diff --git a/README.md b/README.md index 105a826..7e0fd02 100644 --- a/README.md +++ b/README.md @@ -1,26 +1,44 @@ -# Git-Analytics -[![PyPI](https://img.shields.io/pypi/v/git-analytics.svg?color=green)](https://pypi.org/project/git-analytics/) -[![Python Versions](https://img.shields.io/pypi/pyversions/git-analytics.svg)](https://pypi.org/project/git-analytics/) -[![code quality check](https://github.com/n0rfas/git-analytics/actions/workflows/code-check-dev.yml/badge.svg?branch=dev)](https://github.com/n0rfas/git-analytics/tree/dev) -[![python versions check](https://github.com/n0rfas/git-analytics/actions/workflows/python-matrix-main.yml/badge.svg?branch=main)](https://github.com/n0rfas/git-analytics/tree/main) - -The detailed analysis tool for git repositories. +The detailed analysis tool for git repositories - forked version from [here](https://github.com/n0rfas/git-analytics) with SQLite acting as the data store to provide better experience. ## Installation The latest stable version can be installed directly from PyPI: ```sh -pip install git-analytics +pip install git-analytics-sqlite ``` ## Usage -To run, enter the command and open the browser at [http://localhost:8000/](http://localhost:8000/). +Run from inside any git repository, then open [http://localhost:8000/](http://localhost:8000/) in your browser: + +```sh +git-analytics-sqlite +``` + +On first launch the dashboard will show a **Scan repository** button. Click it to parse the git history and populate the local database. Subsequent launches load instantly from the cached data. Use the **Scan** button in the header at any time to pick up new commits. + +### SQLite database + +Commit data is stored in a per-repository SQLite database under your platform's user data directory: + +| Platform | Location | +|----------|----------| +| Linux | `~/.local/share/git-analytics//_.db` | +| macOS | `~/Library/Application Support/git-analytics//_.db` | +| Windows | `%APPDATA%\git-analytics\\_.db` | + +The `` is a short fingerprint of the repository's full path, so two repositories with the same folder name never share a database. + +Use the folder icon button in the header to open the database directory in your file manager. + +### Custom database path + +Pass `--db` to override the default location: ```sh -git-analytics +git-analytics-sqlite --db /path/to/my.db ``` ## Screenshots @@ -38,7 +56,20 @@ poetry install --with dev ### Running ```bash -poetry run git-analytics +poetry run git-analytics-sqlite +``` + +### Building + +```bash +poetry build +# produces dist/git_analytics--py3-none-any.whl +``` + +Install the built wheel directly with pip: + +```bash +pip install dist/git_analytics-*.whl ``` ### Tests diff --git a/git_analytics/main.py b/git_analytics/main.py index 94b4ab0..d7bd577 100644 --- a/git_analytics/main.py +++ b/git_analytics/main.py @@ -1,8 +1,8 @@ +import argparse import os +from pathlib import Path from wsgiref.simple_server import make_server -from git import InvalidGitRepositoryError, Repo - from git_analytics.analyzers import ( AuthorsStatisticsAnalyzer, CommitsSummaryAnalyzer, @@ -12,7 +12,7 @@ LinesAnalyzer, ) from git_analytics.engine import CommitAnalyticsEngine, FileAnalyticsEngine -from git_analytics.sources import GitCommitSource +from git_analytics.sources import SqliteCommitSource, repo_db_path from git_analytics.web_app import create_web_app @@ -28,23 +28,27 @@ def make_analyzers(): def run(): - try: - path_repo = os.getenv("PATH_REPO", ".") - repo = Repo(path_repo) - name_branch = repo.active_branch.name - except InvalidGitRepositoryError: - print("Error: Current directory is not a git repository.") - return + parser = argparse.ArgumentParser(prog="git-analytics") + parser.add_argument( + "--db", + metavar="PATH", + help="path to SQLite database (default: user data directory)", + ) + args = parser.parse_args() - extension_stats = FileAnalyticsEngine().run() + path_repo = os.getenv("PATH_REPO", ".") + db_path = Path(args.db) if args.db else repo_db_path(path_repo) + + repo = _try_get_repo(path_repo) + additional_data = _build_additional_data(repo) engine = CommitAnalyticsEngine( - source=GitCommitSource(repo), + source=SqliteCommitSource(db_path), analyzers_factory=make_analyzers, - additional_data={"name_branch": name_branch, "extension_stats": extension_stats}, + additional_data=additional_data, ) - web_app = create_web_app(engine=engine) + web_app = create_web_app(engine=engine, db_path=db_path, repo=repo) with make_server("", 8000, web_app) as httpd: print("Web service started at http://localhost:8000/") @@ -58,5 +62,27 @@ def run(): print("Web service stopped") +def _try_get_repo(path: str): + try: + from git import InvalidGitRepositoryError, Repo + return Repo(path) + except Exception: + return None + + +def _build_additional_data(repo): + data = {} + if repo is not None: + try: + data["name_branch"] = repo.active_branch.name + except Exception: + pass + try: + data["extension_stats"] = FileAnalyticsEngine().run() + except Exception: + pass + return data + + if __name__ == "__main__": run() diff --git a/git_analytics/sources/__init__.py b/git_analytics/sources/__init__.py index 0aa904a..74a866d 100644 --- a/git_analytics/sources/__init__.py +++ b/git_analytics/sources/__init__.py @@ -1,7 +1,11 @@ from .git_commit_adapter import GitCommitSource from .git_log_adapter import GitLogSource +from .sqlite_commit_source import SqliteCommitSource, populate_sqlite, repo_db_path __all__ = [ "GitCommitSource", "GitLogSource", + "SqliteCommitSource", + "populate_sqlite", + "repo_db_path", ] diff --git a/git_analytics/sources/sqlite_commit_source.py b/git_analytics/sources/sqlite_commit_source.py new file mode 100644 index 0000000..ce8b1d7 --- /dev/null +++ b/git_analytics/sources/sqlite_commit_source.py @@ -0,0 +1,115 @@ +import hashlib +import os +import sqlite3 +import sys +from datetime import datetime +from pathlib import Path +from typing import Iterator + +from git_analytics.entities import AnalyticsCommit, FileChangeStats +from git_analytics.interfaces import CommitSource + + +def user_data_dir() -> Path: + if sys.platform == "win32": + base = Path(os.environ.get("APPDATA", Path.home())) + elif sys.platform == "darwin": + base = Path.home() / "Library" / "Application Support" + else: + base = Path(os.environ.get("XDG_DATA_HOME", Path.home() / ".local" / "share")) + return base / "git-analytics" + + +def repo_db_path(repo_path: str = ".") -> Path: + full_path = str(Path(repo_path).resolve()) + name = Path(full_path).name + path_hash = hashlib.sha1(full_path.encode()).hexdigest()[:8] + return user_data_dir() / name / f"{name}_{path_hash}.db" + + +_CREATE_COMMITS = """ +CREATE TABLE IF NOT EXISTS commits ( + sha TEXT PRIMARY KEY, + commit_author TEXT NOT NULL, + committed_datetime TEXT NOT NULL, + lines_insertions INTEGER NOT NULL, + lines_deletions INTEGER NOT NULL, + files_changed INTEGER NOT NULL, + message TEXT NOT NULL +) +""" + +_CREATE_COMMIT_FILES = """ +CREATE TABLE IF NOT EXISTS commit_files ( + sha TEXT NOT NULL REFERENCES commits(sha), + filepath TEXT NOT NULL, + insertions INTEGER NOT NULL, + deletions INTEGER NOT NULL +) +""" + + +def populate_sqlite(source: CommitSource, db_path: Path) -> None: + db_path.parent.mkdir(parents=True, exist_ok=True) + conn = sqlite3.connect(db_path) + try: + conn.execute(_CREATE_COMMITS) + conn.execute(_CREATE_COMMIT_FILES) + with conn: + for commit in source.iter_commits(): + conn.execute( + "INSERT OR REPLACE INTO commits VALUES (?,?,?,?,?,?,?)", + ( + commit.sha, + commit.commit_author, + commit.committed_datetime.isoformat(), + commit.lines_insertions, + commit.lines_deletions, + commit.files_changed, + commit.message, + ), + ) + for filepath, stats in commit.files.items(): + conn.execute( + "INSERT INTO commit_files VALUES (?,?,?,?)", + (commit.sha, filepath, stats.insertions, stats.deletions), + ) + finally: + conn.close() + + +class SqliteCommitSource(CommitSource): + def __init__(self, db_path: Path) -> None: + self._db_path = db_path + + def iter_commits(self) -> Iterator[AnalyticsCommit]: + conn = sqlite3.connect(self._db_path) + try: + rows = conn.execute( + "SELECT sha, commit_author, committed_datetime, lines_insertions, " + "lines_deletions, files_changed, message " + "FROM commits ORDER BY committed_datetime DESC" + ).fetchall() + + for row in rows: + sha, author, dt_str, insertions, deletions, files_changed, message = row + files_rows = conn.execute( + "SELECT filepath, insertions, deletions FROM commit_files WHERE sha=?", + (sha,), + ).fetchall() + files = { + filepath: FileChangeStats(insertions=ins, deletions=dels) + for filepath, ins, dels in files_rows + } + yield AnalyticsCommit( + sha=sha, + commit_author=author, + committed_datetime=datetime.fromisoformat(dt_str), + lines_insertions=insertions, + lines_deletions=deletions, + files_changed=files_changed, + message=message, + files=files, + ) + finally: + conn.close() diff --git a/git_analytics/static/index.html b/git_analytics/static/index.html index 1cf9fa3..9dcc57e 100644 --- a/git_analytics/static/index.html +++ b/git_analytics/static/index.html @@ -27,13 +27,21 @@ Logo

