Commit graph

12 commits

Author SHA1 Message Date
Icereed
caa27c4d53
Support reasoning models (deepseek-r1) in Ollama () 2025-02-05 19:59:08 +00:00
Icereed
90ca87a207
feat: add TOKEN_LIMIT environment variable for controlling maximum to… ()
* feat: add TOKEN_LIMIT environment variable for controlling maximum tokens in prompts

---------

Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
2025-02-02 15:14:18 +00:00
Icereed
32cc3d2794
Add documentation for prompt templates and getSuggestedTitle function () 2025-01-27 09:33:12 +01:00
Ingmar Stein
8b6041a93f
Make the existing document title available in prompt template. ()
This allow a prompt to feed the LLM with the title as determined by
Paperless to improve its suggestion.

This also fixes the case when `getSuggestedTags` or
`getSuggestedCorrespondent` are called, but `GenerateTitles` is off (in
which case an empty title was used).

Fixes 
2025-01-27 09:10:53 +01:00
Icereed
4a71a42ef3
feat: remove paperless-gpt related tags from available tags () 2025-01-13 16:22:54 +00:00
Icereed
e661989829
LLM powered Correspondents handling ()
---------

Co-authored-by: Jonas Hess <Jonas@Hess.pm>
2025-01-13 15:59:29 +01:00
Icereed
e144661dfb
Add RemoveTags field to DocumentSuggestion for explicit tag removal ()
* feat: add functionality to manage suggested and removable tags in document suggestions

* feat: process multiple documents for auto-tagging and OCR
2025-01-13 09:52:56 +00:00
Icereed
95626371d7
Enhance logging ()
* enhance logging

* refactor: enhance logging in document suggestion and title generation processes
2025-01-13 08:31:51 +01:00
Icereed
38321c4ac9
Limit pages for OCR ()
Bonus: Adds logging for image dimensions and size

Closes 
2025-01-10 17:03:53 +01:00
ccrlawrence
5b3373743a
Fix Vision OCR for OpenAI ()
## Summary by CodeRabbit

- **New Features**
	- Updated environment variable descriptions for improved clarity on OCR processing options.
	- Enhanced the `doOCRViaLLM` method to support different LLM providers and improve image data handling.

- **Bug Fixes**
	- Standardized error handling for better reporting across multiple methods. 

- **Documentation**
	- Revised Docker Compose section in `README.md` to reflect updated environment variable options.
2024-11-13 05:47:25 +01:00
Jonas Hess
5dadbcb53d feat: improve auto throughput & logging 2024-10-31 21:17:16 +01:00
Icereed
03364f2741
OCR via LLM () 2024-10-28 17:34:41 +01:00