WFDASH-WEBAPP/scripts/seed-gielie-quotes.py

#!/usr/bin/env python3
"""
Seed test data for GIELIE customer — 5 quotes with various item counts,
mix of catalogue and non-catalogue items, no pricing set.

Run with the PB admin token file at /tmp/pb_token already populated,
or pass it via $PB_TOKEN env. Auth here is non-interactive (we mint a
fresh token from the env file).
"""
import json
import random
import sys
import urllib.error
import urllib.parse
import urllib.request
from datetime import datetime, timezone

PB = "https://pocketbase-gnqc88suy8fqk6wyryjjdpvl.169.239.180.130.sslip.io"


def req(method, path, token=None, body=None):
    url = f"{PB}{path}"
    headers = {"Content-Type": "application/json"}
    if token:
        headers["Authorization"] = token
    data = json.dumps(body).encode() if body is not None else None
    r = urllib.request.Request(url, data=data, headers=headers, method=method)
    try:
        with urllib.request.urlopen(r) as resp:
            return json.loads(resp.read())
    except urllib.error.HTTPError as e:
        body = e.read().decode("utf-8", "replace")
        raise RuntimeError(f"{method} {path} → {e.code}: {body}")


def mint_token():
    auth = req("POST", "/api/collections/users/auth-with-password", body={
        "identity": "david.els1@outlook.com",
        "password": "1234567890",
    })
    return auth["token"]


def get_or_create_customer(token, name="GIELIE", email="gielie@example.com"):
    existing = req("GET", f"/api/collections/customers/records?filter={urllib.parse.quote(f'customer=\"{name}\"')}&fields=id,customer,email&perPage=1", token=token)
    if existing.get("items"):
        print(f"[customer] {name} already exists: {existing['items'][0]['id']}")
        return existing["items"][0]["id"]

    # Pull the first company for optional relation (non-required)
    company_id = None
    try:
        cos = req("GET", "/api/collections/companies/records?perPage=1&fields=id", token=token)
        if cos.get("items"):
            company_id = cos["items"][0]["id"]
    except Exception:
        pass

    payload = {
        "customer": name,
        "email": email,
        "alternative_email": "orders@gielie.co.za",
        "number": "+27 11 555 0100",
        "description": "Seeded test customer for agent auto-fill dry-runs.",
    }
    if company_id:
        payload["company"] = company_id
    created = req("POST", "/api/collections/customers/records", token=token, body=payload)
    print(f"[customer] created {name}: {created['id']}")
    return created["id"]


def next_quote_number(token):
    # PB filter `quote_number~"Q"` plus sort by created desc — we want the latest
    # numeric Q-number so we can produce the next monotonically increasing one.
    raw = req("GET", "/api/collections/quotes/records?sort=-created&perPage=200&fields=quote_number", token=token)
    nums = []
    for it in raw.get("items", []):
        qn = it.get("quote_number") or ""
        # Accept formats: Q1782862717620 (numeric) or anything that ends in digits
        digits = "".join(c for c in qn if c.isdigit())
        if digits:
            nums.append(int(digits))
    return (max(nums) if nums else 0) + 1


def fetch_catalogue_sample(token, n=80):
    """Sample n catalogue rows across categories so the quote items look diverse."""
    rows = []
    page = 1
    while len(rows) < n:
        chunk = req(
            "GET",
            f"/api/collections/price_list/records?perPage=200&page={page}"
            "&fields=id,sku,description,brand,category,cost_price,selling_price"
            "&sort=sku",
            token=token,
        )
        rows.extend(chunk.get("items", []))
        if chunk.get("page", 1) >= chunk.get("totalPages", 1):
            break
        page += 1
    return rows[:n]


# Sentinel embedded in `quotes.notes` to identify rows this script
# created. Used by cleanup scripts — do NOT rename without updating
# the cleanup section. (We can't use `claimed_by` because that field
# drives the UI's "Claimed by X" pill — see the make_quote() comment.)
CLEANUP_TAG = "[gielie-seed]"


