@codebend3r/che 1.1.0 → 1.2.1

package/bin/machine.py

@@ -0,0 +1,345 @@
+"""Facts about the machine che is installed on.
+
+Where che keeps its config, which shells exist and which rc files each one
+reads, which python can run the scripts, which external tools are missing, and
+whether this copy of che is a git clone. The installer, the doctor and the
+dispatcher all ask these questions; the install record is the only thing
+written from here.
+
+Every path is resolved from the environment at call time, never at import
+time, so the tests can point HOME at a temporary directory.
+
+Library module: no shebang, not executable.
+"""
+
+from __future__ import annotations
+
+import json
+import os
+import shutil
+import subprocess
+import sys
+from collections.abc import Callable
+from dataclasses import asdict, dataclass, field, fields, replace
+from pathlib import Path
+
+from commands import PACKAGE_NAME, REPO, required_binaries
+from utils import atomic_write_text, run
+
+# ---------------------------------------------------------------------------
+# External tools
+# ---------------------------------------------------------------------------
+
+# Binary -> the Homebrew formula that provides it, which is what
+# `che install --install-deps` installs.
+BREW_FORMULA: dict[str, str] = {
+    "ffmpeg": "ffmpeg",
+    "ffprobe": "ffmpeg",
+    "mpv": "mpv",
+    "exiftool": "exiftool",
+    "gh": "gh",
+    "btop": "btop",
+    "git": "git",
+}
+
+# Binary -> how to get it. Anything absent from here is either part of macOS or
+# has no obvious one-liner, and doctor says so rather than inventing a command.
+INSTALL_HINTS: dict[str, str] = {
+    **{binary: f"brew install {formula}" for binary, formula in BREW_FORMULA.items()},
+    "brew": 'bash -c "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/HEAD/install.sh)"',
+}
+
+# Binaries that ship with macOS: never suggest installing them, just report.
+SYSTEM_BINARIES = frozenset({"diskutil", "osascript"})
+
+
+def missing_binaries() -> list[str]:
+    """Tools some command needs that are not on PATH, in name order, macOS's own aside."""
+    return [
+        binary
+        for binary in required_binaries()
+        if binary not in SYSTEM_BINARIES and not shutil.which(binary)
+    ]
+
+
+# ---------------------------------------------------------------------------
+# Locations
+# ---------------------------------------------------------------------------
+
+
+def home() -> Path:
+    return Path(os.environ.get("HOME", str(Path.home())))
+
+
+def xdg_config_home() -> Path:
+    return Path(os.environ.get("XDG_CONFIG_HOME") or home() / ".config")
+
+
+def config_dir() -> Path:
+    return xdg_config_home() / "che"
+
+
+def config_file() -> Path:
+    return config_dir() / "config.json"
+
+
+def backup_dir() -> Path:
+    return config_dir() / "backups"
+
+
+def state_dir() -> Path:
+    return Path(os.environ.get("XDG_STATE_HOME") or home() / ".local" / "state") / "che"
+
+
+def shim_path() -> Path:
+    return home() / ".local" / "bin" / "che"
+
+
+def shim_on_path() -> bool:
+    entries = os.environ.get("PATH", "").split(os.pathsep)
+    return str(shim_path().parent) in entries
+
+
+def tildify(path: Path) -> str:
+    """``path`` for display, with the home directory spelled ``~``."""
+    try:
+        return f"~/{path.relative_to(home())}"
+    except ValueError:
+        return str(path)
+
+
+# ---------------------------------------------------------------------------
+# Shells
+# ---------------------------------------------------------------------------
+
+
+def _bash_rc_files() -> list[Path]:
+    """Which bash startup files to write.
+
+    macOS Terminal and iTerm start bash as a *login* shell, which reads
+    ``.bash_profile`` and never ``.bashrc``; most Linux terminals do the
+    opposite. Writing to whichever already exist (and to ``.bashrc`` when
+    neither does) covers both without the usual "why isn't my alias there"
+    hunt. The block is idempotent, so a ``.bash_profile`` that sources
+    ``.bashrc`` just defines the same functions twice, harmlessly.
