05:30
Debanjum Singh Solanky
ca2f962e95 Read, extract information from web pages in parallel to lower response time
- Time reading webpage, extract info from webpage steps for perf
  analysis
- Deduplicate webpages to read gathered across separate google
  searches
- Use aiohttp to make API requests non-blocking, pair with asyncio to
  parallelize all the online search webpage read and extract calls
2024-03-11 18:41:02 +05:30
sabaimran
e323a6d69b
Include additional user context in the image generation flow ()
* Make major improvements to the image generation flow

- Include user context from online references and personal notes for generating images
- Dynamically select the modality that the LLM should respond with
- Retun the inferred context in the query response for the dekstop, web chat views to read

* Add unit tests for retrieving response modes via LLM

* Move output mode unit tests to the actor suite, rather than director

* Only show the references button if there is at least one available

* Rename aget_relevant_modes to aget_relevant_output_modes

* Use a shared method for generating reference sections, simplify some of the prompting logic

* Make out of space errors in the desktop client more obvious
2024-03-06 13:48:41 +05:30
sabaimran
44f8f20ea7
Miscellaneous bugs and fixes for chat sessions ()
* Display given_name field only if it is not None

* Add default slugs in the migration script

* Ensure that updated_at is saved appropriately, make sure most recent chat is returned for default history

* Remove the bin button from the chat interface, given deletion is handled in the drop-down menus

* Refresh the side panel when a new chat is created

* Improveme tool retrieval prompt, don't let /online fail, and improve parsing of extract questions

* Fix ending chat response by offline chat on hitting a stop phrase

Previously the whole phrase wouldn't be in the same response chunk, so
chat response wouldn't stop on hitting a stop phrase

Now use a queue to keep track of last 3 chunks, and to stop responding
when hit a stop phrase

* Make chat on Obsidian backward compatible post chat session API updates

- Make chat on Obsidian get chat history from
  `responseJson.response.chat' when available (i.e when using new api)
- Else fallback to loading chat history from
  responseJson.response (i.e when using old api)

* Fix detecting success of indexing update in khoj.el

When khoj.el attempts to index on a Khoj server served behind an https
endpoint, the success reponse status contains plist with certs. This
doesn't mean the update failed.

Look for :errors key in status instead to determine if indexing API
call failed. This fixes detecting indexing API call success on the
Khoj Emacs client, even for Khoj servers running behind SSL/HTTPS

* Fix the mechanism for populating notes references in the conversation primer for both offline and online chat

* Return conversation.default when empty list for dynamic prompt selection, send all cmds in telemetry

* Fix making chat on Obsidian backward compatible post chat session API updates

New API always has conversation_id set, not `chat' which can be unset
when chat session is empty.

So use conversation_id to decide whether to get chat logs from
`responseJson.response.chat' or `responseJson.response' instead

---------

Co-authored-by: Debanjum Singh Solanky <debanjum@gmail.com>
2024-02-20 13:55:35 -08:00
sabaimran
a3eb17b7d4
Have Khoj dynamically select conversation command(s) in chat ()
* Have Khoj dynamically select which conversation command(s) are to be used in the chat flow
- Intercept the commands if in default mode, and have Khoj dynamically guess which tools would be the most relevant for answering the user's query
* Remove conditional for default to enter online search mode
* Add multiple-tool examples in the prompt, make prompt for tools more specific to info collection
2024-02-11 17:11:32 +05:30
Debanjum Singh Solanky
dd4cf66be1 Improve offline chat system prompt to think step by step 2024-02-06 20:23:19 +05:30
Debanjum Singh Solanky
035165b534 Make offline chat model current date aware. Improve system prompts
- Can now expect date awareness chat quality test to pass
- Prevent offline chat model from printing verbatim user Notes and
  special tokens
- Make it ask follow-up questions if it needs more context
2024-02-06 20:23:19 +05:30
Debanjum Singh Solanky
0d949140f4 Fix actor, director tests using freeze time by ignoring transformers package
transformers package was causing freeze time to fail during setup
2024-02-06 03:00:48 +05:30
sabaimran
b782683e60
Scrape results from Serper results using Olostep ()
* Initailize changes to incporate web scraping logic after getting SERP results
- Do some minor refactors to pass a symptom prompt to the openai model when making a query
- integrate Olostep in order to perform the webscraping
* Fix truncation error with new line, fix typing in olostep code
* Use the authorization header for the token
* Add a small hint/indicator for how to use Khojs other modalities in the welcome prompt
* Add more detailed error message if Olostep query fails
* Add unit tests which invoke Olostep in chat director
* Add test for olostep tool
2024-01-29 14:16:50 +05:30
Debanjum
4d30f7d1d9
Short-circuit API rate limiter for unauthenticated users ()
### Major
- Short-circuit API rate limiter for unauthenticated user
  Calls by unauthenticated users were failing at API rate limiter as it
  failed to access user info object. This is a bug.
  