# Non-catalogue "custom" line items the agent should NOT auto-match.
# Realistic ad-hoc jobs a buyer might request — never in the catalogue.
CUSTOM_DESCRIPTIONS = [
    "Custom fabrication labour 3 hours (workshop)",
    "Site delivery Gauteng region — small van",
    "On-site installation and commissioning half day",
    "Hire of scissor lift 8m working height, 1 day",
    "Cable termination and labelling per point, 24 points",
    "Custom printed banner 3m x 1m, full colour",
    "Reprogram PLC to new ladder logic, supplier attendance",
    "Sandblast and powder-coat steel brackets, batch of 12",
    "Emergency callout after-hours weekend rate",
    "Pressure-test hydraulic lines to 250 bar per circuit",
    "Custom-cut stainless steel sheet 3mm, 600x600",
    "3D printed jigs and fixtures, PLA batch",
]


def make_line(cat):
    """Build a 'catalogue match' line — uses the catalogue description verbatim
    so the search_text fuzzy will hit."""
    return {
        "description": cat["description"],
        "quantity": random.choice([1, 2, 3, 5, 10]),
    }


def make_custom_line():
    return {
        "description": random.choice(CUSTOM_DESCRIPTIONS),
        "quantity": random.choice([1, 2, 3, 4]),
    }


def make_quote(token, customer_id, quote_number_int, line_count, custom_ratio, catalogue):
    """Build a quote + its quote_items. Returns (quote_id, item_ids)."""
    n_catalogue = max(0, int(round(line_count * (1 - custom_ratio))))
    n_custom = line_count - n_catalogue
    chosen = random.sample(catalogue, k=min(n_catalogue, len(catalogue)))
    lines = [make_line(c) for c in chosen]
    for _ in range(n_custom):
        lines.append(make_custom_line())

    # IMPORTANT: do not set `claimed_by` here — even though status is `new`
    # (so the Claim button still works), the quote detail page renders a
    # "Claimed by X" pill as soon as `claimed_by` is truthy, and that
    # signals to the operator that someone else has their hands on it.
    # These seeds have no agent hold on them, so the field must stay
    # empty. Use `notes` as the cleanup marker instead — see CLEANUP_TAG.
    quote_payload = {
        "quote_number": f"Q{quote_number_int}",
        "quote_reference": f"GIELIE-TEST-{quote_number_int}",
        "customer": customer_id,
        "status": "new",
        "notes": (
            f"{CLEANUP_TAG} — {line_count} line items, {n_custom} non-catalogue. "
            "No pricing set."
        ),
        "description": f"Seed quote {quote_number_int} — {line_count} line items, {n_custom} non-catalogue.",
    }
    quote = req("POST", "/api/collections/quotes/records", token=token, body=quote_payload)
    quote_id = quote["id"]

    created_items = []
    for i, line in enumerate(lines, 1):
        item_payload = {
            "quote": quote_id,
            "description": line["description"],
            "quantity": line["quantity"],
            # intentionally leave cost_price, selling_price, supplier, sku, markup empty
            # so the quote lands in the "unpriced" bucket the agent picks up.
            "is_processed": False,
            "user_edited": True,  # seed as user-edited so the AI rewrite gate doesn't block dry-runs
        }
        it = req("POST", "/api/collections/quote_items/records", token=token, body=item_payload)
        created_items.append(it["id"])

    return {
        "quote_id": quote_id,
        "quote_number": f"Q{quote_number_int}",
        "line_count": line_count,
        "n_custom": n_custom,
        "n_catalogue": n_catalogue,
        "item_ids": created_items,
    }


def main():
    token = mint_token()
    customer_id = get_or_create_customer(token, "GIELIE")
    catalogue = fetch_catalogue_sample(token, n=80)
    print(f"[catalogue] sampled {len(catalogue)} rows")

    start_q = next_quote_number(token)
    print(f"[quote_number] next = Q{start_q}")

    # Layout: 1×1 line, 1×3 lines, 1×4 lines, 1×10 lines, 1×7 lines (heavy custom)
    shapes = [
        (1, 0.0),
        (3, 0.34),  # ~1 of 3 is custom
        (4, 0.50),  # 2 of 4 custom
        (10, 0.30), # 3 of 10 custom
        (7, 0.71),  # 5 of 7 custom — heavy non-match to stress Tier 3 + no-match
    ]

    random.seed(42)
    results = []
    for i, (size, custom_ratio) in enumerate(shapes):
        qn = start_q + i
        out = make_quote(token, customer_id, qn, size, custom_ratio, catalogue)
        results.append(out)
        print(f"[quote] {out['quote_number']} id={out['quote_id']} lines={out['line_count']} (custom {out['n_custom']} / catalogue {out['n_catalogue']})")

    print("\n=== SUMMARY ===")
    for r in results:
        print(json.dumps(r, indent=2))


if __name__ == "__main__":
    main()