+    """
+    candidates = [home() / ".bashrc", home() / ".bash_profile"]
+    existing = [path for path in candidates if path.exists()]
+    return existing or [home() / ".bashrc"]
+
+
+def _zsh_rc_files() -> list[Path]:
+    zdotdir = os.environ.get("ZDOTDIR")
+    base = Path(zdotdir) if zdotdir else home()
+    return [base / ".zshrc"]
+
+
+def _fish_rc_files() -> list[Path]:
+    return [xdg_config_home() / "fish" / "config.fish"]
+
+
+def _sh_rc_files() -> list[Path]:
+    return [home() / ".profile"]
+
+
+def login_shell() -> str:
+    return Path(os.environ.get("SHELL", "")).name
+
+
+@dataclass(frozen=True)
+class ShellSpec:
+    """A shell che can install into, how, and how it stands on this machine.
+
+    The properties are read from the environment each time they are asked, so
+    a test that moves HOME or ZDOTDIR sees the move.
+    """
+
+    name: str
+    find_rc_files: Callable[[], list[Path]]
+    # The generated file the block sources, `shell/che.<flavor>`, and so the
+    # syntax the block itself is written in.
+    flavor: str
+    # Never installed into unless asked for, and a missing block only warns.
+    opt_in: bool = False
+
+    @property
+    def rc_files(self) -> list[Path]:
+        """The rc files an install into this shell writes."""
+        return self.find_rc_files()
+
+    @property
+    def installed(self) -> bool:
+        """Whether the shell itself exists on this machine."""
+        return shutil.which(self.name) is not None
+
+    @property
+    def has_rc(self) -> bool:
+        return any(path.exists() for path in self.rc_files)
+
+    @property
+    def is_login_shell(self) -> bool:
+        return self.name == login_shell()
+
+
+SHELLS: dict[str, ShellSpec] = {
+    spec.name: spec
+    for spec in (
+        ShellSpec("zsh", _zsh_rc_files, "zsh"),
+        ShellSpec("bash", _bash_rc_files, "bash"),
+        ShellSpec("fish", _fish_rc_files, "fish"),
+        # `sh` reuses the bash wrapper file: it is plain POSIX shell apart from
+        # the completion tail, which is guarded by a $BASH_VERSION test. Opt-in,
+        # because `.profile` is read by every POSIX shell on the machine, which
+        # is rarely what someone means.
+        ShellSpec("sh", _sh_rc_files, "bash", opt_in=True),
+    )
+}
+
+
+def default_shell_selection() -> list[str]:
+    """What to install into when the user does not say.
+
+    The login shell always, plus any other shell that is both installed and
+    already has an rc file - those are shells the user demonstrably uses.
+    Opt-in shells are never chosen here.
+    """
+    specs = [spec for spec in SHELLS.values() if not spec.opt_in]
+    chosen = [spec.name for spec in specs if spec.is_login_shell]
+    chosen += [
+        spec.name for spec in specs if spec.name not in chosen and spec.installed and spec.has_rc
+    ]
+    return chosen or ["zsh"]
+
+
+# ---------------------------------------------------------------------------
+# The install record
+# ---------------------------------------------------------------------------
+
+
+@dataclass(frozen=True)
+class InstallRecord:
+    """What ``config.json`` keeps about an install.
+
+    ``che update`` repeats the install from it, ``che doctor`` checks against
+    it, and its absence is how the dispatcher knows to run the first-run
+    wizard. Every field has a default so a record an older che wrote, missing
+    a key, still loads.
+    """
+
+    version: str = ""
+    home: str = ""
+    python: str = ""
+    shells: list[str] = field(default_factory=list)
+    rc_files: list[str] = field(default_factory=list)
+    sounds: bool = True
+    path_shim: bool = True
+    installed_at: str = ""
+
+
+def load_record() -> InstallRecord | None:
+    """The recorded install, or None when there is none.
+
+    A missing, unreadable or empty file, or one that does not hold a JSON
+    object, all read as "not installed". Keys this che does not know (older
+    records carried ``completions``) are ignored, and an empty ``shells``
+    means the default selection - so ``che update``, which reinstalls into
+    them, and ``che doctor``, which checks them, agree on what it means.
