Description

What is this PR

• Bug fix
• Addition of a new feature
• Other

Why is this PR needed?

Loading data with load_dataset() currently requires users to pass source_software explicitly, even when the file format is distinctive enough to infer it automatically. This adds boilerplate to common workflows and makes the API a bit less convenient for users who just want to load a supported file.

At the same time, some DLC-style CSV files can match both DeepLabCut and LightningPose, so inference should be helpful without silently guessing in ambiguous cases.

What does this PR do?

• adds infer_source_software(file) to infer the source software from the input file
• adds support for source_software="auto" in load_dataset()
• makes load_dataset() default to automatic source inference
• uses registered validators to infer supported formats
• handles the DeepLabCut vs LightningPose CSV overlap with a small heuristic
• raises a clear ValueError when a DLC-style CSV is genuinely ambiguous instead of silently defaulting
• exports infer_source_software from movement.io
• adds unit tests covering supported formats and an ambiguous DLC-style CSV case
• updates the user guide to document automatic inference and ambiguous CSV behavior

References

Closes #919

How has this PR been tested?

This PR was tested locally with:

pytest tests/test_unit/test_io/test_load.py

This covers:

• direct testing of infer_source_software()
• loading via load_dataset(..., source_software="auto")
• supported formats including DeepLabCut, LightningPose, SLEAP, Anipose, VIA-tracks, and NWB
• the ambiguous DLC-style CSV case, which now raises a clear error

I also ran the repository pre-commit checks as part of committing the changes.

Is this a breaking change?

• Yes
• No

Existing explicit uses of load_dataset(..., source_software=...) are unchanged. This PR adds automatic inference as a convenience feature. In ambiguous DLC-style CSV cases, inference now raises a clear error instead of silently guessing, but that behavior only applies to the new auto-inference path.

Does this PR require an update to the documentation?

• Yes
• No

The user guide has been updated to:

• describe automatic source software inference
• show usage with source_software="auto"
• explain that ambiguous DLC-style CSV files may still require an explicit source_software

Checklist:

• The code has been tested locally
• Tests have been added to cover all new functionality
• The documentation has been updated to reflect any changes
• The code has been formatted with pre-commit