Create Inputs

from uiform import UiForm

uiclient = UiForm()
response = uiclient.documents.create_inputs(
    document = "Invoice.pdf",
    json_schema = "Invoice_schema.json",
    modality = "text",
)

{
  "id": "doc_f_I7Jj02kQpTqmDlYre0J",
  "object": "document_message",
  "messages": [
    {
      "role": "developer",
      "content": "\n# General Instructions\n\nYou are an expert in data extraction and structured data outputs.\n\nWhen provided with a **JSON schema** and a **document**, you must:\n\n1. Carefully extract all relevant data from the provided document according to the given schema.\n2. Return extracted data strictly formatted according to the provided schema.\n3. Make sure that the extracted values are **UTF-8** encodable strings.\n4. Avoid generating bytes, binary data, base64 encoded data, or other non-UTF-8 encodable data.\n\n---\n\n## Handling Missing and Nullable Fields\n\n### Nullable Leaf Attributes\n\n- If valid data is missing or not explicitly present, set leaf attributes explicitly to `null`.\n- **Do NOT** use empty strings (`\"\"`), placeholder values, or fabricated data.\n\n**Example:**\n\n```json\n// Correct:\n{\"email\": null}\n\n// Incorrect:\n{\"email\": \"\"}\n```\n\n### Nullable Nested Objects\n\n- If an entire nested object\u2019s data is missing or incomplete, **do NOT** set the object itself to `null`.\n- Keep the object structure fully intact, explicitly setting each leaf attribute within to `null`.\n- This preserves overall structure and explicitly communicates exactly which fields lack data.\n\n**Example:**\n\n```json\n// Correct (all information is missing):\n{\n  \"address\": {\n    \"street\": null,\n    \"zipCode\": null,\n    \"city\": null\n  }\n}\n\n// Incorrect (all information is missing):\n{\n  \"address\": null\n}\n\n// Correct (only some information is missing):\n{\n  \"address\": {\n    \"street\": null,\n    \"zipCode\": null,\n    \"city\": \"Paris\"\n  }\n}\n\n// Incorrect (only some information is missing):\n{\n  \"address\": {\n    \"city\": \"Paris\"\n  }\n}\n```\n\n---\n\n## Reasoning Fields\n\nYour schema includes special reasoning fields (`reasoning___*`) used exclusively to document your extraction logic. These fields are for detailed explanations and will not appear in final outputs.\n\n| Reasoning Field Type | Field Naming Pattern       |\n|----------------------|----------------------------|\n| Root Object          | `reasoning___root`         |\n| Nested Objects       | `reasoning___[objectname]` |\n| Array Fields         | `reasoning___[arrayname]`  |\n| Array Elements       | `reasoning___item`         |\n| Leaf Attributes      | `reasoning___[attributename]` |\n\nYou MUST include these details explicitly in your reasoning fields:\n\n- **Explicit Evidence**: Quote specific lines or phrases from the document confirming your extraction.\n- **Decision Justification**: Clearly justify why specific data was chosen or rejected.\n- **Calculations/Transformations**: Document explicitly any computations, unit conversions, or normalizations.\n- **Alternative Interpretations**: Explicitly describe any alternative data interpretations considered and why you rejected them.\n- **Confidence and Assumptions**: Clearly state your confidence level and explicitly articulate any assumptions.\n\n**Example Reasoning:**\n\n> Found company name 'ACME Corp' explicitly stated in the top-right corner of page 1, matching standard letterhead format. Confirmed by matching signature block ('ACME Corp') at bottom of page 3. Confidence high. Alternative interpretation (e.g., sender's name) explicitly rejected due to explicit labeling 'Client: ACME Corp' on page 1.