amazon-po-parser/parse-mbox.py

#!/usr/bin/env python3.12
"""Parse Coupa PO emails from an MBOX file and extract line items + ship-to data."""

import mailbox
import email
import re
import json
import sys
import os

def parse_amount(s):
    """Parse amount string, handling formats like '1.0 EACH x 3,000.00' or '10,000.00'."""
    s = s.strip()
    if ' x ' in s.lower():
        s = s.split(' x ')[-1].strip()
    elif ' X ' in s:
        s = s.split(' X ')[-1].strip()
    return float(s.replace(',', ''))


def parse_po_email(body):
    """Extract PO data from a Coupa 'issued' email plaintext body."""
    lines = body.split('\n')
    lines = [l.strip() for l in lines]

    # PO number from body
    po_match = re.search(r'Purchase Order #?((?:2D|B187|FK)-\d+)', body)
    if not po_match:
        return None
    po_number = po_match.group(1)

    # Line items: between "Items" line and the coupahost URL
    items_start = None
    items_end = None
    for i, line in enumerate(lines):
        if line == 'Items' and items_start is None:
            items_start = i + 1
        if items_start and 'supplier.coupahost.com/orders/' in line:
            items_end = i
            break

    line_items = []
    if items_start and items_end:
        item_lines = [l for l in lines[items_start:items_end] if l]
        i = 0
        while i < len(item_lines):
            desc = item_lines[i]
            # Skip if this line looks like a number/currency
            if re.match(r'^[\d,]', desc) or desc == 'USD':
                i += 1
                continue
            amount = None
            # Look ahead for amount
            for j in range(i + 1, min(i + 4, len(item_lines))):
                candidate = item_lines[j]
                if re.match(r'^[\d,]', candidate):
                    try:
                        amount = parse_amount(candidate)
                    except ValueError:
                        pass
                    break
            line_items.append({
                'description': desc,
                'amount': amount,
            })
            # Skip past amount + currency lines
            i = j + 1 if amount is not None else i + 1
            while i < len(item_lines) and item_lines[i] == 'USD':
                i += 1

    # Key-value pairs after "More Detail"
    kv = {}
    detail_start = None
    for i, line in enumerate(lines):
        if line == 'More Detail':
            detail_start = i + 1
            break

    if detail_start:
        kv_keys = ['PO ID', 'Department', 'Status', 'Last Opened', 'Order Date',
                    'Acknowledged At', 'Revision Date', 'Payment Term', 'Req #']
        for i in range(detail_start, len(lines)):
            for k in kv_keys:
                if lines[i] == k and i + 1 < len(lines):
                    kv[k] = lines[i + 1]

    # Ship-to address: second "Shipping" section
    shipping_count = 0
    ship_to_lines = []
    capturing = False
    skipping_blanks = False
    for i, line in enumerate(lines):
        if line == 'Shipping':
            shipping_count += 1
            if shipping_count == 2:
                capturing = True
                skipping_blanks = True
                continue
        elif capturing:
            if skipping_blanks and line == '':
                continue
            skipping_blanks = False
            if line in ('', 'Ship To Address') or line.startswith('---') or line.startswith('http'):
                break
            if line in kv_keys or line == 'Supplier' or line == 'More Detail':
                break
            ship_to_lines.append(line)

    ship_to_raw = '\n'.join(ship_to_lines) if ship_to_lines else None

    # Extract site code from ship-to
    site_code = None
    if ship_to_raw:
        # (CODE) pattern — e.g. "Amazon.com Services LLC (XSF2)"
        m = re.search(r'\(([A-Z0-9]{3,5})\)', ship_to_raw)
        if m:
            site_code = m.group(1)
        # "LLC - CODE" — e.g. "Amazon.com Services LLC - WUT9"
        if not site_code:
            m = re.search(r'(?:LLC|Inc)\s*-\s*([A-Z0-9]{3,5})\b', ship_to_raw)
            if m:
                site_code = m.group(1)
        # "ATTN: ... Station CODE" or "ATTN: ... DS - CODE"
        if not site_code:
            m = re.search(r'ATTN:.*?(?:Station|DS)\s*[-–]?\s*([A-Z0-9]{3,5})\b', ship_to_raw)
            if m:
                site_code = m.group(1)
        # "CODE - Amazon" at start of first line
        if not site_code:
            m = re.match(r'^([A-Z0-9]{3,5})\s*-\s*Amazon', ship_to_raw)
            if m:
                site_code = m.group(1)
    if not site_code and line_items:
        for li in line_items:
            m = re.match(r'^([A-Z]{1,4}[0-9]{1,2}|[A-Z]{4,5})\b', li['description'])
            if m:
                site_code = m.group(1)
                break

    return {
        'po_number': po_number,
        'site_code': site_code,
        'status': kv.get('Status'),
        'order_date': kv.get('Order Date'),
        'revision_date': kv.get('Revision Date'),
        'payment_term': kv.get('Payment Term'),
        'req_number': kv.get('Req #'),
        'ship_to_raw': ship_to_raw,
        'line_items': line_items,
    }


def process_mbox(mbox_path, target_pos=None, output_path=None):
    """Process an MBOX file and return parsed PO records."""
    mbox = mailbox.mbox(mbox_path)
    results = {}
    processed = 0
    matched = 0

    for msg in mbox:
        processed += 1
        subject = msg.get('Subject', '')

        if 'issued' not in subject.lower() or 'Purchase Order' not in subject:
            continue

        if msg.is_multipart():
            body = None
            for part in msg.walk():
                if part.get_content_type() == 'text/plain':
                    body = part.get_payload(decode=True).decode('utf-8', errors='replace')
                    break
            if not body:
                continue
        else:
            body = msg.get_payload(decode=True).decode('utf-8', errors='replace')

        record = parse_po_email(body)
        if not record:
            continue

        po = record['po_number']

        if target_pos and po not in target_pos:
            continue

        # Keep latest version if duplicate
        if po not in results:
            results[po] = record
            matched += 1

        if matched % 500 == 0 and matched > 0:
            print(f"  Matched {matched} POs so far... ({processed} messages scanned)", file=sys.stderr)

    mbox.close()

    result_list = list(results.values())
    print(f"Processed {processed} messages, matched {matched} unique POs", file=sys.stderr)

    if output_path:
        with open(output_path, 'w') as f:
            json.dump(result_list, f, indent=2)
        print(f"Saved to {output_path}", file=sys.stderr)

    return result_list


if __name__ == '__main__':
    mbox_path = sys.argv[1] if len(sys.argv) > 1 else '/tmp/coupa-emails/coupa-po-dump--info@seahavenind.com-wHxmHD.mbox'
    po_list_path = sys.argv[2] if len(sys.argv) > 2 else 'output/po-list.json'
    output_path = sys.argv[3] if len(sys.argv) > 3 else 'output/email-parsed-pos.json'

    target_pos = None
    if os.path.exists(po_list_path):
        with open(po_list_path) as f:
            target_pos = set(json.load(f))
        print(f"Targeting {len(target_pos)} specific POs", file=sys.stderr)

    process_mbox(mbox_path, target_pos, output_path)