Add multiple different locations as training data

[ogamingSCV]

I think it would be cool to be able to add multiple locations as a variable for the training data. This way I would be able to just add paths to my stored documents, notes etc. Maybe even a remote source, like from a NAS serve.

With this it might be possible to start a new ingrest every night when files have changed and I am always able to access my newest documents.

Also good idea is to skip fails during ingrest and only showing a list of failed files at the end (already a pull request available).

[sime2408]

You gave me great ideas :). I am planning to create a feature for Chroma over IP and for documents NAS as you said with CRON schedulers to pick new ones up. For now, in my fork, I created sub-directories and Chroma collections to granulate or separate multiple databases. So you have max 2 levels of sources: the parent name of the database and the sub-category stored as Chroma collection/tag. Examples of database psychology:

psychology
-  parenting
-  ...