#!/usr/bin/env python3
"""Ledger — independent re-derivation check for the git commit census.

Re-derives commits-census.csv and commits-by-day.csv cell-by-cell from git
at HEAD (not from the committed tables), then asserts the committed tables
match exactly. Also checks commits-full.csv has one row per commit.

Exit code is the verdict:
  0  tables match git
  1  mismatch
  2  ungradable (git unreadable)
"""
from pathlib import Path
import csv
import subprocess
import sys

REPO = Path(__file__).resolve().parents[2]
OUT = Path(__file__).resolve().parents[2] / 'wizard-work' / 'ledger'

# Pinned revision (must match extract_commits_census.py PIN).
PIN = "d8d8e3292db6c435b174e39a3ca18792874f9c19"

SEP = "\x1f"
FMT = "%H" + SEP + "%an" + SEP + "%ae" + SEP + "%aI" + SEP + "%s"

CENSUS = "commits-census.csv"
BY_DAY = "commits-by-day.csv"
FULL = "commits-full.csv"


def git_log():
    cmd = ["git", "-C", REPO, "log",
           "--pretty=format:" + FMT, "--date=iso-strict", PIN]
    out = subprocess.run(cmd, check=True, capture_output=True, text=True).stdout
    rows = []
    for c in out.splitlines():
        if not c:
            continue
        h, an, ae, ad, s = c.split(SEP)
        rows.append({"hash": h, "author": an, "email": ae,
                     "author_date": ad, "subject": s})
    return rows


def read_csv(name):
    path = f"{OUT}/{name}"
    with open(path, newline="", encoding="utf-8") as f:
        return list(csv.reader(f))


def close(a, b):
    try:
        return float(a) == float(b)
    except (TypeError, ValueError):
        return a == b


def main():
    rows = git_log()
    total = len(rows)
    by_author = {}
    for r in rows:
        by_author[r["author"]] = by_author.get(r["author"], 0) + 1
    by_day = {}
    for r in rows:
        d = r["author_date"][:10]
        by_day[d] = by_day.get(d, 0) + 1

    problems = []

    # --- commits-census.csv ---
    try:
        census = read_csv(CENSUS)
    except Exception as e:
        print("UNGRADABLE: cannot read commits-census.csv:", e)
        return 2
    if census[0] != ["author", "commits"]:
        problems.append("census header: %r" % census[0])
    expected_census = sorted(by_author.items(), key=lambda kv: (-kv[1], kv[0]))
    if len(census) - 1 != len(expected_census):
        problems.append("census rows %d != expected %d"
                        % (len(census) - 1, len(expected_census)))
    for (act, exp) in zip(census[1:], expected_census):
        if not close(act[1], str(exp[1])):
            problems.append("census author %r commits %r != %r"
                            % (act[0], act[1], exp[1]))

    # --- commits-by-day.csv ---
    try:
        days = read_csv(BY_DAY)
    except Exception as e:
        print("UNGRADABLE: cannot read commits-by-day.csv:", e)
        return 2
    if days[0] != ["date", "commits"]:
        problems.append("by-day header: %r" % days[0])
    expected_days = sorted(by_day.items())
    if len(days) - 1 != len(expected_days):
        problems.append("by-day rows %d != expected %d"
                        % (len(days) - 1, len(expected_days)))
    for (act, exp) in zip(days[1:], expected_days):
        if not close(act[1], str(exp[1])):
            problems.append("by-day %s commits %r != %r"
                            % (act[0], act[1], exp[1]))

    # --- commits-full.csv row count ---
    try:
        full = read_csv(FULL)
    except Exception as e:
        print("UNGRADABLE: cannot read commits-full.csv:", e)
        return 2
    if len(full) - 1 != total:
        problems.append("full rows %d != commit count %d" % (len(full) - 1, total))

    if problems:
        print("FAIL (%d problems)" % len(problems))
        for p in problems[:30]:
            print(" -", p)
        return 1

    print("PASS: commits-census.csv and commits-by-day.csv match git at HEAD; "
          "commits-full.csv has %d rows; total %d commits" % (len(full) - 1, total))
    return 0


if __name__ == "__main__":
    sys.exit(main())