WFDASH-WEBAPP/scripts/smoke/cleanup-junk-data.py

#!/usr/bin/env python3
"""
WFDash PB data cleanup — removes junk placeholder/smoke-test records,
fixes two quote status anomalies, and clears {0,0} (null-island)
coordinates so the map doesn't render ocean pins.

Safe: checks for references before deleting. DRY-RUN by default; pass
--apply to actually mutate.
"""
import sys, json, urllib.request, urllib.error

PB = "https://pocketbase-gnqc88suy8fqk6wyryjjdpvl.169.239.180.130.sslip.io"
ADMIN_EMAIL = "david.els1@outlook.com"
ADMIN_PASS = "1234567890"
APPLY = "--apply" in sys.argv

def req(method, path, body=None, token=None):
    url = f"{PB}{path}"
    data = json.dumps(body).encode() if body is not None else None
    r = urllib.request.Request(url, data=data, method=method)
    r.add_header("Content-Type", "application/json")
    if token:
        r.add_header("Authorization", token)
    try:
        with urllib.request.urlopen(r) as resp:
            return json.loads(resp.read()) if resp.status != 204 else {}
    except urllib.error.HTTPError as e:
        return {"__error": e.code, "__body": e.read().decode()[:300]}

# 1. Auth
auth = req("POST", "/api/collections/_superusers/auth-with-password",
           {"identity": ADMIN_EMAIL, "password": ADMIN_PASS})
token = auth.get("token", "")
if not token:
    print("AUTH FAILED:", auth); sys.exit(1)
print(f"[auth] admin token ok ({len(token)} chars)")

def list_all(collection, filter_str="", per_page=500):
    out = []
    page = 1
    while True:
        f = f"&filter={urllib.parse.quote(filter_str)}" if filter_str else ""
        r = req("GET", f"/api/collections/{collection}/records?page={page}&perPage={per_page}{f}", token=token)
        if "__error" in r:
            print(f"  [warn] list {collection} failed: {r['__error']}"); break
        items = r.get("items", [])
        out.extend(items)
        if page >= r.get("totalPages", 1): break
        page += 1
    return out

import urllib.parse

# 2. Find junk records
print("\n=== Finding junk records ===")
companies = list_all("companies")
customers = list_all("customers")

junk_company_ids = [c["id"] for c in companies
                    if c.get("company","").startswith("SELECT/CREATE") or c.get("company","").startswith("SMOKE-CO")]
junk_customer_ids = [c["id"] for c in customers
                     if c.get("customer","").startswith("SELECT/CREATE")]
print(f"  junk companies: {junk_company_ids}")
print(f"  junk customers: {junk_customer_ids}")

# 3. Check references before deleting
print("\n=== Checking references ===")
# Quotes that reference junk customers or companies
quotes = list_all("quotes")
ref_customers = set(junk_customer_ids)
ref_companies = set(junk_company_ids)
quotes_blocked = [q for q in quotes if q.get("customer") in ref_customers or q.get("company") in ref_companies]
print(f"  quotes referencing junk customer/company: {len(quotes_blocked)}")
if quotes_blocked:
    for q in quotes_blocked[:5]:
        print(f"    {q.get('quote_number')} (customer={q.get('customer')}, company={q.get('company')})")

# Orders referencing junk
orders = list_all("orders")
orders_blocked = [o for o in orders if o.get("customer") in ref_customers or o.get("company") in ref_companies]
print(f"  orders referencing junk customer/company: {len(orders_blocked)}")

# Customers referencing junk company
cust_blocked = [c for c in customers if c.get("company") in ref_companies and c["id"] not in junk_customer_ids]
print(f"  non-junk customers referencing junk company: {len(cust_blocked)}")

# 4. Delete junk (customers first, then companies — avoids dangling customer→company ref)
print("\n=== Deleting junk ===")
deleted = 0
for cid in junk_customer_ids:
    if quotes_blocked or orders_blocked or cust_blocked:
        # only block if THIS customer is referenced
        this_ref = any(q.get("customer")==cid for q in quotes_blocked) or any(o.get("customer")==cid for o in orders_blocked)
        if this_ref:
            print(f"  SKIP customer {cid} — referenced by quotes/orders"); continue
    if APPLY:
        r = req("DELETE", f"/api/collections/customers/records/{cid}", token=token)
        if "__error" in r: print(f"  FAIL customer {cid}: {r['__error']} {r['__body']}")
        else: print(f"  DELETED customer {cid}"); deleted += 1
    else:
        print(f"  [dry-run] would delete customer {cid}"); deleted += 1

for cid in junk_company_ids:
    # check no remaining references
    still_ref = any(q.get("company")==cid for q in quotes) or any(o.get("company")==cid for o in orders) or any(c.get("company")==cid for c in customers if c["id"] not in junk_customer_ids)
    if still_ref:
        print(f"  SKIP company {cid} — still referenced"); continue
    if APPLY:
        r = req("DELETE", f"/api/collections/companies/records/{cid}", token=token)
        if "__error" in r: print(f"  FAIL company {cid}: {r['__error']} {r['__body']}")
        else: print(f"  DELETED company {cid}"); deleted += 1
    else:
        print(f"  [dry-run] would delete company {cid}"); deleted += 1

# 5. Fix quote status anomalies
print("\n=== Fixing quote status anomalies ===")
fixed = 0
for q in quotes:
    qid = q["id"]; qnum = q.get("quote_number",""); st = q.get("status",""); cb = q.get("claimed_by","")
    if st == "":
        print(f"  {qnum}: empty status -> 'new'")
        if APPLY:
            req("PATCH", f"/api/collections/quotes/records/{qid}", {"status":"new"}, token=token)
        fixed += 1
    elif st == "claimed" and not cb:
        print(f"  {qnum}: claimed w/o claimer -> 'new'")
        if APPLY:
            req("PATCH", f"/api/collections/quotes/records/{qid}", {"status":"new","claimed_by":""}, token=token)
        fixed += 1

# 6. Clear {0,0} coordinates on companies (null-island)
print("\n=== Clearing {0,0} coordinates ===")
cleared = 0
for c in companies:
    co = c.get("coordinates")
    if isinstance(co, dict) and co.get("lon")==0 and co.get("lat")==0:
        if APPLY:
            req("PATCH", f"/api/collections/companies/records/{c['id']}", {"coordinates":None}, token=token)
        print(f"  {c.get('company','?')}: cleared {{0,0}}")
        cleared += 1

print(f"\n=== Summary ({'APPLIED' if APPLY else 'DRY-RUN'}) ===")
print(f"  deleted: {deleted}")
print(f"  quotes fixed: {fixed}")
print(f"  coordinates cleared: {cleared}")
if not APPLY:
    print("\n  Re-run with --apply to make changes.")