  API rate limiter should short-circuit for unauthenicated users so a
  proper Forbidden response can be returned by API
  
  Add regression test to verify that unauthenticated users get 403
  response when calling the /chat API endpoint
  
### Minor
- Remove trailing slash to normalize khoj url in obsidian plugin settings
- Move used /api/config API controllers into separate module
- Delete unused /api/beta API endpoint
- Fix error message rendering in khoj.el, khoj obsidian chat
- Handle deprecation warnings for subscribe renew date, langchain, pydantic & logger.warn
2024-01-17 00:59:52 +05:30
Debanjum Singh Solanky
d26a4ffcea Only run the OpenAI chat client, /online test when API keys are set 2024-01-17 00:36:03 +05:30
Debanjum Singh Solanky
7039c202c8 Merge branch 'master' into short-circuit-api-rate-limiter 2024-01-16 18:18:34 +05:30
Debanjum Singh Solanky
6ded4c1d75 Merge branch 'master' into fix-1000-file-index-update-limit 2024-01-16 16:50:58 +05:30
Debanjum Singh Solanky
e0b381d523 Only run /online command offline chat director test when SERPER KEY present 2024-01-16 13:09:38 +05:30
Debanjum Singh Solanky
7dfbcd2e5a Handle subscribe renew date, langchain, pydantic & logger.warn warnings
- Ensure langchain less than 0.2.0 is used, to prevent breaking
  ChatOpenAI, PyMuPDF usage due to their deprecation after 0.2.0
- Set subscription renewal date to a timezone aware datetime
- Use logger.warning instead of logger.warn as latter is deprecated
- Use `model_dump' not deprecated dict to get all configured content_types
2024-01-12 01:46:52 +05:30
Debanjum Singh Solanky
ba99089a12 Short-circuit API rate limiter for unauthenticated user
Calls by unauthenticated users were failing at API rate limiter as it
failed to access user info object. This is a bug.

API rate limiter should short-circuit for unauthenicated users so a
proper Forbidden response can be returned by API

Add regression test to verify that unauthenticated users get 403
response when calling the /chat API endpoint
2024-01-12 00:23:50 +05:30
Debanjum Singh Solanky
4ded32cc64 Test 1000 file upload limit to index/update API endpoint
Due to FastAPI limitation
2024-01-03 22:14:36 +05:30
sabaimran
79913d4c17
Add isort to the pre-commit configuration and apply it to the whole project ()
* Apply isort to the entire repository
* Fix missing import issues in text_to_entries
* Fix imports in migration files
2023-12-28 18:04:02 +05:30
sabaimran
6dd2b05bf5 Rebase with master 2023-12-19 21:02:49 +05:30
sabaimran
ef21d78c99 Initial changes to support multiple search model configurations
- All search models are loaded into memory, and stored in a dictionary indexed by name
- Still need to add database migrations and create a UI for user to select their choice. Presently, it uses the default option
2023-12-05 00:35:40 -05:00
Debanjum Singh Solanky
2b09caa237 Make online results an optional argument to the gpt converse method 2023-12-04 12:15:29 -05:00
sabaimran
6e1ba11e59 Resolve merge conflicts for rendering chat response 2023-11-27 11:33:13 -08:00
sabaimran
e438853b09 Add additional unit tests to verify behavior of unsubscribed/subscribed users 2023-11-26 13:09:00 -08:00
Debanjum Singh Solanky
a0a7ab7ec8 Rename conversation.gpt4all package to conversation.offline 2023-11-26 04:19:32 -08:00
sabaimran
73e38fccf3 Explicitly set billing to off in the test for being able to index a large set of data 2023-11-25 20:48:32 -08:00
sabaimran
b2afbaa315 Add support for rate limiting the amount of data indexed
- Add a dependency on the indexer API endpoint that rounds up the amount of data indexed and uses that to determine whether the next set of data should be processed
- Delete any files that are being removed for adminstering the calculation
- Show current amount of data indexed in the config page
2023-11-25 20:28:04 -08:00
sabaimran
60c23d9e3a Add online search chat director tests 2023-11-21 23:08:36 -08:00
sabaimran
c652a7fd2d Move text_to_entries under the new content folder 2023-11-21 22:25:17 -08:00
sabaimran
1e2af083f0 Rename the data_sources module to content 2023-11-21 22:11:32 -08:00
sabaimran
2bb989e9d8 Resolve merge conflicts and fix some import ordering 2023-11-21 12:30:43 -08:00
sabaimran
a474c31e02 Move the django app into the src/khoj folder for better organization and functionality
- Our pypi package currently does not work because the django app and associated database is not included. To remedy this issue, move the app into the src/khoj folder. This has the added benefit of improved organization of the codebase, as all server related code is now in a single folder
- Update associated file paths and system references
2023-11-21 10:56:04 -08:00
sabaimran
b8e6883a81 Merge branch 'master' of github.com:khoj-ai/khoj into features/internet-enabled-search 2023-11-19 16:20:08 -08:00
sabaimran
4def8cce36
Merge pull request from asim-shrestha/patch-1
Add test separators
2023-11-19 14:14:34 -08:00
Debanjum
71799add0b
Index Parent Headings of Org-Mode Entries to Improve Search Context ()
### Overview
The parent hierarchy of org-mode entries can store important context. 
This change updates OrgNode to track parent headings for each org entry and adds the parent outline for each entry to the index

