Version upped

sqlite: avoid warnings on too short vectors

This should not happen in the real world. But when embeddings were
created with a shorter vector model than the model that is used to
embed the query, the

sqlite: avoid warnings on too short vectors

This should not happen in the real world. But when embeddings were
created with a shorter vector model than the model that is used to
embed the query, the cosineSimilarity method threw a whole bunch of
warnings. We now stop the comparison at the vector length.

In the real world the same model for embeddings and the query should be
used, results are unpredictable otherwise. So this is mostly a cosmetic
change for messed up states during development.

show more ...

added "similar" endpoint to the remote api

add remote component to ask questions to the bot

The endpoint allows to override model and language settings on demand.

Version upped

Adjust OpenAI API calls. fixes #34

* max_token has been replaced by max_completion_tokens
* temperature setting is not suported by reasoning models

Version upped

ollama: remove thinking part from deepseek answers

ollama: fix embed endpoint and error in model file

We expect /api to be part of the configured base URL so it should not
be part of the call itself.

Update EmbeddingModel.php to remove the :lastest suffix

Update models.json with new llama3.3 70b

Update EmbeddingModel.php for new Ollama API

Update models.json

Remove deprecated Groq models

Version upped

some model updates

still not 100% uptodate

allow to set arbitrary models

We now initialize a model configuration even if we have no info in
model.json using some default values for the token limits.

Models can implement the loadUnkonwModelI

allow to set arbitrary models

We now initialize a model configuration even if we have no info in
model.json using some default values for the token limits.

Models can implement the loadUnkonwModelInfo() method to fetch the info
from an API if such a thing exist. Implemented for gemini and ollama
currently.

show more ...

actually use the configured rephrase model

stupid copy n paste error

translation update

Version upped

translatable default message

When no message is given in the chat syntax, the default is now taken
from the language file.

Version upped

mark some more settings that need reindexing

added bge-large embedding model for ollama

syntax component to ignore parts of a page while indexing