#!/usr/bin/env python3
"""Extract the Boti Education app API endpoints from decompiled Capacitor JS chunks.

Usage: python3 extract_endpoints.py <dir-with-chunks> [--markdown]

Scans *.js for the Angular ApiService call signature:
    <http>.post(<payload>, "endpoint")     /     <http>.get(<payload>, "endpoint")
and emits a deduplicated endpoint -> methods table.
"""
import re
import sys
import glob
import os
import argparse

CALL = re.compile(
    r'\.(post|get|getBlob)\(\s*'
    r'(?:"[^"]*"|`[^`]*`|[A-Za-z_$][\w$]*(?:\s*\.\s*[\w$]+)*(?:\([^()]*\))?|\{[^{}]*\})'
    r'\s*,\s*"([^"]*)"'
)


def scan(directory):
    found = {}  # endpoint -> {method: [chunks]}
    for path in sorted(glob.glob(os.path.join(directory, '*.js'))):
        chunk = os.path.basename(path)
        src = open(path, encoding='utf8', errors='replace').read()
        for m in CALL.finditer(src):
            method, ep = m.group(1), m.group(2)
            d = found.setdefault(ep, {})
            d.setdefault(method, []).append(chunk)
    return found


def main():
    ap = argparse.ArgumentParser()
    ap.add_argument('directory')
    ap.add_argument('--markdown', action='store_true')
    args = ap.parse_args()

    found = scan(args.directory)
    if not found:
        sys.exit('no endpoints found, wrong directory?')

    if args.markdown:
        print(f'# API endpoints extracted from {os.path.basename(args.directory)}')
        print()
        print(f'{len(found)} unique endpoints.')
        print()
        print('| endpoint | methods | seen in chunk(s) |')
        print('|---|---|---|')
        for ep in sorted(found):
            methods = ', '.join(sorted(found[ep]))
            chunks = ', '.join(sorted({c for ms in found[ep].values() for c in ms})[:3])
            print(f'| {ep} | {methods} | {chunks} |')
    else:
        for ep in sorted(found):
            print(ep, sorted(found[ep]))
        print(f'TOTAL endpoints: {len(found)}', file=sys.stderr)


if __name__ == '__main__':
    main()