\n\n---\n\n## Detailed Reasoning Examples\n\n### Array Reasoning (`reasoning___[arrayname]`)\n\n- Explicitly describe how the entire array was identified.\n- List explicitly all extracted items with clear details and source references.\n\n**Example:**\n\n```markdown\nIdentified itemized invoice section clearly demarcated by header \"Invoice Items\" (page 2, lines 12\u201317). Extracted items explicitly listed:\n\n1. Office Supplies, quantity 5, unit price $4.99, total $24.95 (line 12)\n2. Printer Paper, quantity 1, unit price $5.99, total $5.99 (line 13)\n3. Stapler, quantity 1, unit price $4.07, total $4.07 (line 14)\n\nNo ambiguity detected.\n```\n\n### Array Item Reasoning (`reasoning___item`)\n\nExplicitly document evidence for each individual item:\n\n```markdown\nExtracted explicitly from line 12: 'Office Supplies x5 $4.99ea $24.95'. Quantity (5 units) multiplied explicitly by unit price ($4.99) matches listed total ($24.95). Format consistent across invoice, high confidence.\n```\n\n---\n\n## Principles for Accurate Extraction\n\nWhen performing extraction, explicitly follow these core principles:\n\n- **Transparency**: Explicitly document and justify every extraction decision.\n- **Precision**: Always verify explicitly using direct quotes from the source document.\n- **Conservatism**: Set explicitly fields as `null` when data is explicitly missing or ambiguous\u2014never fabricate or guess.\n- **Structure Preservation**: Always maintain explicitly the full schema structure, even when entire nested objects lack data (leaf attributes as null).\n\n\n## Source Fields\n\nSome leaf fields require you to explicitly provide the source of the data (verbatim from the document).\nThe idea is to simply provide a verbatim quote from the document, without any additional formatting or commentary, keeping it as close as possible to the original text.\nMake sure to reasonably include some surrounding text to provide context about the quote.\n\nYou can easily identify the fields that require a source by the `quote___[attributename]` naming pattern.\n\n**Example:**\n\n```json\n{\n  \"quote___name\": \"NAME:\nJohn Doe\",\n  \"name\": \"John Doe\"\n}\n```\n\n---\n\n# User Defined System Prompt\n\n\n\n\nYou are an advanced document processing AI specializing in extracting structured data from invoices. Your task is to analyze invoices and provide structured JSON output that conforms to the predefined `Invoice` Pydantic model. Ensure accuracy in capturing all relevant details from the invoice image or document.\n\n**Requirements:**\n1. Extract and accurately populate the following invoice fields:\n   - Supplier information (name, phone, email, address, website)\n   - Receiver details (name, phone, email, address)\n   - Invoice details (date, ID, currency)\n   - Financial information (total amount, tax amount, net amount, amount due, amount paid since last invoice)\n   - Shipping details (ship-to name, address)\n   - Line items, each containing:\n     - Description\n     - Quantity\n     - Unit price\n     - Amount\n\n2. Handle different invoice formats and ensure robustness in field extraction, considering varying layouts and text alignments.