"""Archive full-resolution chunks from an authorized public streamed-SOG scene."""

import argparse
import gzip
import hashlib
import json
import re
import subprocess
import sys
import urllib.request
from pathlib import Path
from urllib.parse import urljoin

parser = argparse.ArgumentParser()
parser.add_argument("--metadata-url", required=True)
parser.add_argument("--output", type=Path, required=True)
parser.add_argument("--author", required=True)
parser.add_argument("--license", required=True)
parser.add_argument("--source-url", required=True)
args = parser.parse_args()
if not args.metadata_url.startswith("https://") or not args.metadata_url.endswith("/lod-meta.json"):
    parser.error("Use an observed public HTTPS lod-meta.json URL")
args.output.mkdir(parents=True, exist_ok=False)
with urllib.request.urlopen(args.metadata_url, timeout=30) as response:
    raw = response.read(1024**2 + 1)
if len(raw) > 1024**2:
    raise ValueError("LOD manifest too large")
if raw[:2] == b"\x1f\x8b":
    raw = gzip.decompress(raw)
if len(raw) > 1024**2:
    raise ValueError("LOD manifest too large")
meta = json.loads(raw)
if meta["version"] != 1 or not 1 <= meta["counts"][0] <= 20_000_000:
    raise ValueError("Unsupported scene size or LOD version")
indices = set()


def visit(node):
    if "0" in node.get("lods", {}):
        indices.add(node["lods"]["0"]["file"])
    for child in node.get("children", []):
        visit(child)


visit(meta["tree"])
if not 1 <= len(indices) <= 100:
    raise ValueError("Unexpected full-resolution chunk count")
(args.output / "lod-meta.json").write_bytes(raw)
for index in sorted(indices):
    name = meta["filenames"][index]
    if not re.fullmatch(r"[a-zA-Z0-9_-]+/meta\.json", name):
        raise ValueError("Unexpected chunk reference")
    command = [
        sys.executable,
        str(Path(__file__).with_name("fetch_sog.py")),
        "--metadata-url",
        urljoin(args.metadata_url, name),
        "--output",
        str(args.output / Path(name).parent),
        "--author",
        args.author,
        "--license",
        args.license,
        "--source-url",
        args.source_url,
    ]
    subprocess.run(command, check=True)
manifest = {
    "source_url": args.source_url,
    "metadata_url": args.metadata_url,
    "author": args.author,
    "license": args.license,
    "lod": 0,
    "splat_count": meta["counts"][0],
    "files": [
        {
            "path": p.relative_to(args.output).as_posix(),
            "sha256": hashlib.sha256(p.read_bytes()).hexdigest(),
            "byte_length": p.stat().st_size,
        }
        for p in sorted(args.output.rglob("*"))
        if p.is_file()
    ],
}
(args.output / "source-manifest.json").write_text(json.dumps(manifest, indent=2))
print(json.dumps({"count": meta["counts"][0], "chunks": len(indices)}))