### Details
- Test search uses ancestor headings as context for improved results
- Add ancestor headings of each org-mode entry to their compiled form
- Track ancestor headings for each org-mode entry in org-node parser

Resolves 
2023-11-19 13:18:19 -08:00
sabaimran
e398a76779 Fix test word filter 2023-11-19 13:14:58 -08:00
sabaimran
33a9304428 Resolve merge conflicts 2023-11-19 12:57:55 -08:00
sabaimran
ef5e9d66c1 Resolve merge conflicts in dependency imports 2023-11-19 11:42:20 -08:00
sabaimran
f688529150 Update the default configuration for the AppConfig 2023-11-17 19:26:31 -08:00
Debanjum Singh Solanky
ca87b4ede9 Wrap common API query parameters into shared class to deduplicate code
- Upgrade FastAPI to >= latest version. Required upgrade of FastAPI.
  Earlier version didn't support wrapping common query params in class

- Use per fixture app instead of a global FastAPI app in conftest

- Upgrade minimum required Django version

- Fix no notes chat director test with updated no notes message
  No notes message was updated in commit 118f1143
2023-11-17 18:43:49 -08:00
Debanjum Singh Solanky
33ad9b8e64 Update text search test since indexing ancestor hierarchy added 2023-11-17 15:26:55 -08:00
Debanjum Singh Solanky
55785d50c3 Use title, when present, as root ancestor of entries instead of file path 2023-11-17 15:03:27 -08:00
sabaimran
ec06d2c446 Move data indexer files into a separate folder under processor. Update assoc UTs 2023-11-16 17:19:55 -08:00
Debanjum Singh Solanky
ddb07def0d Test search uses ancestor headings as context for improved results
- Update test data to add deeper outline hierarchy for testing
  hierarchy as context
- Update collateral tests that need count of entries updated, deleted
  asserts to be updated
2023-11-16 03:05:19 -08:00
Debanjum Singh Solanky
74403e3536 Add ancestor headings of each org-mode entry to their compiled form
Resolves 
2023-11-16 02:54:41 -08:00
Debanjum Singh Solanky
305c25ae1a Track ancestor headings for each org-mode entry in org-node parser 2023-11-16 02:39:14 -08:00
Debanjum Singh Solanky
922983bd53 Set max cos distance to 0.18. Test search API query with max distance 2023-11-15 20:26:21 -08:00
Debanjum Singh Solanky
08a057bdd5 Rename SearchModel to SearchModelConfig DB model, Require Cross-Encoder 2023-11-15 17:31:50 -08:00
Debanjum Singh Solanky
5a6ab9cc85 Fix failing client tests 2023-11-15 00:17:44 -08:00
Debanjum Singh Solanky
8f200cf53f Remove unused parameter from configure_search_type method 2023-11-14 19:09:35 -08:00
Debanjum Singh Solanky
4af194d74b Make search model configurable on server
- Expose ability to modify search model via Django admin interface
- Previously the bi_encoder and cross_encoder models to use were set
  in code
- Now it's user configurable but with a default config generated by
  default
2023-11-14 19:09:35 -08:00
Debanjum Singh Solanky
e98141f4c3 Subscribe default user to standard plan with a far away renewal date
Self hosted users in anonymous mode have all capabilities unlocked
2023-11-14 16:31:39 -08:00
Asim Shrestha
0bfc094e18
Add test separators 2023-11-11 17:08:58 -08:00
Debanjum Singh Solanky
c4364b9100 Weaken asking follow-up qs and q&a mode in notes prompt to OpenAI models
- Notes prompt doesn't need to be so tuned to question answering. User
could just want to talk about life. The notes need to be used to
response to those, not necessarily only retrieve answers from notes

- System and notes prompts were forcing asking follow-up questions a
  little too much. Reduce strength of follow-up question asking
2023-11-10 23:36:43 -08:00
sabaimran
e2e96f9aa4 Add default settings to let new users be subscribed on trial
- Add the default user to a subscription trial
- Update associated unit tests
2023-11-10 22:38:28 -08:00
Debanjum Singh Solanky
c9c0ba67c6 Fix chat_client configurations for OpenAI chat director tests 2023-11-10 17:29:23 -08:00
sabaimran
262a8574d1 Add a test to verify that a user without data sucessfully returns a respones to the /search endpoint 2023-11-10 14:00:58 -08:00
Debanjum Singh Solanky
404d47f1a1 Bubble up content indexing errors to notify user on client apps 2023-11-07 05:28:13 -08:00
Debanjum Singh Solanky
9ab327a2b6 Store the data source of each entry in database
This will be useful for updating, deleting entries by their data
source. Data source can be one of Computer, Github or Notion for now

Store each file/entries source in database
2023-11-07 02:18:48 -08:00
Debanjum Singh Solanky
a08b152358 Improve log messages in text_entries and memory leak unit test 2023-11-06 19:27:31 -08:00
Debanjum
38f24a037d
Improve Indexing Text Entries ()
Major
- Ensure search results logic consistent across migration to DB, multi-user
- Manually verified search results for sample queries look the same across migration
 - Flatten indexing code for better indexing progress tracking and code readability

