Pydantic projection

result.to_pydantic("entity", model=...) projects canonical records into Pydantic model instances. Two modes:

Bring your own model

Useful when the downstream pipeline already has a Pydantic model and you just want to fit the extracted data into it.

from pydantic import BaseModel

class Order(BaseModel):
    order_id: str
    quantity: int

orders = result.to_pydantic("order", model=Order)

Field matching is opportunistic by attribute name. Fields the model doesn't declare are dropped silently; only the fields you care about are populated.

Situation	Behaviour
Model has field, canonical has it, types match	Carries over
Model has field, canonical has it, types mismatch	Raises crease.ValidationError
Model has required field, canonical doesn't	Raises crease.ValidationError
Model doesn't have field, canonical has it	Silently dropped

Auto-generated model

If you don't pass model=, crease builds one from the template's field declarations:

orders = result.to_pydantic("order")    # uses Template.model("order")

This is useful for one-off scripts and notebooks where the template is already the source of truth. The generated model uses ConfigDict(extra="ignore", strict=False, arbitrary_types_allowed=True) — it accepts Pydantic's safe coercions (ISO date strings → date, "5" → 5) and ignores extra keys, but rejects truly incompatible types.

Halt by default, opt-in to partial

Both modes halt when the report has errors:

try:
    orders = result.to_pydantic("order", model=Order)
except crease.ValidationError as e:
    e.errors()                  # full error list

Pass allow_partial=True to recover what you can:

orders = result.to_pydantic("order", model=Order, allow_partial=True)
# rows that didn't project are absent from `orders`;
# they're listed in result.report.errors()