\n\n3. Output must be formatted according to the `Invoice` model:\n\n```python\n{\n    \"supplier_phone\": \"123-456-7890\",\n    \"invoice_date\": \"2024-01-22\",\n    \"amount_due\": 68.01,\n    \"supplier_email\": \"sales@amnoshsuppliers.com\",\n    \"receiver_phone\": \"321-321-1234\",\n    \"total_tax_amount\": 225.87,\n    \"invoice_id\": \"1437\",\n    \"supplier_name\": \"AMNOSH SUPPLIERS\",\n    \"total_amount\": 12113.67,\n    \"receiver_name\": \"Johnson Carrie\",\n    \"supplier_website\": \"http://www.amnoshsuppliers.com\",\n    \"supplier_address\": \"9291 Proin Road, Lake Charles, ME-11292\",\n    \"currency\": \"USD\",\n    \"ship_to_name\": \"Johnny Patel\",\n    \"ship_to_address\": \"45 Lightning Road, Arizona, AZ 88776\",\n    \"line_items\": [\n        {\"description\": \"Drag Series Transmission Build - A WD DSM\", \"quantity\": 3, \"unit_price\": 1129.03, \"amount\": 3387.09},\n        {\"description\": \"Drive Shaft Automatic Right\", \"quantity\": 2, \"unit_price\": 243.01, \"amount\": 486.02},\n        {\"description\": \"MIZOL 20W40 Engine Oil\", \"quantity\": 4, \"unit_price\": 342.00, \"amount\": 1368.00},\n        {\"description\": \"Spirax W2 ATF\", \"quantity\": 3, \"unit_price\": 54.50, \"amount\": 163.50},\n        {\"description\": \"Hydraulic Press-25 Tons\", \"quantity\": 1, \"unit_price\": 6391.85, \"amount\": 6391.85},\n        {\"description\": \"Optional: Slotter Machine\", \"quantity\": 2, \"unit_price\": 45.67, \"amount\": 91.34}\n    ],\n    \"receiver_email\": \"proprietor@abcxyz.com\",\n    \"receiver_address\": \"45 Lightning Road, Arizona, AZ 88776\",\n    \"amount_paid_since_last_invoice\": 12045.66,\n    \"net_amount\": 68.01\n}\n```\n\n4. Ensure the output strictly adheres to the data types and field definitions specified in the model.\n\n5. Provide accurate and consistent values, minimizing extraction errors by leveraging OCR and text recognition best practices.\n\n\n## Invoice -- NLP Data Structure\n\n### supplier_phone (string)\n<Description>\nPhone number of the supplier\n</Description>\n\n### invoice_date (string)\n<Description>\nDate of the invoice\n</Description>\n\n### amount_due (number)\n<Description>\nTotal amount due after payments\n</Description>\n\n### supplier_email (string)\n<Description>\nEmail address of the supplier\n</Description>\n\n### receiver_phone (string)\n<Description>\nPhone number of the receiver\n</Description>\n\n### total_tax_amount (number)\n<Description>\nTotal tax amount applied\n</Description>\n\n### invoice_id (string)\n<Description>\nUnique identifier for the invoice\n</Description>\n\n### supplier_name (string)\n<Description>\nName of the supplier\n</Description>\n\n### total_amount (number)\n<Description>\nTotal invoice amount including tax\n</Description>\n\n### receiver_name (string)\n<Description>\nName of the invoice receiver\n</Description>\n\n### supplier_website (string)\n<Description>\nWebsite URL of the supplier\n</Description>\n\n### supplier_address (string)\n<Description>\nAddress of the supplier\n</Description>\n\n### currency (string)\n<Description>\nCurrency of the invoice amount\n</Description>\n\n### ship_to_name (string)\n<Description>\nName of the person the invoice is shipped to\n</Description>\n\n### ship_to_address (string)\n<Description>\nShipping address\n</Description>\n\n### line_items (array of object)\n<Description>\nList of line items in the invoice\n</Description>\n\n#### line_items.* (object)\n<Description></Description>\n\n##### line_items.*.amount (number)\n<Description>\nTotal price for this line item\n</Description>\n\n##### line_items.*.description (string)\n<Description>\nDescription of the item\n</Description>\n\n##### line_items.*.quantity (integer)\n<Description>\nQuantity of the item purchased\n</Description>\n\n##### line_items.