Minor
- a4f407f Test memory leak on MPS device when generating vector embeddings
- ef24485 Improve Khoj with DB setup instructions in the Django app readme (for now)
- f212cc7 Arrange remaining text search tests in arrange, act, assert order
- 022017d Fix text search tests to test updated indexing log messages
2023-11-06 16:01:53 -08:00
sabaimran
3d6e8d53fe Try adding dependencies for libgl in order to run OCR in github action unit tests 2023-11-05 15:09:40 -08:00
sabaimran
fdd727712f Rename test files from x_to_jsonl to x_to_entries 2023-11-05 14:33:07 -08:00
Debanjum Singh Solanky
a4f407f595 Test memory leak on MPS device when generating vector embeddings
Slope threshold of 2.0 determined qualitatively on local Mac device
Minor unused import and clean-up
2023-11-05 03:48:54 -08:00
Debanjum Singh Solanky
f212cc7174 Arrange remaining text search tests in arrange, act, assert order 2023-11-05 02:04:52 -08:00
Debanjum Singh Solanky
022017dd0f Fix text search tests to test updated indexing log messages 2023-11-05 02:04:52 -08:00
sabaimran
d1d210605e Merge branch 'features/multi-user-support-khoj' of github.com:khoj-ai/khoj into features/multi-user-support-khoj 2023-11-04 14:29:34 -07:00
sabaimran
3678aa5614 Add tests to validate expected behaviors in the multi-user scenario 2023-11-04 14:29:30 -07:00
Debanjum Singh Solanky
2f1756cc15 Do not use icon for each file, folder to index in desktop app.
Other minor fixes based on PR feedback
2023-11-04 00:13:10 -07:00
Debanjum Singh Solanky
345856e7be Merge branch 'master' of github.com:khoj-ai/khoj into features/multi-user-support-khoj
Merge changes to use latest GPT4All with GPU, GGUF model support into
khoj multi-user support rearchitecture branch
2023-11-02 22:44:25 -07:00
sabaimran
fe6720fa06
[Multi-User Part 8]: Make conversation processor settings server-wide ()
- Rather than having each individual user configure their conversation settings, allow the server admin to configure the OpenAI API key or offline model once, and let all the users re-use that code.
- To configure the settings, the admin should go to the `django/admin` page and configure the relevant chat settings. To create an admin, run `python3 src/manage.py createsuperuser` and enter in the details. For simplicity, the email and username should match.
- Remove deprecated/unnecessary endpoints and views for configuring per-user chat settings
2023-11-02 10:43:27 -07:00
Debanjum Singh Solanky
d92a2d03a7 Rename Files, Classes from X_To_JSONL to more appropriate X_To_Entries
These content processors are converting content into entries in DB
instead of entries in JSONL file
2023-11-01 14:51:33 -07:00
Debanjum Singh Solanky
87e6b1eab9 Rename TextEmbeddings to TextEntries for improved readability
Improves readability as name has closer match to underlying
constructs
2023-10-31 18:55:59 -07:00
Debanjum Singh Solanky
bcbee05a9e Rename DbModels Embeddings, EmbeddingsAdapter to Entry, EntryAdapter
Improves readability as name has closer match to underlying
constructs

- Entry is any atomic item indexed by Khoj. This can be an org-mode
  entry, a markdown section, a PDF or Notion page etc.

- Embeddings are semantic vectors generated by the search ML model
  that encodes for meaning contained in an entries text.

- An "Entry" contains "Embeddings" vectors but also other metadata
  about the entry like filename etc.
2023-10-31 18:50:54 -07:00
sabaimran
54a387326c
[Multi-User Part 6]: Address small bugs and upstream PR comments ()
- 08654163cb: Add better parsing for XML files
- f3acfac7fb: Add a try/catch around the dateparser in order to avoid internal server errors in app
- 7d43cd62c0: Chunk embeddings generation in order to avoid large memory load
- e02d751eb3: Addresses comments from PR  
- a3f393edb4: Addresses comments from PR  
- 66eb078286: Addresses comments from PR  
- Address various items in https://github.com/khoj-ai/khoj/issues/527
2023-10-31 17:59:53 -07:00
sabaimran
5f3f6b7c61
[Multi-User Part 5]: Add a production Docker file and use a gunicorn configuration with it ()
- Add a productionized setup for the Khoj server using `gunicorn` with multiple workers for handling requests
- Add a new Dockerfile meant for production config at `ghcr.io/khoj-ai/khoj:prod`; the existing Docker config should remain the same
2023-10-26 13:15:31 -07:00
Debanjum
9acc722f7f
[Multi-User Part 4]: Authenticate using API Tokens ()
###  New
- Use API keys to authenticate from Desktop, Obsidian, Emacs clients
- Create API, UI on web app config page to CRUD API Keys
- Create user API keys table and functions to CRUD them in Database

### 🧪 Improve
- Default to better search model, [gte-small](https://huggingface.co/thenlper/gte-small), to improve search quality
- Only load chat model to GPU if enough space, throw error on load failure
- Show encoding progress, truncate headings to max chars supported
- Add instruction to create db in Django DB setup Readme

### ⚙️ Fix
- Fix error handling when configure offline chat via Web UI
- Do not warn in anon mode about Google OAuth env vars not being set
- Fix path to load static files when server started from project root
2023-10-26 12:33:03 -07:00
sabaimran
4b6ec248a6
[Multi-User Part 3]: Separate chat sesssions based on authenticated users ()
- Add a data model which allows us to store Conversations with users. This does a minimal lift over the current setup, where the underlying data is stored in a JSON file. This maintains parity with that configuration.
- There does _seem_ to be some regression in chat quality, which is most likely attributable to search results.

