Pular para conteúdo

FAQ

Answers to common questions about using and developing this library. Add project-specific entries as they come up in issues.

See also: Usage · Examples · Contributing.


How do I install it?

pip install filings-rfb
pip install "filings-rfb[parquet]"   # optional Parquet export

Or with Poetry:

poetry add filings-rfb
poetry add "filings-rfb[parquet]"   # optional Parquet export

Until the first PyPI release is published, install from source: pip install git+https://github.com/guilhermegor/filings-rfb.git.

How do I add or update a dependency?

Use Poetry so the lock file stays authoritative:

poetry add <package>          # runtime dependency
poetry add --group dev <package>   # dev-only tool

Every package the code imports must be a direct dependency — never rely on it arriving transitively through another package.

Which Python versions are supported?

See the python = "..." constraint in pyproject.toml; CI runs the test matrix against each.

How is the version determined?

The version is the git tag (via poetry-dynamic-versioning); pyproject.toml carries the version for Commitizen (1.0.0 at the first release) but the built wheel takes the tag. Cut a release from the release workflow — see Contributing.

What is the cnpj.tar.gz at the root of the RFB share, and does the library use it?

No, the library ignores it. It is a 63,954,782,749-byte (59.6 GiB) gzip of a tar (probably a single gzip member), last modified 2026-01-28 and never refreshed since. It appears to hold only the nine monthly folders 2025-05 to 2026-01 (inferred from the gzip trailer and byte-range checks, not a full listing), and the sampled files are byte-identical to the per-month zips that remain on the share, so it adds no data, cannot back-fill anything before 2025-05, and is impractical to read selectively (a gzip stream offers no known seek points). Download the YYYY-MM/<file>.zip parts you need instead. The evidence is in .specs/features/cnpj-tarball/design.md.