Maybe we should do a review of all locations where we filter pipelines/training configs from VIAME?

Where pipelines are allowed/disallowed using RegEx within the Web version: https://github.com/Kitware/dive/blob/main/server/dive_tasks/pipeline_discovery.py#L20-L34

Where we get user-trained models, we filter by extension type here on the desktop:
https://github.com/Kitware/dive/blob/main/client/platform/desktop/backend/native/common.ts#L382

On the web for user trained models it looks like we are looking for [detector|tracker].pipe in their trained confg folders:
https://github.com/Kitware/dive/blob/main/server/dive_server/crud_rpc.py#L65
More specifically: https://github.com/Kitware/dive/blob/main/server/dive_server/crud_rpc.py#L98-105

Training Configs on the desktop web are specified here:
https://github.com/Kitware/dive/blob/main/client/platform/desktop/backend/native/common.ts#L453-L455

Training Configs on the web just seem to be filtered by '*.conf":
https://github.com/Kitware/dive/blob/main/server/dive_tasks/pipeline_discovery.py#L68

For Fine-Tuned Training on Desktop:
https://github.com/Kitware/dive/blob/main/client/platform/desktop/backend/native/common.ts#L455

This is looking inside of the user trained model folders for items for fine tune training:
https://github.com/Kitware/dive/blob/main/server/dive_server/crud_rpc.py#L128
These models are also added during the static training config import after using the CSV downloader for other models: https://github.com/Kitware/dive/blob/main/server/dive_tasks/pipeline_discovery.py#L97-L100
Training Extension matching uses this source:
https://github.com/Kitware/dive/blob/main/server/dive_utils/constants.py#L142