This will help us with . It should become much easier to maintain multiple Conversations in a given table in the backend now. We will have to do some thinking on the UI.
2023-10-26 11:37:41 -07:00
sabaimran
a8a82d274a
[Multi-User Part 2]: Add login pages and gate access to application behind login wall ()
- Make most routes conditional on authentication *if anonymous mode is not enabled*. If anonymous mode is enabled, it scaffolds a default user and uses that for all application interactions.
- Add a basic login page and add routes for redirecting the user if logged in
2023-10-26 10:17:29 -07:00
sabaimran
216acf545f
[Multi-User Part 1]: Enable storage of settings for plaintext files based on user account ()
- Partition configuration for indexing local data based on user accounts
- Store indexed data in an underlying postgres db using the `pgvector` extension
- Add migrations for all relevant user data and embeddings generation. Very little performance optimization has been done for the lookup time
- Apply filters using SQL queries
- Start removing many server-level configuration settings
- Configure GitHub test actions to run during any PR. Update the test action to run in a containerized environment with a DB.
- Update the Docker image and docker-compose.yml to work with the new application design
2023-10-26 09:42:29 -07:00
Debanjum Singh Solanky
0f1ebcae18 Upgrade to latest GPT4All. Use Mistral as default offline chat model
GPT4all now supports gguf llama.cpp chat models. Latest
GPT4All (+mistral) performs much at least 3x faster.

On Macbook Pro at ~10s response start time vs 30s-120s earlier.
Mistral is also a better chat model, although it hallucinates more
than llama-2
2023-10-22 19:04:23 -07:00
sabaimran
963cd165eb Resolve merge conflicts 2023-10-19 14:39:05 -07:00
Debanjum Singh Solanky
e3cd8b4150 Only index files returned by input-filter globs in fs_syncer
Ignore .org, .pdf etc. suffixed directories under `input-filter' from
being evaluated as files.

Explicitly filter results by input-filter globs to only index files,
not directory for each text type

Add test to prevent regression

Closes 
2023-10-17 23:32:10 -07:00
Debanjum Singh Solanky
d9d133dfb9 Read text files as utf-8, instead of default os locale
On Windows, the default locale isn't utf8. Khoj had regressed to
reading files in OS specified locale encoding, e.g cp1252, cp949 etc.

It now explicitly uses utf8 encoding to read text files for indexing

Resolves , resolves 
2023-10-17 21:47:19 -07:00
Debanjum
ecc6fbfeb2
Push Files to Index from Emacs, Obsidian & Desktop Clients using Multi-Part Forms ()
### Overview
- Add ability to push data to index from the Emacs, Obsidian client
- Switch to standard mechanism of syncing files via HTTP multi-part/form. Previously we were streaming the data as JSON
  - Benefits of new mechanism
    - No manual parsing of files to send or receive on clients or server is required as most have in-built mechanisms to send multi-part/form requests
    - The whole response is not required to be kept in memory to parse content as JSON. As individual files arrive they're automatically pushed to disk to conserve memory if required
    - Binary files don't need to be encoded on client and decoded on server

### Code Details
### Major
- Use multi-part form to receive files to index on server
- Use multi-part form to send files to index on desktop client
- Send files to index on server from the khoj.el emacs client
  - Send content for indexing on server at a regular interval from khoj.el
- Send files to index on server from the khoj obsidian client
- Update tests to test multi-part/form method of pushing files to index

#### Minor
- Put indexer API endpoint under /api path segment
- Explicitly make GET request to /config/data from khoj.el:khoj-server-configure method
- Improve emoji, message on content index updated via logger
- Don't call khoj server on khoj.el load, only once khoj invoked explicitly by user
- Improve indexing of binary files
  - Let fs_syncer pass PDF files directly as binary before indexing
  - Use encoding of each file set in indexer request to read file 
- Add CORS policy to khoj server. Allow requests from khoj apps, obsidian & localhost
- Update indexer API endpoint URL to` index/update` from `indexer/batch`

Resolves  
2023-10-17 06:05:15 -07:00
Debanjum Singh Solanky
7b1c62ba53 Mark test_get_configured_types_via_api unit test as flaky
It passes locally on running individually but fails when run in
parallel on local or CI
2023-10-17 05:56:00 -07:00
Debanjum Singh Solanky
5efae1ad55 Update indexer API endpoint query params for force, content type
New URL query params, `force' and `t' match name of query parameter in
existing Khoj API endpoints

