Add datatrove tag to synthetic dataset cards

[JoelNiklaus]

Problem

Datasets generated with DataTrove have no way to be discovered or filtered by the datatrove tag on the Hugging Face Hub.

Solution

Add "datatrove" as a permanent tag in the dataset card generator's tag set, alongside the existing "synthetic" tag. Every new synthetic dataset card will now include - datatrove in its YAML frontmatter tags.

Testing

Existing tests pass — none assert on specific tag values in the rendered card.

Made with Cursor

---

Note

Low Risk
Low risk: adds a single constant tag to the dataset card YAML frontmatter, with no changes to inference, uploads, or data processing logic.

Overview
Synthetic dataset cards generated by InferenceDatasetCardGenerator now always include the datatrove tag (in addition to synthetic) in the README YAML frontmatter, improving discoverability/filtering on the Hugging Face Hub.

^{Written by Cursor Bugbot for commit 5fd2d3f. Configure here.}