Built-in Schemas¶

strutex includes ready-to-use Pydantic schemas for common document types. Use these directly with DocumentProcessor for instant structured extraction.

Quick Start¶

from strutex import DocumentProcessor
from strutex.schemas import INVOICE_US

processor = DocumentProcessor(provider="gemini")
invoice = processor.process(
    "invoice.pdf",
    "Extract invoice details",
    model=INVOICE_US
)

print(f"Invoice #{invoice.invoice_number}")
print(f"Total: ${invoice.total}")
print(f"Items: {len(invoice.line_items)}")

Available Schemas¶

Invoices¶

Schema	Description	Key Fields
`INVOICE_GENERIC`	Universal invoice format	invoice_number, vendor, customer, line_items, total
`INVOICE_US`	US-style with sales tax	+ sales_tax, state_tax_rate, ein
`INVOICE_EU`	EU-style with VAT	+ vat_number, vat_rate, reverse_charge

from strutex.schemas import INVOICE_GENERIC, INVOICE_US, INVOICE_EU

Receipts¶

Schema	Description	Key Fields
`RECEIPT`	Retail/restaurant receipts	merchant_name, items, subtotal, tax, total, payment_method

from strutex.schemas import RECEIPT

receipt = processor.process("receipt.jpg", "Extract receipt", model=RECEIPT)
print(f"Store: {receipt.merchant_name}, Total: ${receipt.total}")

Purchase Orders¶

Schema	Description	Key Fields
`PURCHASE_ORDER`	B2B purchase orders	po_number, vendor_name, line_items, ship_to, payment_terms

from strutex.schemas import PURCHASE_ORDER

Order Confirmations¶

Schema	Description	Key Fields
`ORDER_CONFIRMATION`	Quotations / Order Responses	quotation_number, vendor, buyer, items, delivery, totals

from strutex.schemas import ORDER_CONFIRMATION

Shipping Documents¶

Schema	Description	Key Fields
`BILL_OF_LADING`	Ocean freight B/L	bl_number, shipper, consignee, containers, cargo, port_of_loading, port_of_discharge

Container Fields: container_number, seal_number, type, weight Cargo Fields: description, quantity, gross_weight, volume, hs_code, marks

from strutex.schemas import BILL_OF_LADING

bol = processor.process("bl.pdf", "Extract B/L", model=BILL_OF_LADING)
print(f"{bol.port_of_loading} → {bol.port_of_discharge}")

Financial Documents¶

Schema	Description	Key Fields
`BANK_STATEMENT`	Bank account statements	account_number, opening_balance, closing_balance, transactions

from strutex.schemas import BANK_STATEMENT

Resumes/CVs¶

Schema	Description	Key Fields
`RESUME`	Professional resumes	name, email, skills, work_experience, education, certifications

from strutex.schemas import RESUME

resume = processor.process("resume.pdf", "Extract resume", model=RESUME)
print(f"{resume.name}: {len(resume.work_experience)} positions")

Legal Documents¶

Schema	Description	Key Fields
`CONTRACT_CLAUSE`	Contract key terms	title, parties, effective_date, payment_terms, governing_law

from strutex.schemas import CONTRACT_CLAUSE

Extending Schemas¶

Create custom schemas by inheriting from built-in ones:

from strutex.schemas import InvoiceGeneric
from pydantic import Field

class MyInvoice(InvoiceGeneric):
    """Custom invoice with company-specific fields."""
    internal_code: str = Field(..., description="Internal tracking code")
    department: str = Field(None, description="Department to charge")

# Use your custom schema
result = processor.process("invoice.pdf", "Extract invoice", model=MyInvoice)

Schema Reference¶

InvoiceGeneric Fields¶

Field	Type	Description
`invoice_number`	str	Invoice number/ID
`invoice_date`	str?	Invoice date
`due_date`	str?	Payment due date
`vendor`	Party?	Seller information
`customer`	Party?	Buyer information
`line_items`	list	List of line items
`subtotal`	float?	Subtotal before tax
`tax_amount`	float?	Tax amount
`total`	float	Grand total
`payment_terms`	str?	Payment terms

Receipt Fields¶

Field	Type	Description
`merchant_name`	str	Store name
`receipt_number`	str?	Receipt/transaction number
`date`	str?	Transaction date
`items`	list	Purchased items
`subtotal`	float?	Before tax
`tax`	float?	Tax amount
`total`	float	Total paid
`payment_method`	str?	Cash, Card, etc.

BillOfLading Fields¶

Field	Type	Description
`bl_number`	str	B/L number
`shipper`	str?	Shipper name/address
`consignee`	str?	Consignee name/address
`carrier`	str?	Shipping line
`vessel_name`	str?	Ship name
`port_of_loading`	str?	POL
`port_of_discharge`	str?	POD
`containers`	list	Container details
`cargo`	list	Cargo details

Best Practices¶

Use specific schemas when possible - INVOICE_US over INVOICE_GENERIC for better accuracy
Extend for custom fields - Inherit and add company-specific fields
Validate after extraction - Use Pydantic's validation:

try:
    invoice = processor.process("doc.pdf", "Extract", model=INVOICE_US)
except ValidationError as e:
    print(f"Extraction incomplete: {e}")