JEI_Import_Documents-App/scripts/extract_one.py

"""Single-PDF invoice extractor. Prints JSON to stdout.

Reads a single PDF, parses it via the same regex logic as extract_invoices.py,
and emits a single JSON record (header + lines) to stdout.

Usage:
    python scripts/extract_one.py <path-to-pdf>

Output JSON shape:
{
  "ok": true,
  "invoice_no": "...",
  "invoice_date": "YYYY-MM-DD",
  "buyer_name_variant": "...",
  "account_ref": "...",
  "customer_order_no": "...",
  "sales_order_no": "...",
  "deliver_to_name": "...",
  "deliver_to_address": "...",
  "supplier_name": "...",
  "total_net": 1470.0,
  "carriage_net": 550.0,
  "invoice_total": 2020.0,
  "document_type": "invoice",
  "lines": [
    {"stock_code": "DRILL-RDE36/2", "description": "...", "quantity": 1,
     "unit_price": 1385.0, "net_amount": 1385.0,
     "commodity_code": "84592900", "country_of_origin": "PL"},
    ...
  ],
  "warnings": []   // any parsing anomalies the user should know about
}

On parse failure, emits {"ok": false, "error": "..."}.
"""
import json
import re
import sys
from pathlib import Path

# Import the shared parser from extract_invoices.py (same module SJ has been using)
sys.path.insert(0, str(Path(__file__).resolve().parent))
from extract_invoices import parse_invoice_page  # noqa: E402

import pdfplumber  # noqa: E402


def extract(pdf_path: str) -> dict:
    pdf_path = str(pdf_path)
    if not Path(pdf_path).is_file():
        return {"ok": False, "error": f"File not found: {pdf_path}"}

    # Combine text from all pages — invoices that span multiple pages need it
    page_texts = []
    try:
        with pdfplumber.open(pdf_path) as pdf:
            for i, page in enumerate(pdf.pages):
                try:
                    txt = page.extract_text() or ""
                except Exception as e:
                    txt = ""
                    page_texts = None
                    return {"ok": False, "error": f"Failed to read page {i + 1}: {e}"}
                page_texts.append(txt)
    except Exception as e:
        return {"ok": False, "error": f"Failed to open PDF: {e}"}

    if not page_texts or all(not t.strip() for t in page_texts):
        return {"ok": False, "error": "No text could be extracted from PDF (scanned? needs OCR via ocr_invoices.py first)."}

    # Parse each page, then dedupe by (invoice_no, document_type) — same merge logic
    records = [parse_invoice_page(t, Path(pdf_path).name) for t in page_texts]
    # Group by (invoice_no, document_type) — multi-page proformas share these
    merged: dict[tuple, dict] = {}
    for r in records:
        key = (r.get("invoice_no", ""), r.get("document_type", "invoice"))
        if key not in merged:
            merged[key] = {
                "invoice_no": r.get("invoice_no"),
                "invoice_date": r.get("invoice_date"),
                "buyer_name_variant": r.get("buyer_name_variant"),
                "account_ref": r.get("account_ref"),
                "customer_order_no": r.get("customer_order_no"),
                "sales_order_no": r.get("sales_order_no"),
                "deliver_to_name": r.get("deliver_to_name"),
                "deliver_to_address": r.get("deliver_to_address"),
                "supplier_name": r.get("supplier_name"),
                "total_net": r.get("total_net"),
                "carriage_net": r.get("carriage_net"),
                "invoice_total": r.get("invoice_total"),
                "document_type": r.get("document_type"),
                "lines": [],
            }
        merged[key]["lines"].extend(r.get("lines", []))

    if not merged:
        return {"ok": False, "error": "No invoice records found in PDF."}

    # Take the first record (multi-invoice PDFs aren't supported here — use split_invoices.py first)
    rec = next(iter(merged.values()))

    # Warnings: anything the user should review
    warnings = []
    line_sum = sum(ln.get("net_amount", 0) or 0 for ln in rec["lines"])
    total = rec.get("total_net") or 0
    if rec["lines"] and abs(line_sum - total) > 0.05:
        warnings.append(
            f"Line net_amount sum ({line_sum:.2f}) doesn't match total_net ({total:.2f}) — review the line items."
        )
    if not rec.get("lines"):
        warnings.append("No line items were parsed from this PDF.")
    if not rec.get("invoice_no"):
        warnings.append("Could not detect invoice_no from PDF.")

    # Format date as YYYY-MM-DD (PB date field expects this)
    raw_date = rec.get("invoice_date") or ""
    if raw_date and len(raw_date) >= 10:
        iso_date = raw_date[:10]
    else:
        iso_date = ""

    return {
        "ok": True,
        "invoice_no": rec.get("invoice_no") or "",
        "invoice_date": iso_date,
        "buyer_name_variant": rec.get("buyer_name_variant") or "",
        "account_ref": rec.get("account_ref") or "",
        "customer_order_no": rec.get("customer_order_no") or "",
        "sales_order_no": rec.get("sales_order_no") or "",
        "deliver_to_name": rec.get("deliver_to_name") or "",
        "deliver_to_address": rec.get("deliver_to_address") or "",
        "supplier_name": rec.get("supplier_name") or "",
        "total_net": rec.get("total_net") or 0,
        "carriage_net": rec.get("carriage_net") or 0,
        "invoice_total": rec.get("invoice_total") or 0,
        "document_type": rec.get("document_type") or "invoice",
        "lines": [
            {
                "stock_code": ln.get("stock_code") or "",
                "description": ln.get("description") or "",
                "quantity": ln.get("quantity") or 0,
                "unit_price": ln.get("unit_price") or 0,
                "net_amount": ln.get("net_amount") or 0,
                "commodity_code": ln.get("commodity_code") or "",
                "country_of_origin": ln.get("country_of_origin") or "",
            }
            for ln in rec["lines"]
        ],
        "warnings": warnings,
    }


if __name__ == "__main__":
    if len(sys.argv) != 2:
        print(json.dumps({"ok": False, "error": "Usage: python extract_one.py <path-to-pdf>"}))
        sys.exit(1)
    try:
        result = extract(sys.argv[1])
    except Exception as e:
        result = {"ok": False, "error": f"{type(e).__name__}: {e}"}
    print(json.dumps(result, ensure_ascii=False))