new-site/scripts/formation/test_name_search.py

#!/usr/bin/env python3
"""
Test name search automation for any state.

Usage:
  python -m scripts.formation.test_name_search WY "Test Business LLC"
  python -m scripts.formation.test_name_search CO "Acme Holdings"
  python -m scripts.formation.test_name_search DE "First State Corp"
  python -m scripts.formation.test_name_search --all "Test Business LLC"

Reports:
  - Whether name search succeeded
  - Whether the portal was reachable
  - What selectors worked
  - What barriers were encountered (CAPTCHA, login, WAF, etc.)
"""

from __future__ import annotations

import asyncio
import json
import logging
import sys
import time
from datetime import datetime

logging.basicConfig(
    level=logging.INFO,
    format="%(asctime)s [%(name)s] %(levelname)s %(message)s",
)
LOG = logging.getLogger("test_name_search")


async def test_state(state_code: str, name: str) -> dict:
    """Test name search for a single state."""
    from scripts.formation.states import get_adapter, get_config, STATES

    state_code = state_code.upper()
    if state_code not in STATES:
        return {"state": state_code, "status": "ERROR", "error": f"Unknown state: {state_code}"}

    config = get_config(state_code)
    state_name = STATES[state_code]["name"]
    search_method = config.get("search_method", "playwright")

    LOG.info("Testing %s (%s) — method: %s, name: '%s'", state_code, state_name, search_method, name)
    start = time.time()

    result = {
        "state": state_code,
        "state_name": state_name,
        "search_method": search_method,
        "portal_url": config.get("name_search_url", ""),
        "test_name": name,
        "started_at": datetime.utcnow().isoformat(),
    }

    try:
        adapter = get_adapter(state_code)
        search_result = await adapter.search_name(name)

        result["status"] = "SUCCESS" if search_result.searched_name else "PARTIAL"
        result["available"] = search_result.available
        result["exact_match"] = search_result.exact_match
        result["similar_names"] = search_result.similar_names[:5]
        result["raw_response_length"] = len(search_result.raw_response)

        if "not yet configured" in search_result.raw_response.lower():
            result["status"] = "NOT_CONFIGURED"
            result["barrier"] = "Selectors not configured in config.py"
        elif "captcha" in search_result.raw_response.lower():
            result["status"] = "BLOCKED_CAPTCHA"
            result["barrier"] = "CAPTCHA detected"
        elif "error" in search_result.raw_response.lower() and not search_result.available:
            result["status"] = "ERROR"
            result["error"] = search_result.raw_response[:200]

    except Exception as e:
        result["status"] = "ERROR"
        result["error"] = str(e)
        LOG.error("Test failed for %s: %s", state_code, e)

    elapsed = time.time() - start
    result["elapsed_seconds"] = round(elapsed, 2)
    result["finished_at"] = datetime.utcnow().isoformat()

    # Status emoji for readability
    status_icon = {
        "SUCCESS": "✓",
        "PARTIAL": "~",
        "NOT_CONFIGURED": "○",
        "BLOCKED_CAPTCHA": "⚠",
        "BLOCKED_LOGIN": "🔒",
        "ERROR": "✗",
    }.get(result["status"], "?")

    LOG.info(
        "%s %s (%s): %s — available=%s, elapsed=%.1fs",
        status_icon, state_code, state_name, result["status"],
        result.get("available", "?"), elapsed,
    )

    return result


async def test_all_states(name: str) -> list[dict]:
    """Test name search for all 51 states sequentially."""
    from scripts.formation.states import STATES

    results = []
    for code in sorted(STATES.keys()):
        result = await test_state(code, name)
        results.append(result)

        # Brief pause between states
        if result["status"] not in ("NOT_CONFIGURED", "ERROR"):
            await asyncio.sleep(2)

    return results


def print_summary(results: list[dict]):
    """Print a summary table of test results."""
    print("\n" + "=" * 80)
    print("STATE NAME SEARCH TEST RESULTS")
    print("=" * 80)
    print(f"{'State':<6} {'Name':<25} {'Status':<20} {'Available':<10} {'Time':<8} {'Barrier'}")
    print("-" * 80)

    for r in results:
        barrier = r.get("barrier", r.get("error", ""))[:30]
        print(
            f"{r['state']:<6} {r.get('state_name','?'):<25} {r['status']:<20} "
            f"{str(r.get('available','?')):<10} {r.get('elapsed_seconds',0):>6.1f}s  {barrier}"
        )

    # Summary counts
    statuses = {}
    for r in results:
        s = r["status"]
        statuses[s] = statuses.get(s, 0) + 1

    print("-" * 80)
    print("Summary:", " | ".join(f"{k}: {v}" for k, v in sorted(statuses.items())))
    print(f"Total: {len(results)} states tested")


def main():
    args = sys.argv[1:]

    if len(args) < 2:
        print("Usage:")
        print("  python -m scripts.formation.test_name_search WY 'Test Business LLC'")
        print("  python -m scripts.formation.test_name_search --all 'Test Business LLC'")
        sys.exit(1)

    if args[0] == "--all":
        name = " ".join(args[1:])
        results = asyncio.run(test_all_states(name))
        print_summary(results)

        # Save results to JSON
        output_path = "/tmp/name_search_test_results.json"
        with open(output_path, "w") as f:
            json.dump(results, f, indent=2)
        print(f"\nFull results saved to: {output_path}")
    else:
        state_code = args[0]
        name = " ".join(args[1:])
        result = asyncio.run(test_state(state_code, name))
        print(json.dumps(result, indent=2))


if __name__ == "__main__":
    main()