"""Record installed public source, route branches and exact copy-ready snippets.

No network calls. All writes stay in this lab directory.
"""
import hashlib
import importlib
import importlib.metadata
import inspect
import json
import os
import platform
import runpy
import shutil
import subprocess
import sys
import tempfile
import time
from datetime import datetime
from pathlib import Path
from unittest.mock import patch

import pdfplumber

LAB = Path(__file__).resolve().parent
SOURCES = LAB / "sources"
SOURCES.mkdir(exist_ok=True)
converter = importlib.import_module("markitdown.converters._pdf_converter")
source_files = {
    "markitdown-pdf-converter.py": Path(inspect.getsourcefile(converter)),
    "pdfplumber-table.py": Path(inspect.getsourcefile(pdfplumber.table.Table)),
    "lab-score.py": LAB.parents[2] / "tools/lab_score.py",
    "lab-run.py": LAB.parents[2] / "tools/lab_run.py",
}
manifest = {"date": datetime.now().astimezone().isoformat(), "python": platform.python_version(),
            "versions": {p: importlib.metadata.version(p) for p in
                         ["pdfplumber", "pdfminer.six", "markitdown", "pymupdf4llm", "PyMuPDF", "pymupdf-layout"]},
            "source_files": [], "markitdown_branches": [], "snippet_checks": []}
for name, source in source_files.items():
    shutil.copyfile(source, SOURCES / name)
    manifest["source_files"].append({"saved": "sources/" + name,
                                     "sha256": hashlib.sha256(source.read_bytes()).hexdigest()})
for source in sorted((LAB / "corpus").glob("*.pdf")):
    with pdfplumber.open(source) as pdf:
        pages = [converter._extract_form_content_from_words(page) for page in pdf.pages]
    manifest["markitdown_branches"].append({"sample": source.stem,
        "page_form_branch": [p is not None for p in pages],
        "document_pdfminer_fallback": all(p is None for p in pages)})

EXAMPLE_FILES = {"pdfplumber": "extract-text-and-cells.py",
                 "markitdown": "convert-with-markitdown.py",
                 "pymupdf4llm": "convert-with-pymupdf4llm.py",
                 "markovo": "convert-with-markovo.py"}
# Run as standalone files in fresh interpreters, matching a downloaded example.
for tool in ["pdfplumber", "markitdown", "pymupdf4llm"]:
    with tempfile.TemporaryDirectory(prefix="snippet-", dir=LAB) as work:
        shutil.copyfile(LAB / "sample.pdf", Path(work) / "input.pdf")
        started = time.perf_counter()
        command = [sys.executable, str(LAB / "examples" / EXAMPLE_FILES[tool])]
        result = subprocess.run(command, cwd=work, capture_output=True, text=True, check=True,
                                env={**os.environ, "PYTHONDONTWRITEBYTECODE": "1"})
        (LAB / "examples" / (tool + "-sample-stdout.txt")).write_text(result.stdout + result.stderr)
        content = (Path(work) / "output.md").read_text()
        assert len(content.strip()) > 0, tool
        saved_output = LAB / "examples" / (tool + "-sample-output.md")
        saved_output.write_text(content)
        manifest["snippet_checks"].append({"tool": tool, "input": "sample.pdf",
             "result": "PASS", "output": "examples/" + saved_output.name,
             "execution": "fresh interpreter, standalone example file",
             "command": command,
             "elapsed_s": round(time.perf_counter() - started, 4),
             "output_sha256": hashlib.sha256(content.encode()).hexdigest()})

# The API example is validated against a local transport stub, never a live API.
class Response:
    def __init__(self, value):
        self.value = value if isinstance(value, bytes) else json.dumps(value).encode()
    def __enter__(self): return self
    def __exit__(self, *args): return False
    def read(self, *args): return self.value

calls = []
def stub(request, timeout):
    calls.append({"method": request.get_method(), "url": request.full_url})
    assert request.get_header("Authorization") == "Bearer offline-placeholder"
    if len(calls) == 1:
        assert request.get_method() == "POST"
        assert b'name="max_credit_units"' in request.data
        assert b'name="capability_id"' in request.data
        assert b"%PDF" in request.data
        return Response({"jobs": [{"job_id": "job_offline"}]})
    if len(calls) == 2:
        return Response({"job": {"status": "succeeded"}})
    return Response(b"# Offline transport fixture\n")

with tempfile.TemporaryDirectory(prefix="snippet-", dir=LAB) as work:
    shutil.copyfile(LAB / "corpus/table-simple.pdf", Path(work) / "input.pdf")
    old_cwd = os.getcwd()
    try:
        os.chdir(work)
        with patch.dict(os.environ, {"MARKOVO_API_KEY": "offline-placeholder", "MARKOVO_MAX_CREDIT_UNITS": "1"}), patch("urllib.request.urlopen", stub):
            runpy.run_path(str(LAB / "examples" / EXAMPLE_FILES["markovo"]), run_name="__main__")
    finally:
        os.chdir(old_cwd)
    assert (Path(work) / "output.md").read_text() == "# Offline transport fixture\n"
manifest["snippet_checks"].append({"tool": "markovo", "result": "PASS (local HTTP stub only)", "calls": calls})
(LAB / "source-diagnostics.json").write_text(json.dumps(manifest, indent=2) + "\n")
print(json.dumps(manifest, indent=2))