Update Desktop, Obsidian and Emacs client to call using these new API
query params. Set `client' query param from each client for telemetry
visibility
2023-10-17 04:58:13 -07:00
Debanjum Singh Solanky
84654ffc5d Update indexer API endpoint URL to index/update from indexer/batch
New URL follows action oriented endpoint naming convention used for
other Khoj API endpoints

Update desktop, obsidian and emacs client to call this new API
endpoint
2023-10-17 04:58:13 -07:00
Debanjum Singh Solanky
541cd59a49 Let fs_syncer pass PDF files directly as binary before indexing
No need to do unneeded base64 encoding/decoding to pass pdf contents
for indexing from fs_syncer to pdf_to_jsonl
2023-10-17 04:58:13 -07:00
Debanjum Singh Solanky
1a9023d396 Update Chat Actor test to not incept with prior world knowledge 2023-10-15 17:22:44 -07:00
Debanjum Singh Solanky
feb4f17e3d Update chat config schema. Make max_prompt, chat tokenizer configurable
This provides flexibility to use non 1st party supported chat models

- Create migration script to update khoj.yml config
  - Put `enable_offline_chat' under new `offline-chat' section
    Referring code needs to be updated to accomodate this change
  - Move `offline_chat_model' to `chat-model' under new `offline-chat' section
  - Put chat `tokenizer` under new `offline-chat' section
  - Put `max_prompt' under existing `conversation' section
    As `max_prompt' size effects both openai and offline chat models
2023-10-15 16:35:11 -07:00
sabaimran
c125995d94
[Multi-User]: Part 0 - Add support for logging in with Google ()
* Add concept of user authentication to the request session via GoogleUser
2023-10-14 19:39:13 -07:00
Debanjum Singh Solanky
56bd69d5af Improve Llama v2 extract questions actor and associated prompt
- Format extract questions prompt format with newlines and whitespaces
- Make llama v2 extract questions prompt consistent

- Remove empty questions extracted by offline extract_questions actor
- Update implicit qs extraction unit test for offline search actor
2023-10-13 22:48:56 -07:00
Debanjum Singh Solanky
bed3aff059 Update tests to test multi-part/form method of pushing files to index
Instead of using the previous method to push data as json payload of POST request
pass it as files to upload via the multi-part/form to the batch indexer API endpoint
2023-10-12 20:58:52 -07:00
Debanjum Singh Solanky
6aa69da3ef Put indexer API endpoint under /api path segment
Update FastAPI app router, desktop app and to use new url path to
batch indexer API endpoint

All api endpoints should exist under /api path segment
2023-10-09 21:35:58 -07:00
Debanjum Singh Solanky
13b16a4364 Use default Llama 2 supported by GPT4All
Remove custom logic to download custom Llama 2 model.
This was added as GPT4All didn't support Llama 2 when it was added to Khoj
2023-10-03 19:01:54 -07:00
sabaimran
96a9fa07f0 Fix conf test setup for offline chat 2023-09-18 15:05:15 -07:00
sabaimran
2dd15e9f63
Resolve issues with GPT4All and fix prompt for yesterday extract questions date filter ()
- GPT4All integration had ceased working with 0.1.7 specification. Update to use 1.0.12. At a later date, we should also use first party support for llama v2 via gpt4all
- Update the system prompt for the extract_questions flow to add start and end date to the yesterday date filter example.
- Update all setup data in conftest.py to use new client-server indexing pattern
2023-09-18 14:41:26 -07:00
sabaimran
8141be97f6 Update date filter test to use compiled rather than raw key 2023-09-18 11:24:56 -07:00
sabaimran
343854752c
Improve docker builds for local hosting ()
* Remove GPT4All dependency in pyproject.toml and use multiplatform builds in the dockerization setup in GH actions
* Move configure_search method into indexer
* Add conditional installation for gpt4all
* Add hint to go to localhost:42110 in the docs. Addresses 
2023-09-08 17:07:26 -07:00
sabaimran
dccfae3853
Remove PySide dependency and deprecate desktop builds ()
* Remove PySide, gui option from code
* Remove pyside 6 dependency from code
* Remove workflows which build desktop applications
* Update unit tests and update line in documentation
* Remove additional references to pyinstaller, gui
* Add uninstall steps to normal uninstall instructions
2023-09-07 11:36:27 -07:00
sabaimran
76562f4250
Add front-end Electron application for Khoj local file syncing ()
* Initial version - setup a file-push architecture for generating embeddings with Khoj
* Use state.host and state.port for configuring the URL for the indexer
* Fix parsing of PDF files
* Read markdown files from streamed data and update unit tests
* On application startup, load in embeddings from configurations files, rather than regenerating the corpus based on file system
* Init: refactor indexer/batch endpoint to support a generic file ingestion format
* Add features to better support indexing from files sent by the desktop client
* Initial commit with Electron application
- Adds electron app
* Add import for pymupdf, remove import for pypdf
* Allow user to configure khoj host URL
* Remove search type configuration from index.html
* Use v1 path for current indexer routes
2023-09-06 12:04:18 -07:00
sabaimran
4854258047
Move to a push-first model for retrieving embeddings from local files ()
* Initial version - setup a file-push architecture for generating embeddings with Khoj
* Update unit tests to fix with new application design
* Allow configure server to be called without regenerating the index; this no longer works because the API for indexing files is not up in time for the server to send a request
* Use state.host and state.port for configuring the URL for the indexer
* On application startup, load in embeddings from configurations files, rather than regenerating the corpus based on file system
2023-08-31 12:55:17 -07:00
Debanjum Singh Solanky
01b310635e Enable passing search query filters via chat and test it 2023-08-28 09:24:32 -07:00
Debanjum Singh Solanky
794bad8bcb Make date_filter.extract_date_range method always return a list type 2023-08-28 00:55:28 -07:00
Debanjum Singh Solanky
d5a2de6222 Add method to extract filter terms from query to all filters
- Test the get_filter_term method in all 3 word, file, date filters
- Make the existing can_filter method by default in base filter abstract class
2023-08-28 00:55:28 -07:00
Debanjum Singh Solanky
319f066aec Test general and notes slash commands in offline chat director tests 2023-08-28 00:47:02 -07:00
Debanjum Singh Solanky
edffbad837 Make Khoj ask clarifying questions when answer not in provided context
Previously it would just refuse ask for clarification. This improves
the chat quality score for the existing director tests
2023-08-28 00:47:02 -07:00
sabaimran
b45e1d8c0d
Fix plaintext HTML parsing and rendering ()
* Store conversation command options in an Enum
* Move to slash commands instead of using @ to specify general commands
* Calculate conversation command once & pass it as arg to child funcs
* Add /notes command to respond using only knowledge base as context
This prevents the chat model to try respond using it's general world
knowledge only without any references pulled from the indexed
knowledge base
* Test general and notes slash commands in openai chat director tests
---------

