How to export JSON Schemas

The schemas/ directory holds JSON Schema files generated from the package’s Pydantic models. This guide shows how to regenerate them after the models change, and how consumers use them.

Export from the CLI

Write one <model>.schema.json file per model into a directory:

researcher-profiles schema export schemas/

Expected output:

wrote schemas/profile_metadata.schema.json
wrote schemas/profile_config.schema.json
wrote schemas/papers_file.schema.json
wrote schemas/summary_file.schema.json
wrote schemas/question.schema.json

The command creates the output directory if it does not exist and overwrites any existing schema files. Run it whenever you change a model in researcher_profiles.schema so the checked-in files stay in sync.

Export programmatically

Use export_schemas to write the files, or build_schemas to get the schema dicts in memory:

from researcher_profiles.schema_export import export_schemas, build_schemas

written = export_schemas("schemas")     # returns list[Path]
print([p.name for p in written])

schemas = build_schemas()               # returns {name: json_schema_dict}
print(sorted(schemas))

What each file validates

File	Validates	Source model
`profile_metadata.schema.json`	`profile.yaml`	`ProfileMetadata`
`profile_config.schema.json`	`config.json`	`ProfileConfig`
`papers_file.schema.json`	`sources/papers.yaml`	`PapersFile`
`summary_file.schema.json`	`sources/summaries/*.summary.md` frontmatter	`SummaryFile`
`question.schema.json`	`meta/questions.yaml` entries	`Question`

See the schema reference for the fields in each model.

Using the schemas without the package

The point of the checked-in files is that a consumer can validate a profile without installing this package, using any JSON Schema validator in any language. Load the schema file and the target artifact (parsing YAML to JSON-compatible data first) and validate. See Validate a profile for a Python example with the jsonschema package.