flatcitybuf

A from-scratch, pure-Python reader for FlatCityBuf, a cloud-optimized binary format for 3D city models: CityJSON's semantics in FlatBuffers, with a packed Hilbert R-tree for spatial queries, a static B+tree for attribute queries, and HTTP range requests so a client fetches only the bytes it needs. No FFI and no compiled extension — a single py3-none-any wheel on CPython 3.9+, with flatbuffers as its only required dependency. Reader only: write .fcb files with the Rust CLI or the C++ writer.

pip install flatcitybuf          # or: uv pip install flatcitybuf
pip install "flatcitybuf[numpy]" # optional: ~2.4x faster bulk decoding

import json
import flatcitybuf as fcb

reader = fcb.FcbReader.open_file("city.fcb")

# The CityJSONSeq header line.
print(json.dumps(fcb.to_cityjson_metadata(reader.header)))

# Every feature, in stored (Hilbert) order.
for feature in reader.select_all():
    cj = fcb.to_cityjson_feature(feature, reader.header)

# Over HTTP, byte-range by byte-range (synchronously).
remote = fcb.FcbReader.open(fcb.HttpRangeReader("https://example.com/city.fcb"))

Attribute and spatial queries, the optional-numpy story, development commands, and the migration notes for users of the retired PyO3 extension (0.2.0 and earlier, whose API this does not drop-in replace) are all in the guide:

• Python guide — install, full API tour, tooling and testing.
• Runnable examples — eight scripts, one per capability, each with its real output. They run as part of the test suite, so they cannot drift.
• Format specification
• Project README
• Issue tracker

License

MIT — see LICENSE.