Co-authored-by: Debanjum Singh Solanky <debanjum@gmail.com>
2023-08-27 11:24:30 -07:00
Debanjum
7919787fb7
Use Slash Commands and Add Notes Slash Command ()
* Store conversation command options in an Enum

* Move to slash commands instead of using @ to specify general commands

* Calculate conversation command once & pass it as arg to child funcs

* Add /notes command to respond using only knowledge base as context

This prevents the chat model to try respond using it's general world
knowledge only without any references pulled from the indexed
knowledge base

* Test general and notes slash commands in openai chat director tests

* Update gpt4all tests to use md configuration

* Add a /help tooltip

* Add dynamic support for describing slash commands. Remove default and treat notes as the default type

---------

Co-authored-by: sabaimran <narmiabas@gmail.com>
2023-08-26 18:11:18 -07:00
sabaimran
0ea901c7c1
Allow indexing to continue even if there's an issue parsing a particular org file ()
* Allow indexing to continue even if there's an issue parsing a particular org file
* Use approximation in pytorch comparison in text_search UT, skip additional file parser errors for org files
* Change error of expected failure
2023-08-14 07:56:33 -07:00
sabaimran
7b907add77
Add support for indexing plaintext files ()
* Add support for indexing plaintext files
- Adds backend support for parsing plaintext files generically (.html, .txt, .xml, .csv, .md)
- Add equivalent frontend views for setting up plaintext file indexing
- Update config, rawconfig, default config, search API, setup endpoints
* Add a nifty plaintext file icon to configure plaintext files in the Web UI
* Use generic glob path for plaintext files. Skip indexing files that aren't in whitelist
2023-08-09 15:44:40 -07:00
Debanjum Singh Solanky
95acb1583d Update local Chat Actor and Director tests expected to fail 2023-08-01 20:52:00 -07:00
Debanjum Singh Solanky
c2b7a14ed5 Fix context, response size for Llama 2 to stay within max token limits
Create regression text to ensure it does not throw the prompt size
exceeded context window error
2023-08-01 20:52:00 -07:00
sabaimran
d8fa967b43 Update chat actor unit tests for greater accuracy and benchmarking 2023-08-01 12:24:43 -07:00
sabaimran
48363ec861 Add additional check for chat_messages length in UT 2023-08-01 09:25:52 -07:00
sabaimran
90efc2ea7a Update comments and add explanations 2023-08-01 09:24:03 -07:00
sabaimran
e55e9a7b67 Fix unit tests and truncation logic 2023-07-31 21:37:59 -07:00
sabaimran
95c7b07c20 Make the fake message longer 2023-07-31 20:55:19 -07:00
sabaimran
8dd5756ce9 Add new director tests for the offline chat model with llama v2 2023-07-31 20:24:52 -07:00
sabaimran
5ccb01343e
Add Offline chat to Obsidian ()
* Add support for configuring/using offline chat from within Obsidian
* Fix type checking for search type
* If Github is not configured, /update call should fail
* Fix regenerate tests same as the update ones
* Update help text for offline chat in obsidian
* Update relevant description for Khoj settings in Obsidian
* Simplify configuration logic and use smarter defaults
2023-07-28 18:47:56 -07:00
Debanjum Singh Solanky
9b1048caf7 Remove asymmetric from name of remaining text search tests
Asymmetric search is the only search type used now in khoj.el. So
making distinction between between symmetric and asymmetric search
isn't necessary anymore
2023-07-28 15:33:22 -07:00
sabaimran
124d97c26d
Replace Falcon 🦅 model with Llama V2 🦙 for offline chat ()
* Working example with LlamaV2 running locally on my machine

