Skip to content

feat: add the Cuybernautas/airaware connector for Lima PM2.5 forecasting - #796

Draft
keneth-arias wants to merge 3 commits into
OpenTSLM:mainfrom
keneth-arias:feat/airaware-senamhi-lima
Draft

keneth-arias wants to merge 3 commits into
OpenTSLM:mainfrom
keneth-arias:feat/airaware-senamhi-lima

Conversation

@keneth-arias

Copy link
Copy Markdown

Dataset

Add airaware/lima-pm25-forecast@0.1.0 for retrospective hourly PM2.5 forecasting at seven provisional station identities in Lima, Peru. Each task uses 24 measured hours to predict the next six. The connector pins the original SENAMHI CSV by SHA-256, consolidates equal repeats, removes ambiguous hourly values, and purges windows that cross the fixed chronological split boundaries.

The full release contains 237,397 training, 50,706 validation, and 50,789 test tasks. The source database remains ODC-By-1.0 with SENAMHI attribution. The newly written connector follows TimeNet's MIT license.

Verification

  • Built the full release on the pinned TimeNet revision with networking disabled and the verified original in the cache.
  • TimeF read-back matched all 338,892 task rows and targets in the frozen AirAware benchmark release.
  • Connector lint and type checks passed on the supported Python versions. All 974 selected TimeNet tests passed, including the three new connector tests.
  • The repository-wide type check requires optional Torch, Zarr, and documentation dependencies that were outside the local connector environment.

The CSV and generated TimeF artifacts are kept out of Git.

@keneth-arias keneth-arias changed the title Add the Lima hourly PM2.5 forecasting benchmark feat: add the Lima hourly PM2.5 forecasting benchmark Nov 18, 2026
@keneth-arias keneth-arias changed the title feat: add the Lima hourly PM2.5 forecasting benchmark feat: add the Cuybernautas/airaware connector for Lima PM2.5 forecasting Nov 18, 2026
…nector API

Main changed the connector API. A label set carries its name as an
annotation, a Hugging Face connector passes its Hub settings as class
keywords and reads a mapping from part to path, BaseConnector names the
whole download result, and Annotation takes its value by keyword.

This branch has not been deployed

No deployments
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants