"""Build data/timings.csv and data/paths.xyz.gz from time_geodesic.py output.

    python reproduce/build_tables.py OUTDIR      # OUTDIR holds times_{before,after}{1,2,3}.csv, paths_*.xyz.gz

Each path's sha256 is over its text block in paths_*.xyz.gz (all images, repr floats),
so equal hashes mean bit-identical coordinates.
"""
import csv, gzip, hashlib, io, shutil, sys
from pathlib import Path

HERE = Path(__file__).resolve().parent.parent


def path_hashes(fn):
    blocks = {}
    lines = gzip.open(fn, "rt").read().splitlines()
    i = 0
    while i < len(lines):
        n = int(lines[i]); head = lines[i + 1].split()
        key = (head[0], head[1].split("=")[1])
        blocks.setdefault(key, []).extend(lines[i:i + 2 + n])
        i += 2 + n
    return {k: hashlib.sha256("\n".join(v).encode()).hexdigest() for k, v in blocks.items()}


def main():
    src = Path(sys.argv[1])
    rows = []
    for version in ("before", "after"):
        for rnd in (1, 2, 3):
            h = path_hashes(src / f"paths_{version}{rnd}.xyz.gz")
            for r in csv.DictReader(open(src / f"times_{version}{rnd}.csv")):
                rows.append({"idx": r["idx"], "nimages": r["nimages"], "version": version, "round": rnd,
                             "seconds": f"{float(r['seconds']):.4f}", "path_sha256": h[(r["idx"], r["nimages"])]})
    (HERE / "data").mkdir(exist_ok=True)
    with open(HERE / "data" / "timings.csv", "w", encoding="utf-8", newline="") as out:
        w = csv.DictWriter(out, fieldnames=list(rows[0])); w.writeheader(); w.writerows(rows)
    with gzip.open(src / "paths_after1.xyz.gz", "rt") as fi, \
            io.TextIOWrapper(gzip.GzipFile(HERE / "data" / "paths.xyz.gz", "wb", compresslevel=9, mtime=0),
                             encoding="utf-8") as fo:
        shutil.copyfileobj(fi, fo)
    print(len(rows), "timing rows")


if __name__ == "__main__":
    main()