Git-Analytics

-
+
Site GitHub + +
@@ -68,6 +76,19 @@

Git-Analytics

+ +
+
+
+ +

No data yet

+

Scan your repository to populate the database and view analytics.

+ +
+
+
@@ -382,5 +403,17 @@

Commits

+ + + \ No newline at end of file diff --git a/git_analytics/static/js/git-analytics.js b/git_analytics/static/js/git-analytics.js index a02d641..5347c79 100644 --- a/git_analytics/static/js/git-analytics.js +++ b/git_analytics/static/js/git-analytics.js @@ -35,24 +35,85 @@ const BOOTSTRAP_COLORS = [ // main logic -document.addEventListener("DOMContentLoaded", () => { +document.addEventListener("DOMContentLoaded", async () => { const menu = document.getElementById("rangeMenu"); const btn = document.getElementById("rangeDropdownBtn"); - loadAndRender("months", 1, "Last month"); - menu.addEventListener("click", (e) => { const item = e.target.closest(".dropdown-item"); if (!item) return; e.preventDefault(); - const type = item.dataset.type; - const value = item.dataset.value; - const label = item.textContent.trim(); btn.textContent = item.textContent.trim(); - loadAndRender(type, value, label); + loadAndRender(item.dataset.type, item.dataset.value, item.textContent.trim()); }); + + await checkStatusAndBoot(); }); +async function checkStatusAndBoot() { + try { + const res = await fetch("/api/status"); + const status = await res.json(); + + if (status.can_scan) { + document.getElementById("scanBtn").classList.remove("d-none"); + } + + if (status.db_folder) { + const btn = document.getElementById("openFolderBtn"); + const label = document.getElementById("dbFolderLabel"); + btn.classList.remove("d-none"); + btn.title = status.db_folder; + // show only the last path segment to keep the header compact + label.textContent = status.db_folder.split(/[\\/]/).pop(); + } + + if (!status.has_data) { + document.getElementById("noDataBanner").classList.remove("d-none"); + return; + } + + loadAndRender("months", 1, "Last month"); + } catch (err) { + console.error("Failed to fetch status:", err); + } +} + +async function openDbFolder() { + try { + await fetch("/api/open-folder", { method: "POST" }); + } catch (err) { + console.error("Failed to open folder:", err); + } +} + +async function scanRepository() { + const scanBtn = document.getElementById("scanBtn"); + const scanIcon = document.getElementById("scanIcon"); + const modalEl = document.getElementById("scanningModal"); + const modal = new bootstrap.Modal(modalEl, { backdrop: "static", keyboard: false }); + + scanBtn.disabled = true; + scanIcon.className = "spinner-border spinner-border-sm"; + modal.show(); + + try { + const res = await fetch("/api/scan", { method: "POST" }); + if (!res.ok) throw new Error("Scan failed"); + + document.getElementById("noDataBanner").classList.add("d-none"); + modal.hide(); + loadAndRender("months", 1, "Last month"); + } catch (err) { + modal.hide(); + console.error("Scan error:", err); + alert("Scan failed. Check the server logs."); + } finally { + scanBtn.disabled = false; + scanIcon.className = "bi bi-arrow-repeat"; + } +} + async function loadAndRender(type, value, timeIntervalLabel) { try { // fetch stats diff --git a/git_analytics/web_app.py b/git_analytics/web_app.py index da0cf8e..c5354e4 100644 --- a/git_analytics/web_app.py +++ b/git_analytics/web_app.py @@ -1,18 +1,49 @@ import os +import sqlite3 +import subprocess +import sys from datetime import datetime +from pathlib import Path import falcon # type: ignore from git_analytics.engine import CommitAnalyticsEngine +from git_analytics.sources import GitCommitSource, populate_sqlite class GitAnalyticsResource: - def __init__(self, engine: CommitAnalyticsEngine) -> None: + def __init__(self, engine: CommitAnalyticsEngine, db_path: Path, repo=None) -> None: self._engine = engine + self._db_path = db_path + self._repo = repo def on_get_index(self, req, resp): raise falcon.HTTPMovedPermanently("index.html") + def on_get_status(self, req, resp): + resp.media = { + "has_data": self._has_data(), + "can_scan": self._repo is not None, + "db_folder": str(self._db_path.parent), + } + + def on_post_open_folder(self, req, resp): + folder = self._db_path.parent + folder.mkdir(parents=True, exist_ok=True) + if sys.platform == "win32": + subprocess.Popen(["explorer", str(folder)]) + elif sys.platform == "darwin": + subprocess.Popen(["open", str(folder)]) + else: + subprocess.Popen(["xdg-open", str(folder)]) + resp.media = {"status": "ok"} + + def on_post_scan(self, req, resp): + if self._repo is None: + raise falcon.HTTPBadRequest("scan_unavailable", "No git repository available") + populate_sqlite(GitCommitSource(self._repo), self._db_path) + resp.media = {"status": "ok", "commit_count": self._count_commits()} + def on_get_statistics(self, req, resp): start_str = req.get_param("start_date") stop_str = req.get_param("stop_date") @@ -28,14 +59,37 @@ def on_get_statistics(self, req, resp): resp.media = result + def _has_data(self) -> bool: + if not self._db_path.exists(): + return False + try: + conn = sqlite3.connect(self._db_path) + count = conn.execute("SELECT COUNT(*) FROM commits").fetchone()[0] + conn.close() + return count > 0 + except Exception: + return False + + def _count_commits(self) -> int: + try: + conn = sqlite3.connect(self._db_path) + count = conn.execute("SELECT COUNT(*) FROM commits").fetchone()[0] + conn.close() + return count + except Exception: + return 0 + -def create_web_app(engine: CommitAnalyticsEngine) -> falcon.App: +def create_web_app(engine: CommitAnalyticsEngine, db_path: Path, repo=None) -> falcon.App: app = falcon.App() - analytics_resource = GitAnalyticsResource(engine) + resource = GitAnalyticsResource(engine=engine, db_path=db_path, repo=repo) static_path = os.path.dirname(os.path.abspath(__file__)) + "/static/" app.add_static_route("/", static_path) - app.add_route("/", analytics_resource, suffix="index") - app.add_route("/api/statistics", analytics_resource, suffix="statistics") + app.add_route("/", resource, suffix="index") + app.add_route("/api/status", resource, suffix="status") + app.add_route("/api/scan", resource, suffix="scan") + app.add_route("/api/open-folder", resource, suffix="open_folder") + app.add_route("/api/statistics", resource, suffix="statistics") return app diff --git a/pyproject.toml b/pyproject.toml index e8c6d47..75293f4 100644 --- a/pyproject.toml +++ b/pyproject.toml @@ -1,5 +1,5 @@ [tool.poetry] -name = "git-analytics" +name = "git-analytics-sqlite" version = "0.1.14" description = "Advanced analytics for Git repositories — commits, authors, code churn, lines of code, trends, and