+    """
+    try:
+        data = json.loads(config_file().read_text(encoding="utf-8"))
+    except (OSError, ValueError):
+        return None
+    if not isinstance(data, dict) or not data:
+        return None
+
+    known = {entry.name for entry in fields(InstallRecord)}
+    record = InstallRecord(**{key: value for key, value in data.items() if key in known})
+    return record if record.shells else replace(record, shells=default_shell_selection())
+
+
+def save_record(record: InstallRecord, *, dry_run: bool = False) -> None:
+    if dry_run:
+        return
+    atomic_write_text(config_file(), json.dumps(asdict(record), indent=2) + "\n")
+
+
+# ---------------------------------------------------------------------------
+# Python interpreter
+# ---------------------------------------------------------------------------
+
+MIN_PYTHON = (3, 12)
+PYTHON_HINT = "brew install python@3.13"
+
+
+def version_text(version: tuple[int, ...]) -> str:
+    return ".".join(str(part) for part in version)
+
+
+MIN_PYTHON_TEXT = version_text(MIN_PYTHON)
+
+
+def python_ok(version: tuple[int, ...]) -> bool:
+    return tuple(version[:2]) >= MIN_PYTHON
+
+
+def python_version(executable: str) -> tuple[int, ...] | None:
+    """Version of another interpreter, or None if it cannot be asked."""
+    if executable == sys.executable:
+        return sys.version_info[:3]
+    try:
+        output = run(
+            [executable, "-c", "import sys; print('.'.join(str(p) for p in sys.version_info[:3]))"],
+            capture=True,
+            check=True,
+            timeout=10,
+        ).stdout
+    except (OSError, subprocess.SubprocessError):
+        return None
+    try:
+        return tuple(int(part) for part in output.strip().split("."))
+    except ValueError:
+        return None
+
+
+def find_python() -> tuple[str, tuple[int, ...]]:
+    """Pick the interpreter the wrappers should use.
+
+    The interpreter running this script wins when it is new enough - it is the
+    one the user just proved works. Otherwise fall back to the newest suitable
+    ``python3.X`` on PATH. macOS matters here: ``/usr/bin/python3`` is whatever
+    the Command Line Tools shipped, often years behind the Homebrew python that
+    is also installed. With nothing suitable anywhere, the running interpreter
+    comes back anyway, and the caller reports it as too old.
+    """
+    if python_ok(sys.version_info):
+        return sys.executable, sys.version_info[:3]
+
+    for candidate in ("python3.14", "python3.13", "python3.12", "python3"):
+        resolved = shutil.which(candidate)
+        if not resolved:
+            continue
+        version = python_version(resolved)
+        if version and python_ok(version):
+            return resolved, version
+
+    return sys.executable, sys.version_info[:3]
+
+
+# ---------------------------------------------------------------------------
+# Where this copy of che came from
+# ---------------------------------------------------------------------------
+
+
+def is_git_checkout() -> bool:
+    """Whether this copy of che is a clone, which ``che update`` can pull.
+
+    The npm package is the same tree without ``.git``.
+    """
+    return (REPO / ".git").exists()
+
+
+def update_command() -> str:
+    """The command that fetches a newer che for this kind of install."""
+    if is_git_checkout():
+        return "git pull"
+    return f"npm install -g {PACKAGE_NAME}@latest"

package/bin/media.py

@@ -0,0 +1,127 @@
+"""ffprobe helpers and shared extension lists for the scripts under ``bin/video/``.
+
+Split out of ``utils.py``, which keeps the logging, argument parsing and
+subprocess plumbing every script shares. Scripts pick it up the same way::
+
+    import sys
+    from pathlib import Path
+
+    sys.path.insert(0, str(Path(__file__).resolve().parent.parent))
+    from media import probe_media
+
+Standard library only, like everything else at the ``bin/`` root. ffprobe
+itself is an external binary; callers check for it with ``require_binary``.
+"""
+
+from __future__ import annotations
+
+import json
+from dataclasses import dataclass
+from pathlib import Path
+
+from utils import run
+
+# The two containers the library is kept in: what delete-duplicate-videos,
+# rename-video-file, validate-video-files and video-list all scan.