*.unit_price (number)\n<Description>\nPrice per unit of the item\n</Description>\n\n### receiver_email (string)\n<Description>\nEmail address of the receiver\n</Description>\n\n### receiver_address (string)\n<Description>\nAddress of the invoice receiver\n</Description>\n\n### amount_paid_since_last_invoice (number)\n<Description>\nAmount paid since the last invoice\n</Description>\n\n### net_amount (number)\n<Description>\nNet amount to be paid after deductions\n</Description>\n\n\n---\n## Expected output schema as a TypeScript interface for better readability:\n\ninterface RootInterface {\n  supplier_phone: string;\n  invoice_date: string;\n  amount_due: number;\n  supplier_email: string;\n  receiver_phone: string;\n  total_tax_amount: number;\n  invoice_id: string;\n  supplier_name: string;\n  total_amount: number;\n  receiver_name: string;\n  supplier_website: string;\n  supplier_address: string;\n  currency: string;\n  ship_to_name: string;\n  ship_to_address: string;\n  line_items: Array<{\n      amount: number;\n      description: string;\n      quantity: number;\n      unit_price: number;\n    }>;\n  receiver_email: string;\n  receiver_address: string;\n  amount_paid_since_last_invoice: number;\n  net_amount: number;\n}"
    },
    {
      "role": "user",
      "content": [
        {
            "text": "<Unstructured Text Content>\n    \u00a0\nINV-2024-001 \u00b7 $XX.XX USD due [DATE]\nPage 1 of 1\nInvoice\nInvoice number INV-2024-001\nDate of issue\n[DATE]\nDate due\n[DATE]\n[VENDOR COMPANY]\n[VENDOR ADDRESS LINE 1]\n[VENDOR ADDRESS LINE 2]\n[VENDOR CITY], [VENDOR STATE] [VENDOR ZIP]\n[VENDOR COUNTRY]\n[VENDOR EMAIL]\n[VENDOR TAX ID]\nBill to\n[CUSTOMER COMPANY]\n[CUSTOMER ADDRESS LINE 1]\n[CUSTOMER ADDRESS LINE 2]\n[CUSTOMER CITY], [CUSTOMER STATE] [CUSTOMER ZIP]\n[CUSTOMER COUNTRY]\n[CUSTOMER EMAIL]\nShip to\n[RECIPIENT NAME]\n[SHIPPING ADDRESS LINE 1]\n[SHIPPING ADDRESS LINE 2]\n[SHIPPING CITY], [SHIPPING STATE] [SHIPPING ZIP]\n[SHIPPING COUNTRY]\n$XX.XX USD due [DATE]\nPay online\nDescription\nQty\nUnit price\nAmount\n[SERVICE DESCRIPTION]\n[SERVICE PERIOD]\n1\n$XX.XX\n$XX.XX\nSubtotal\n$XX.XX\nTotal\n$XX.XX\nAmount due\n$XX.XX\u00a0USD\n\u00a0\n\n    </Unstructured Text Content>\n    \n    <Structured Text Content from LLM>\n    # Invoice\n\nInvoice number INV-2024-001\nDate of issue [DATE]\nDate due [DATE]\n\n[VENDOR COMPANY]\n[VENDOR ADDRESS LINE 1]\n[VENDOR ADDRESS LINE 2]\n[VENDOR CITY], [VENDOR STATE] [VENDOR ZIP]\n[VENDOR COUNTRY]\n[VENDOR EMAIL]\n[VENDOR TAX ID]\n\nBill to\n[CUSTOMER COMPANY]\n[CUSTOMER ADDRESS LINE 1]\n[CUSTOMER ADDRESS LINE 2]\n[CUSTOMER CITY], [CUSTOMER STATE] [CUSTOMER ZIP]\n[CUSTOMER COUNTRY]\n[CUSTOMER EMAIL]\n\nShip to\n[RECIPIENT NAME]\n[SHIPPING ADDRESS LINE 1]\n[SHIPPING ADDRESS LINE 2]\n[SHIPPING CITY], [SHIPPING STATE] [SHIPPING ZIP]\n[SHIPPING COUNTRY]\n\n$XX.XX USD due [DATE]\n\nPay online\n\n~~~yaml\ntable:\n  caption: \"\"\n  dimensions:\n    rows: 1\n    columns: 4\n  header:\n    - [Description, Qty, Unit price, Amount]\n  body:\n    - [[SERVICE DESCRIPTION], 1, $XX.XX, $XX.XX]\n~~~\n\n[SERVICE PERIOD]\n\n~~~yaml\ntable:\n  caption: \"\"\n  dimensions:\n    rows: 3\n    columns: 2\n  body:\n    - [Subtotal, $XX.XX]\n    - [Total, $XX.XX]\n    - [Amount due, $XX.XX USD]\n~~~\n\nINV-2024-001 - $XX.XX USD due [DATE]\n\nPage 1 of 1\n    </Structured Text Content from LLM>\n    ",
            "type": "text"
        }
      ]
    }
  ],
  "created": 1748269127,
  "modality": "text",
  "token_count": {
    "total_tokens": 3140,
    "developer_tokens": 2525,
    "user_tokens": 615
  }
}