mysoc-validator-fast

A Rust-backed shadow of mysoc-validator for mySociety democracy data: Popolo, publicwhip XML transcripts, and JSON registers of interests. Rust handles parsing and validation, with a Python API that currently targets mysoc-validator 1.5. Package versions are independent of the upstream compatibility target.

Requires Python 3.11 or newer.

pip install mysoc-validator-fast

Usage

These examples mirror the upstream README with imports changed to mysoc_validator_fast.

Look up a person by identifier or name:

from datetime import date

from mysoc_validator_fast import Chamber, IdentifierScheme, Popolo

popolo = Popolo.from_parlparse()
by_id = popolo.persons.from_identifier("4514", scheme=IdentifierScheme.MNIS)
by_name = popolo.persons.from_name(
    "keir starmer", chamber_id=Chamber.COMMONS, date=date(2022, 7, 31)
)
assert by_id.id == by_name.id

Load a transcript or register of interests:

from pathlib import Path

from mysoc_validator_fast import RegmemRegister, Transcript

transcript = Transcript.from_xml_path(Path("data/debates2023-03-28d.xml"))
interests = RegmemRegister.from_path(Path("data/commons-regmem-2025-01-20.json"))

Validate files from the command line:

uvx mysoc-validator-fast popolo validate people.json
uvx mysoc-validator-fast transcript validate debates.xml
uvx mysoc-validator-fast transcript validate transcripts/
uvx mysoc-validator-fast transcript validate 'transcripts/*.xml' --glob
uvx mysoc-validator-fast interests validate interests.json
uvx mysoc-validator-fast popolo format people.json

mysoc-validator-fast and msv-fast are also CLI entry points. Use --help to see available commands.

Current differences from mysoc-validator

This is not yet a complete drop-in replacement:

• Imports: use mysoc_validator_fast, with models in .popolo, .interests and .transcripts; upstream's mysoc_validator.models.* paths are not provided. Import Pydantic extra schemas such as MembershipExtra from mysoc_validator_fast.popolo_extras.
• Coverage: legacy XML interests (XMLRegister), transcript JSON I/O, info XML models (InfoCollection, PersonInfo, ConsInfo), and the Popolo TUI are not implemented.
• Pydantic integration: core models are Rust extension classes, not Pydantic models. Selected model_validate/model_dump helpers are provided; the full Pydantic API, its keyword options, and ValidationError structure are not. Rust validation failures generally raise ValueError.
• Mutation: Popolo collection lookups return live handles, but appending a standalone model copies it. Retrieve it from the collection for later edits. Names and dates are copies, and changing a collection's returned .root list does not edit the collection; use setters and collection methods to write changes. Register details, annotations, and summaries are also detached values; use entry/detail-group editing methods.
• Typed extras: get_extra_as(CustomExtra) returns a validated Pydantic snapshot. Editing that snapshot does not update the record; write through set_extra or the live extra proxy.
• Transcript editing: item wrappers are copies. Editing an item does not update the transcript; model_dump_xml() returns its retained source XML.

See the repository README for implementation and development details.