Tags

Tags give the ability to mark specific points in history as being important

  • v1.0.0

    9eba0cf5 · Updated version to v1.0. ·
    Release: Stable Release 
    v1.0.0 — Initial stable release

Three-stage satellite data pipeline: file discovery, cache creation,
and downstream processing, built on satpy with a persistent Zarr cache.

Features:
- Stage 1: File discovery from local archive and remote sources (Sky DB),
  with timestamp parsing from filenames and YAML-based precomputed file lists.
- Stage 2: Cache creation via CacheIndex — read, calibrate, resample,
  and transform data, writing entries to a persistent Zarr cache with
  partial-pipeline reuse based on transform-chain hashing.
- Stage 3: Downstream processing (statistics, filtering, ML training)
  reading exclusively from the persistent cache.
- SatelliteQuery: channel/calibration/time-range querying, subset
  matching, and solar zenith angle (sza) computation.
- Transformer pipeline (BaseTransform / ChannelTransform / ResampleTransform)
  for resampling, cutouts, reprojection, and unit conversion, with
  multi-area support (common prefix + per-area suffix chains).
- SatelliteDataset / SatelliteTorchConverter: PyTorch Dataset/DataLoader
  integration with stacked tensor conversion.

See docs/README.md for architecture, workflow, and diagram details.

  • v0.0.1

    c778871f · Looks like reading data works now. ·
    Release: v0.0.1 
    Reading and resampling data works in principle.

Imprint | Privacy Policy