git_candlestick.py

#!/usr/bin/env python3
"""
Git Commit Candlestick Chart
============================
コミット履歴をローソク足チャートで可視化する

各コミットを1本のローソク足として表現:
- 始値 (Open): コミット前の総行数
- 終値 (Close): コミット後の総行数
- 高値 (High): 始値 + 追加行数
- 安値 (Low): 始値 - 削除行数
"""

import argparse
import subprocess
import re
import fnmatch
from datetime import datetime
from pathlib import Path

import pandas as pd
import plotly.graph_objects as go


def run_git(repo_path: str, *args: str) -> str:
    """gitコマンドを実行"""
    result = subprocess.run(
        ["git", "-C", repo_path, *args],
        capture_output=True,
        text=True,
        check=True,
    )
    return result.stdout


def get_commits(
    repo_path: str,
    branch: str | None = None,
    since: str | None = None,
    until: str | None = None,
) -> list[dict]:
    """コミット一覧を取得（古い順）"""
    args = [
        "log",
        "--reverse",
        "--format=%H|%s|%ai|%an",
    ]
    if branch:
        args.append(branch)
    if since:
        args.append(f"--since={since}")
    if until:
        args.append(f"--until={until}")

    output = run_git(repo_path, *args)
    commits = []
    for line in output.strip().split("\n"):
        if not line:
            continue
        parts = line.split("|", 3)
        if len(parts) >= 4:
            commits.append({
                "hash": parts[0],
                "message": parts[1],
                "date": parts[2],
                "author": parts[3],
            })
    return commits


def get_commit_stats(
    repo_path: str,
    commit_hash: str,
    file_pattern: str | None = None,
) -> tuple[int, int]:
    """コミットの追加/削除行数を取得"""
    try:
        output = run_git(repo_path, "show", "--numstat", "--format=", commit_hash)
    except subprocess.CalledProcessError:
        return 0, 0

    added = 0
    deleted = 0

    for line in output.strip().split("\n"):
        if not line:
            continue
        parts = line.split("\t")
        if len(parts) < 3:
            continue

        # バイナリファイルは "-" になる
        if parts[0] == "-" or parts[1] == "-":
            continue

        file_path = parts[2]

        # ファイルパターンでフィルタ
        if file_pattern and not fnmatch.fnmatch(file_path, file_pattern):
            continue

        try:
            added += int(parts[0])
            deleted += int(parts[1])
        except ValueError:
            continue

    return added, deleted


def build_candlestick_data(
    repo_path: str,
    commits: list[dict],
    file_pattern: str | None = None,
) -> pd.DataFrame:
    """ローソク足データを構築"""
    data = []
    total_lines = 0

    for i, commit in enumerate(commits):
        added, deleted = get_commit_stats(repo_path, commit["hash"], file_pattern)

        open_val = total_lines
        close_val = total_lines + added - deleted
        high_val = total_lines + added
        low_val = total_lines - deleted

        data.append({
            "index": i,
            "hash": commit["hash"][:8],
            "full_hash": commit["hash"],
            "message": commit["message"][:50],
            "full_message": commit["message"],
            "date": commit["date"],
            "author": commit["author"],
            "open": open_val,
            "high": high_val,
            "low": low_val,
            "close": close_val,
            "added": added,
            "deleted": deleted,
        })

        total_lines = close_val

        # 進捗表示
        if (i + 1) % 50 == 0 or i + 1 == len(commits):
            print(f"Processing: {i + 1}/{len(commits)} commits")

    return pd.DataFrame(data)


def create_chart(
    df: pd.DataFrame,
    title: str = "Git Commit Candlestick Chart",
    hide_message: bool = False,
):
    """ローソク足チャートを作成"""
    # ホバーテキストを作成
    hover_texts = []
    for _, row in df.iterrows():
        if hide_message:
            text = (
                f"<b>{row['hash']}</b><br>"
                f"Date: {row['date']}<br>"
                f"─────────────<br>"
                f"Added: +{row['added']}<br>"
                f"Deleted: -{row['deleted']}<br>"
                f"Net: {row['close'] - row['open']:+d}<br>"
                f"─────────────<br>"
                f"Total: {row['open']} → {row['close']}"
            )
        else:
            text = (
                f"<b>{row['hash']}</b><br>"
                f"{row['full_message']}<br>"
                f"─────────────<br>"
                f"Date: {row['date']}<br>"
                f"Author: {row['author']}<br>"
                f"─────────────<br>"
                f"Added: +{row['added']}<br>"
                f"Deleted: -{row['deleted']}<br>"
                f"Net: {row['close'] - row['open']:+d}<br>"
                f"─────────────<br>"
                f"Total: {row['open']} → {row['close']}"
            )
        hover_texts.append(text)

