Dataset Viewer

The dataset viewer is not available because its heuristics could not detect any supported data files. You can try uploading some data files, or configuring the data files location manually.

cozip GeoParquet Fixtures

A small, generated collection of Flat-profile cozip archives for testing remote random access and the cozip playground.

Four of the five archives use a GeoParquet __metadata__ manifest, so their features render directly on the playground map. The remaining archive is a plain Flat manifest used to verify the non-geospatial fallback.

All archives conform to cozip specification 0.1.0, binary format version 1, and Flat profile 1. They use the standard .zip extension. None of their stored metadata contains the reader-owned columns cozip:location or taco:location; readers create locations at runtime.

Fixtures

Archive Geometry Purpose
data/cities.zip Point Global city markers and attributes
data/routes.zip LineString Routes connecting city pairs
data/regions.zip Polygon A global grid of synthetic regions
data/geometry-showcase.zip Mixed Points, lines, and polygons in one manifest
data/plain-files.zip None Flat-manifest fallback without GeoParquet

To inspect a fixture, copy its Hugging Face download URL into the playground. For example:

https://huggingface.co/datasets/asterisk-labs/cozip-api-fixtures/resolve/main/data/cities.zip

Repository structure

data/                       generated cozip archives
generate/generate.py        deterministic fixture generator
verify/verify_local.py      structural and GeoParquet checks
verify/verify_huggingface.py HTTP Range and remote-reader checks
manifest.json               expected archive metadata
checksums.sha256            SHA-256 checksums

Reproduce and verify

uv sync
uv run python generate/generate.py
uv run python verify/verify_local.py
uv run python verify/verify_huggingface.py --revision main

For stable automated tests, replace main with a release tag or commit hash.

Data and licensing

The geometries, attributes, and payloads are synthetic and generated entirely by generate/generate.py. City coordinates and population values are rounded demonstration values, not an authoritative geographic dataset. Regeneration produces the same logical records, but ZIP timestamps may change the archive checksums; published revisions and their checksums.sha256 are immutable test targets. The fixture collection is released under CC0 1.0.

Downloads last month
75