visual dashboards." authors = ["n0rfas "] @@ -23,7 +23,7 @@ ruff = "^0.12" freezegun= "^1.5" [tool.poetry.scripts] -git-analytics = 'git_analytics.main:run' +git-analytics-sqlite = 'git_analytics.main:run' [build-system] requires = ["poetry-core"] diff --git a/tests/test_sqlite_source.py b/tests/test_sqlite_source.py new file mode 100644 index 0000000..ca3623f --- /dev/null +++ b/tests/test_sqlite_source.py @@ -0,0 +1,93 @@ +import tempfile +from datetime import datetime, timedelta, timezone +from pathlib import Path + +import pytest + +from git_analytics.entities import AnalyticsCommit, FileChangeStats +from git_analytics.sources import SqliteCommitSource, populate_sqlite +from tests.fakes import FAKE_COMMITS, FakeCommitSource + + +def _round_trip(commits): + with tempfile.TemporaryDirectory() as tmp: + db = Path(tmp) / "test.db" + populate_sqlite(FakeCommitSource(commits), db) + return list(SqliteCommitSource(db).iter_commits()) + + +def test_round_trip_count(): + result = _round_trip(FAKE_COMMITS) + assert len(result) == len(FAKE_COMMITS) + + +def test_round_trip_fields(): + result = _round_trip(FAKE_COMMITS) + # SqliteCommitSource yields newest-first; FAKE_COMMITS is already newest-first + original = FAKE_COMMITS[0] + loaded = result[0] + assert loaded.sha == original.sha + assert loaded.commit_author == original.commit_author + assert loaded.lines_insertions == original.lines_insertions + assert loaded.lines_deletions == original.lines_deletions + assert loaded.files_changed == original.files_changed + assert loaded.message == original.message + + +def test_round_trip_timezone_preserved(): + result = _round_trip(FAKE_COMMITS) + original = FAKE_COMMITS[0] + loaded = result[0] + assert loaded.committed_datetime == original.committed_datetime + assert loaded.committed_datetime.tzinfo is not None + + +def test_round_trip_ordering_newest_first(): + result = _round_trip(FAKE_COMMITS) + datetimes = [c.committed_datetime for c in result] + assert datetimes == sorted(datetimes, reverse=True) + + +def test_round_trip_with_files(): + tz = timezone(timedelta(hours=2)) + commits = [ + AnalyticsCommit( + sha="abc123", + commit_author="Alice", + committed_datetime=datetime(2025, 6, 1, 10, 0, 0, tzinfo=tz), + lines_insertions=10, + lines_deletions=5, + files_changed=2, + message="feat: add files", + files={ + "src/foo.py": FileChangeStats(insertions=8, deletions=3), + "src/bar.py": FileChangeStats(insertions=2, deletions=2), + }, + ) + ] + result = _round_trip(commits) + assert len(result) == 1 + loaded = result[0] + assert set(loaded.files.keys()) == {"src/foo.py", "src/bar.py"} + assert loaded.files["src/foo.py"].insertions == 8 + assert loaded.files["src/foo.py"].deletions == 3 + + +def test_round_trip_empty_files(): + result = _round_trip(FAKE_COMMITS) + # All FAKE_COMMITS have files={}, should round-trip cleanly + for commit in result: + assert commit.files == {} + + +def test_load_nonexistent_db_raises(): + source = SqliteCommitSource(Path("/nonexistent/path/db.sqlite")) + with pytest.raises(Exception): + list(source.iter_commits()) + + +def test_populate_creates_parent_dirs(): + with tempfile.TemporaryDirectory() as tmp: + db = Path(tmp) / "nested" / "dir" / "test.db" + populate_sqlite(FakeCommitSource([]), db) + assert db.exists()