    # 色を決定（陽線=緑、陰線=赤）
    colors = ["#26a69a" if row["close"] >= row["open"] else "#ef5350" for _, row in df.iterrows()]

    fig = go.Figure()

    # ローソク足を描画
    fig.add_trace(go.Candlestick(
        x=df["index"],
        open=df["open"],
        high=df["high"],
        low=df["low"],
        close=df["close"],
        increasing=dict(line=dict(color="#26a69a"), fillcolor="#26a69a"),
        decreasing=dict(line=dict(color="#ef5350"), fillcolor="#ef5350"),
        hoverinfo="text",
        hovertext=hover_texts,
        name="Commits",
    ))

    # レイアウト設定
    fig.update_layout(
        title=dict(
            text=title,
            font=dict(size=20),
        ),
        xaxis=dict(
            title="Commit (chronological order)",
            rangeslider=dict(visible=True, thickness=0.05),
            tickmode="linear",
            dtick=max(1, len(df) // 20),
        ),
        yaxis=dict(
            title="Lines of Code",
            gridcolor="rgba(128, 128, 128, 0.2)",
        ),
        plot_bgcolor="white",
        hovermode="x unified",
        showlegend=False,
    )

    # X軸のラベルをコミットハッシュに
    fig.update_xaxes(
        ticktext=df["hash"].tolist(),
        tickvals=df["index"].tolist(),
    )

    return fig


def main():
    parser = argparse.ArgumentParser(
        description="Git Commit Candlestick Chart - コミット履歴をローソク足で可視化"
    )
    parser.add_argument(
        "repo_path",
        nargs="?",
        default=".",
        help="Gitリポジトリのパス (default: current directory)",
    )
    parser.add_argument(
        "-b", "--branch",
        help="対象ブランチ (default: current branch)",
    )
    parser.add_argument(
        "-s", "--since",
        help="開始日 (e.g., '2024-01-01', '1 month ago')",
    )
    parser.add_argument(
        "-u", "--until",
        help="終了日 (e.g., '2024-12-31', 'yesterday')",
    )
    parser.add_argument(
        "-p", "--pattern",
        help="ファイルパターン (e.g., '*.py', 'src/*.ts')",
    )
    parser.add_argument(
        "-o", "--output",
        help="出力HTMLファイル (指定しない場合はブラウザで表示)",
    )
    parser.add_argument(
        "-n", "--limit",
        type=int,
        help="最新N件のコミットのみ",
    )
    parser.add_argument(
        "--hide-message",
        action="store_true",
        help="コミットメッセージと作者を非表示",
    )
    parser.add_argument(
        "--cdn",
        action="store_true",
        help="plotly.jsをCDNから読み込む（HTMLファイル軽量化）",
    )

    args = parser.parse_args()

    repo_path = str(Path(args.repo_path).resolve())

    # リポジトリ確認
    try:
        run_git(repo_path, "rev-parse", "--git-dir")
    except subprocess.CalledProcessError:
        print(f"Error: '{repo_path}' is not a git repository")
        return 1

    print(f"Repository: {repo_path}")
    if args.branch:
        print(f"Branch: {args.branch}")
    if args.pattern:
        print(f"File pattern: {args.pattern}")
    if args.since:
        print(f"Since: {args.since}")
    if args.until:
        print(f"Until: {args.until}")

    # コミット取得
    print("\nFetching commits...")
    commits = get_commits(repo_path, args.branch, args.since, args.until)

    if not commits:
        print("No commits found")
        return 1

    if args.limit:
        commits = commits[-args.limit:]

    print(f"Found {len(commits)} commits")

    # ローソク足データ構築
    print("\nBuilding candlestick data...")
    df = build_candlestick_data(repo_path, commits, args.pattern)

    if df.empty:
        print("No data to display")
        return 1

    # チャート作成
    title = f"Git Candlestick: {Path(repo_path).name}"
    if args.pattern:
        title += f" ({args.pattern})"

    fig = create_chart(df, title, hide_message=args.hide_message)

    # 出力
    if args.output:
        # CDNオプション: plotly.jsをCDNから読み込み（約3MB→数KB）
        include_plotlyjs = "cdn" if args.cdn else True
        fig.write_html(
            args.output,
            include_plotlyjs=include_plotlyjs,
            full_html=True,
            config={
                "displayModeBar": True,
                "responsive": True,
            },
        )
        print(f"\nChart saved to: {args.output}")
        if args.cdn:
            print("(Using CDN for plotly.js - requires internet connection)")
    else:
        print("\nOpening chart in browser...")
        fig.show()

    # サマリー
    print(f"\n=== Summary ===")
    print(f"Total commits: {len(df)}")
    print(f"Initial lines: {df.iloc[0]['open']}")
    print(f"Final lines: {df.iloc[-1]['close']}")
    print(f"Net change: {df.iloc[-1]['close'] - df.iloc[0]['open']:+d}")

    return 0


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