+LIBRARY_EXTENSIONS = ("mp4", "mkv")
+
+# What show-codecs audits and fix-codecs re-encodes. One list, so fix-codecs
+# can never pass over a file show-codecs reported.
+CODEC_SCAN_EXTENSIONS = ("mp4", "mkv", "mov", "avi")
+
+# Everything any video script reads, so each file is opened once.
+_PROBE_ENTRIES = "stream=codec_type,codec_name,width:stream_tags=language:format=bit_rate"
+
+
+@dataclass(frozen=True)
+class MediaProbe:
+    """ffprobe's view of one file's streams and container.
+
+    The codec and language accessors return what the one-field-per-call
+    ffprobe invocations the scripts used to make printed, so callers read the
+    same values as before. That takes restoring by hand: the JSON writer OMITS
+    a field ffprobe cannot fill, where the csv and default writers printed
+    ``unknown`` or ``N/A``. ``width`` and ``bit_rate`` are numbers instead,
+    ``None`` where ffprobe had nothing to report.
+
+    One difference is deliberate. On ffprobe 9 the csv writer appended a comma
+    per side-data block on the stream, so AC3 or E-AC3 audio in an MP4 came
+    back as ``ac3,`` and its language as ``und,``. JSON has no such artefact.
+    """
+
+    streams: tuple[dict, ...]
+    format: dict
+
+    @classmethod
+    def from_json(cls, data: dict) -> MediaProbe:
+        """Build from ffprobe's parsed ``-of json`` output."""
+        return cls(streams=tuple(data.get("streams", ())), format=data.get("format", {}))
+
+    def first(self, kind: str) -> dict | None:
+        """The first ``"video"``/``"audio"``/``"subtitle"`` stream: ``-select_streams v:0``."""
+        return next((s for s in self.streams if s.get("codec_type") == kind), None)
+
+    def codec(self, kind: str) -> str:
+        """The first ``kind`` stream's codec.
+
+        ``""`` when there is no such stream, ``"unknown"`` when there is one
+        but ffprobe cannot name its codec - the two values the per-field calls
+        printed for those cases.
+        """
+        stream = self.first(kind)
+        if stream is None:
+            return ""
+        return stream.get("codec_name", "unknown")
+
+    def width(self) -> int | None:
+        """The first video stream's width in pixels, or None."""
+        return _count((self.first("video") or {}).get("width", ""))
+
+    def bit_rate(self) -> int | None:
+        """The container's overall bit rate, or None where ffprobe said ``N/A``."""
+        return _count(self.format.get("bit_rate", ""))
+
+    def languages(self, kind: str) -> list[str]:
+        """Every language tagged on a ``kind`` stream, sorted and de-duplicated.
+
+        The key is matched case-insensitively because ``-show_entries
+        stream_tags=language`` matches that way, so a container's ``LANGUAGE``
+        tag arrives under its own spelling.
+        """
+        return sorted(
+            {
+                value.strip()
+                for stream in self.streams
+                if stream.get("codec_type") == kind
+                for key, value in stream.get("tags", {}).items()
+                if key.lower() == "language" and value.strip()
+            }
+        )
+
+
+def _count(value: object) -> int | None:
+    """``value`` as a non-negative integer, or None when it is absent or not numeric."""
+    text = str(value)
+    return int(text) if text.isdecimal() else None
+
+
+def probe_media(path: Path) -> MediaProbe | None:
+    """Probe ``path`` once for everything the video scripts read.
+
+    Returns None when ffprobe fails - a corrupt or unreadable file - which
+    callers have to tell apart from a readable file that lacks a stream or a
+    tag. One process per file, where the scripts used to spawn one per field:
+    on an SMB share each of those re-opened the file and re-read its header.
+    """
+    completed = run(
+        ["ffprobe", "-v", "error", "-show_entries", _PROBE_ENTRIES, "-of", "json", str(path)],
+        check=False,
+        capture=True,
+    )
+    if completed.returncode != 0:
+        return None
+    try:
+        data = json.loads(completed.stdout or "")
+    except json.JSONDecodeError:
+        return None
+    return MediaProbe.from_json(data)