"""Compare captured input maps to exact named Git trees; no builds or checkouts."""
from pathlib import Path, PurePosixPath
import argparse
import hashlib
import json
import os
import subprocess
import zipfile

HERE = Path(__file__).resolve().parent
COMMITS = {"desktop-4": "12a7352ddef5f69a144ead8219a50916233d63a9",
           "desktop-5": "f6a6010baed9e6164a1fcc7ecd58256a562827d7"}
COMPARISON = "Exact committed Git blobs equal every captured input; same complete filtered inventory"
FILES = {".gitattributes", "verify.py", "README.md", "verify-committed-inputs.py",
         "committed-desktop-4.json", "committed-desktop-5.json"}


def need(ok, message):
    if not ok:
        raise ValueError(message)


def pin(data):
    return {"bytes": len(data), "sha256": hashlib.sha256(data).hexdigest()}


def parsed(data):
    def unique(items):
        result = {}
        for key, value in items:
            need(key not in result, "duplicate JSON key: " + key)
            result[key] = value
        return result
    return json.loads(data, object_pairs_hook=unique)


def safe(name):
    return (isinstance(name, str) and bool(name) and "\\" not in name and ":" not in name
            and not PurePosixPath(name).is_absolute()
            and all(part not in ("", ".", "..") for part in name.split("/")))


def relevant(path):
    return path in ("Cargo.toml", "Cargo.lock", "rust-toolchain.toml") or path.startswith(
        ("crates/", "build/", ".cargo/", "scripts/", "docs/", "examples/", "harness/fixtures/", "harness/prototypes/"))


def git(repo, *arguments, data=None):
    environment = dict(os.environ, GIT_NO_LAZY_FETCH="1")
    return subprocess.run(["git", *arguments], cwd=repo, input=data, stdout=subprocess.PIPE,
                          stderr=subprocess.PIPE, check=True, env=environment).stdout


def compare_commit(repo, commit, inputs):
    need(git(repo, "rev-parse", "--verify", commit + "^{commit}").decode().strip() == commit, "named commit unavailable")
    entries = git(repo, "ls-tree", "-r", "--full-tree", "-z", commit).split(b"\0")
    blobs = {}
    for entry in filter(None, entries):
        header, path_bytes = entry.split(b"\t", 1)
        path = path_bytes.decode("utf-8")
        if relevant(path):
            _mode, kind, oid = header.split()
            need(kind == b"blob", "relevant non-blob tree entry: " + path)
            need(path not in blobs, "duplicate tree path")
            blobs[path] = oid
    need(set(blobs) == set(inputs), "complete filtered Git inventory differs; missing="
         + repr(sorted(set(inputs) - set(blobs))) + "; additional=" + repr(sorted(set(blobs) - set(inputs))))
    # Object IDs avoid path quoting and commit:path interpretation in batch input.
    paths = sorted(blobs)
    data = git(repo, "cat-file", "--batch", data=b"".join(blobs[path] + b"\n" for path in paths))
    offset, actual = 0, {}
    for path in paths:
        end = data.index(b"\n", offset)
        oid, kind, size_bytes = data[offset:end].split()
        size = int(size_bytes)
        need(oid == blobs[path] and kind == b"blob" and size >= 0, "Git batch object relationship")
        content = data[end + 1:end + 1 + size]
        need(len(content) == size and data[end + 1 + size:end + 2 + size] == b"\n", "Git batch framing")
        actual[path] = pin(content)
        offset = end + 2 + size
    need(offset == len(data), "trailing Git batch bytes")
    different = [path for path in paths if actual[path] != inputs[path]]
    need(not different, "captured bytes differ from named commit: " + repr(different))


def packaged(checks, runs):
    files = parsed((checks / "files.json").read_bytes())
    need(set(files) == {"prepare.py", "verify.py", "README.md", "manifest.json", "evidence.zip"}, "checks outer inventory")
    for name, expected in files.items():
        need(pin((checks / name).read_bytes()) == expected, "checks outer pin: " + name)
    manifest = parsed((checks / "manifest.json").read_bytes())
    need(manifest["schema"] == 1 and manifest["kind"] == "multiselect-cargo-gates", "checks manifest kind")
    captured = {}
    with zipfile.ZipFile(checks / "evidence.zip") as archive:
        names = archive.namelist()
        need(len(names) == len(set(names)) and set(names) == set(manifest["members"])
             and all(safe(name) for name in names), "checks exact safe members")
        def raw(logical):
            member = manifest["logicalFiles"][logical]
            data = archive.read(member)
            need(pin(data) == manifest["members"][member], "checks member pin: " + logical)
            return data
        for run in runs:
            prefix = "cargo/" + run + "/"
            result = parsed(raw(prefix + "result.json"))
            spec = manifest["runs"][run]
            need(all(result[key] == spec[key] for key in ("command", "head", "exitCode", "inputCount")), run + ": checks run relationships")
            captured[run] = (raw(prefix + "inputs.json"), result)
    return captured


def verify(repo, captured):
    outer = parsed((HERE / "files.json").read_bytes())
    need(set(outer) == FILES, "source evidence file inventory")
    for name, expected in outer.items():
        need(pin((HERE / name).read_bytes()) == expected, "source evidence pin: " + name)
    for run, commit in COMMITS.items():
        receipt = parsed((HERE / ("committed-" + run + ".json")).read_bytes())
        need(receipt["gate"] == run and receipt["commit"] == commit
             and receipt["comparison"] == COMPARISON, run + ": historical identity relationship")
        map_bytes, result = captured[run]
        inputs = parsed(map_bytes)
        need(pin(map_bytes) == receipt["inputMap"] == result["inputs"], run + ": exact captured input map")
        need(result["exitCode"] == 0 and result["inputsUnchanged"] is True, run + ": accepted unchanged gate")
        need(len(inputs) == receipt["inputCount"] == result["inputCount"], run + ": input counts")
        need(all(safe(path) and relevant(path) for path in inputs), run + ": input scope")
        need(sum(path.startswith("crates/vendor/iced_widget/") for path in inputs) == 71, run + ": vendor count")
        compare_commit(repo, commit, inputs)
        print("PASS exact Git tree/blobs:", run, commit, len(inputs), "captured inputs")


def main():
    parser = argparse.ArgumentParser(description=__doc__)
    parser.add_argument("--repo", type=Path, required=True, help="repository with both named commits already present")
    source = parser.add_mutually_exclusive_group(required=True)
    source.add_argument("--checks", type=Path, help="portable checks package directory")
    source.add_argument("--raw-gates", type=Path, help="optional preparation mode: raw gate directory, not a packaged-evidence verification")
    args = parser.parse_args()
    if args.checks:
        captured = packaged(args.checks, COMMITS)
    else:
        captured = {run: ((args.raw_gates / run / "inputs.json").read_bytes(),
                          parsed((args.raw_gates / run / "result.json").read_bytes())) for run in COMMITS}
    verify(args.repo, captured)
    print("Source identity only; no build, checkout, helper execution or native acceptance claim.")


if __name__ == "__main__":
    main()