- Download from huggingface
- Plug in to GPT4All
- Update prompts to fit the llama format

* Add appropriate prompts for extracting questions based on a query based on llama format

* Rename Falcon to Llama and make some improvements to the extract_questions flow

* Do further tuning to extract question prompts and unit tests

* Disable extracting questions dynamically from Llama, as results are still unreliable
2023-07-27 20:51:20 -07:00
Debanjum Singh Solanky
da3f4dc7e4 Fix test config to run OpenAI Chat Actor, Director tests
OpenAI conversation processor schema had updated but conftest hadn't
been updated to reflect the same.

Update conftest setup of conversation processor to fix this
2023-07-27 11:30:04 -07:00
sabaimran
8b2af0b5ef
Add support for our first Local LLM 🤖🏠 ()
* Add support for gpt4all's falcon model as an additional conversation processor
- Update the UI pages to allow the user to point to the new endpoints for GPT
- Update the internal schemas to support both GPT4 models and OpenAI
- Add unit tests benchmarking some of the Falcon performance
* Add exc_info to include stack trace in error logs for text processors
* Pull shared functions into utils.py to be used across gpt4 and gpt
* Add migration for new processor conversation schema
* Skip GPT4All actor tests due to typing issues
* Fix Obsidian processor configuration in auto-configure flow
* Rename enable_local_llm to enable_offline_chat
2023-07-26 16:27:08 -07:00
Debanjum Singh Solanky
5bb42e56a8 Fix formatting of khoj test config and unused references in conftests 2023-07-22 00:29:26 -07:00
Debanjum Singh Solanky
d078e7b1f6 Clean up search type usage in khoj server, tests and Readme 2023-07-18 19:57:55 -07:00
Debanjum Singh Solanky
ef6a0044f4 Drop embeddings of deleted text entries from index
Previously the deleted embeddings would continue to be in the index,
even after the entry was deleted
2023-07-16 03:47:05 -07:00
Debanjum Singh Solanky
c73feebf25 Test index embeddings are stable on incremental update & no norm
Ensure order of new embedding insertion on incremental update
does not affect the order and value of existing embeddings when
normalization is turned off
2023-07-16 02:22:28 -07:00
Debanjum Singh Solanky
1482fd4d4d Test index is stable sorted on incremental update with new entry
Ensure order of new embedding, entry insertion on incremental update
is stable
2023-07-16 01:45:53 -07:00
Debanjum Singh Solanky
b02323ade6 Improve name of text search test functions
Asymmetric was older name used to differentiate between symmetric,
asymmetric search.

Now that text search just uses asymmetric search stick to simpler name
2023-07-16 01:45:53 -07:00
Debanjum Singh Solanky
7669b85da6 Test index is stable sorted on regenerate with new entry 2023-07-16 01:45:53 -07:00
Debanjum Singh Solanky
6e70b914c2 Remove unused dump_jsonl method
The entries index is stored ingzipped jsonl files for each content type
2023-07-16 01:45:53 -07:00
Debanjum Singh Solanky
88d1a29a84 Test index is stable for duplicate entries across regenerate, update
- Current incorrect behavior:
  All entries with duplicate compiled form are kept on regenerate
  but on update only the last of the duplicated entries is kept

This divergent behavior is not ideal to prevent index corruption
across reconfigure and update
2023-07-16 01:45:53 -07:00
Debanjum Singh Solanky
da98b92dd4 Create helper function to test value, order of entries & embeddings
This helper should be used to observe if the current embeddings are
stable sorted on regenerate and incremental update of index in text
search tests
2023-07-16 01:45:53 -07:00
Debanjum Singh Solanky
b9fb656657 Update Tests to setup both content_index, search_models before testing
This is required by the updated structure of Khoj setup

- Add content_config pytest fixture, pass bi_encoder from
  search_models.[text|image]_search
2023-07-14 01:29:48 -07:00
Debanjum Singh Solanky
f664a74e77 Update Khoj server to run on non standard port, 42110 instead of 8000
Resolves 
2023-07-10 21:27:58 -07:00
sabaimran
4c135ea316
Make streaming optional for the /chat endpoint ()
* Update the /chat endpoint to conditionally support streaming

- If streams are enabled, return the threadgenerator as it does currently
- If stream is disabled, return a JSON response with the response/compiled references separated out
- Correspondingly, update the chat.html UI to use the streamed API, as well as Obsidian
- Rename chat/init/ to chat/history

* Update khoj.el to use the /history endpoint

- Update corresponding unit tests to use stream=true

* Remove & from call to /chat for obsidian

* Abstract functions out into a helpers.py file and clean up some of the error-catching
2023-07-09 10:12:09 -07:00
Debanjum Singh Solanky
171ce19e1f Update date filter to allow quoting values in single quotes 2023-07-07 17:13:47 -07:00