tgit/app/repo/git.py

"""Local clone of the uploads repo and read-only queries against `origin/<branch>` (ARCHITECTURE §2.3).

git runs as a subprocess with an argument list — never through a shell — and with the bot's own deploy key.
"""

import asyncio
import base64
import logging
import os
import tempfile
import time
from dataclasses import dataclass
from datetime import datetime
from pathlib import Path
from typing import Literal

from app.config import GitSettings

log = logging.getLogger(__name__)

EntryKind = Literal["dir", "file"]


class GitError(Exception):
    pass


@dataclass(frozen=True)
class Entry:
    path: str  # repo-relative, "/"-separated
    kind: EntryKind
    size: int  # bytes; 0 for directories

    @property
    def name(self) -> str:
        return self.path.rsplit("/", 1)[-1]


@dataclass(frozen=True)
class CommitInfo:
    sha: str
    author: str
    date: datetime
    subject: str


class UploadsRepo:
    def __init__(self, settings: GitSettings) -> None:
        self._s = settings
        # absolute: git clone runs from the parent dir, a relative path would land one level deeper
        self._dir = settings.clone_dir.resolve()
        self._lock = asyncio.Lock()
        self._fetched_at: float | None = None
        self._key_path = self._materialize_key()

    @property
    def branch(self) -> str:
        return self._s.branch

    @property
    def ref(self) -> str:
        return f"origin/{self._s.branch}"

    @property
    def path(self) -> Path:
        return self._dir

    def _materialize_key(self) -> Path | None:
        if self._s.ssh_key_b64 is None:
            return self._s.ssh_key_path
        key = base64.b64decode(self._s.ssh_key_b64.get_secret_value())
        fd, name = tempfile.mkstemp(prefix="tgit-ssh-")  # created 0600, owned by the bot's user
        with os.fdopen(fd, "wb") as f:
            f.write(key if key.endswith(b"\n") else key + b"\n")
        return Path(name)

    def _env(self) -> dict[str, str]:
        env = {k: v for k, v in os.environ.items() if k in ("PATH", "HOME", "LANG", "LC_ALL", "TMPDIR")}
        env["GIT_TERMINAL_PROMPT"] = "0"
        if self._key_path is not None:
            known_hosts = self._dir.parent / "known_hosts"
            env["GIT_SSH_COMMAND"] = (
                # -F /dev/null: ignore the host's ~/.ssh/config — otherwise a personal key configured there for the
                # same host gets offered too and silently masks a broken deploy key (happened in #7).
                f"ssh -F /dev/null -i {self._key_path} -o IdentitiesOnly=yes -o BatchMode=yes "
                f"-o StrictHostKeyChecking=accept-new -o UserKnownHostsFile={known_hosts}"
            )
        return env

    async def git(self, *args: str, cwd: Path | None = None) -> bytes:
        proc = await asyncio.create_subprocess_exec(
            "git",
            *args,
            cwd=cwd or self._dir,
            env=self._env(),
            stdout=asyncio.subprocess.PIPE,
            stderr=asyncio.subprocess.PIPE,
        )
        try:
            out, err = await asyncio.wait_for(proc.communicate(), timeout=self._s.timeout)
        except TimeoutError:
            proc.kill()
            await proc.wait()
            raise GitError(f"git {args[0]}: timeout") from None
        if proc.returncode != 0:
            raise GitError(f"git {args[0]}: {err.decode(errors='replace').strip()[-500:]}")
        return out

    async def ensure(self) -> None:
        """Clone on first use; afterwards just make sure the clone is there."""
        async with self._lock:
            if (self._dir / ".git").is_dir():
                return
            self._dir.parent.mkdir(parents=True, exist_ok=True)
            log.info("cloning uploads repo into %s", self._dir)
            await self.git(
                "clone", "--branch", self._s.branch, self._s.uploads_url, str(self._dir), cwd=self._dir.parent
            )
            self._fetched_at = time.monotonic()

    async def refresh(self, force: bool = False) -> None:
        """git fetch, but not more often than fetch_interval unless forced."""
        await self.ensure()
        async with self._lock:
            now = time.monotonic()
            if not force and self._fetched_at is not None and now - self._fetched_at < self._s.fetch_interval:
                return
            await self.git("fetch", "--prune", "origin")
            self._fetched_at = now

    async def head(self) -> str:
        return (await self.git("rev-parse", self.ref)).decode().strip()

    async def list_dir(self, path: str = "") -> list[Entry]:
        """Directories first, then files, each alphabetically."""
        spec = f"{self.ref}:{path}" if path else self.ref
        try:
            out = await self.git("ls-tree", "-z", "-l", spec)
        except GitError as e:
            raise GitError(f"no such directory: /{path}") from e
        entries: list[Entry] = []
        for record in out.split(b"\0"):
            if not record:
                continue
            meta, name = record.split(b"\t", 1)
            _mode, kind, _obj, size = meta.split()
            full = f"{path}/{name.decode()}" if path else name.decode()
            if kind == b"tree":
                entries.append(Entry(full, "dir", 0))
            elif kind == b"blob":
                entries.append(Entry(full, "file", int(size)))
        return sorted(entries, key=lambda e: (e.kind != "dir", e.name.lower()))

    async def read(self, path: str) -> bytes:
        return await self.git("cat-file", "blob", f"{self.ref}:{path}")

    async def size(self, path: str) -> int | None:
        try:
            return int((await self.git("cat-file", "-s", f"{self.ref}:{path}")).decode().strip())
        except GitError:
            return None

    async def last_commit(self, path: str) -> CommitInfo | None:
        out = await self.git("log", "-1", "--format=%H%x00%an%x00%aI%x00%s", self.ref, "--", path)
        if not out.strip():
            return None
        sha, author, date, subject = out.decode().rstrip("\n").split("\0", 3)
        return CommitInfo(sha=sha, author=author, date=datetime.fromisoformat(date), subject=subject)