Extract - NDI Platform API

Exactly one of json_schema or schema_id is required. not_found on a field is an answer about the document, not an HTTP error.

from ndi_sdk import NdiClient, ParseResultSource

schema = {
    "type": "object",
    "properties": {
        "invoice_number": {"type": "string"},
        "total": {"type": "number"},
    },
    "required": ["invoice_number", "total"],
}

with NdiClient() as client:
    upload = client.documents.create_upload("invoice.pdf")
    parsed = client.jobs.wait(client.documents.parse(upload, wait_seconds=30).job_id)
    job = client.jobs.wait(
        client.documents.extract(
            ParseResultSource(job_id=parsed.job_id),
            json_schema=schema,
            wait_seconds=30,
        ).job_id
    )
    print(job.result.data)
    for field in job.result.fields:
        print(field.path, field.status, field.value)

Method signature

def extract(
    source: DocumentSource | UploadResponse,
    *,
    json_schema: JsonSchema | None = None,
    schema_id: str | None = None,
    instructions: str | None = None,
    page_ranges: list[PageRange] | None = None,
    citations: ExtractCitationOptions | None = None,
    wait_seconds: int = 0,
    idempotency_key: str | None = None,
) -> Job

Parameters

Parameter	Type	Required	Description
`source`	`DocumentSource` or `UploadResponse`	Yes	File or prior parse
`json_schema`	JSON Schema object	One of	Inline schema
`schema_id`	`str`	One of	Saved schema id
`instructions`	`str`	No	Extra extraction guidance
`page_ranges`	`list[PageRange]`	No	Not available on a `parse_result` source
`citations`	`ExtractCitationOptions`	No	Default on
`wait_seconds`	`int`	No	Hold the HTTP response open (max 300)

Returns

A Job. On success, result.result_type == "extract". result.data is the filled object; result.fields is the per-path status list.

Validate a schema

from ndi_sdk import NdiClient

schema = {"type": "object", "properties": {"total": {"type": "number"}}}

with NdiClient() as client:
    check = client.documents.validate_extract_schema(json_schema=schema)
    print(check.valid, check.errors)

Register reusable schemas over HTTP (POST /v1/extraction-schemas) and pass schema_id. There is no SDK registry namespace. See Extract guide and Extract response.