#!/usr/bin/env python3
"""Derivation core for the wizard-work git worktree census (extract lane).

Shared `derive()` — the single source of truth for every census value.
Used by the extractor runner AND the verifier so both read the live git tree;
nothing is hardcoded, nothing is a stub. Exit-gated by the CHECK.
"""
import subprocess
from collections import Counter

REPO = "/root/troupe-standup"


def git(*args):
    return subprocess.run(
        ["git", "-C", REPO, *args], capture_output=True, text=True, check=True
    ).stdout


def derive():
    data = {}
    data["stamp"] = "2026-11-02"
    data["repo"] = REPO
    data["branch"] = git("rev-parse", "--abbrev-ref", "HEAD").strip()
    data["branch_head"] = git("rev-parse", "--short", "HEAD").strip()
    data["head"] = git("rev-parse", "HEAD").strip()
    data["commit_count"] = int(git("rev-list", "--count", "HEAD").strip())
    data["tracked_file_count"] = len(git("ls-files").splitlines())
    data["total_unique_emails"] = len({e for e in git("log", "--format=%ae").splitlines() if e})
    data["ahead_by_origin_main"] = int(git("rev-list", "--count", "origin/main..HEAD").strip())
    data["behind_by_origin_main"] = int(git("rev-list", "--count", "HEAD..origin/main").strip())
    data["remote_heads"] = {
        l: git("rev-parse", "--short", l).strip()
        for l in git("branch", "-r", "--format=%(refname:short)").splitlines()
        if l.strip()
    }

    # Author census: group commits by (name, email) with first/last date.
    by_pair = {}
    for line in git("log", "--format=%an%x00%ae%x00%ad", "--date=iso-strict").splitlines():
        if not line.strip():
            continue
        name, email, date = line.split("\x00")
        by_pair.setdefault((name, email), []).append(date)
    rows = []
    for (name, email), dates in by_pair.items():
        rows.append({"name": name, "email": email, "count": len(dates), "first": min(dates), "last": max(dates)})
    nc = Counter(r["name"] for r in rows)
    ec = Counter(r["email"] for r in rows)
    for r in rows:
        r["name_collides"] = nc[r["name"]] > 1
        r["email_collides"] = ec[r["email"]] > 1
    rows.sort(key=lambda r: (-r["count"], r["name"].lower(), r["email"]))
    data["author_census"] = rows

    # Top-modified files: commit count per path from --name-only.
    names = git("log", "--pretty=format:", "--name-only").splitlines()
    fc = Counter(n for n in names if n.strip())
    data["top_modified_files"] = [{"path": p, "commits": c} for p, c in fc.most_common()]
    return data