renoise-kit 0.5.0__tar.gz → 0.5.1__tar.gz

{renoise_kit-0.5.0/renoise_kit.egg-info → renoise_kit-0.5.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: renoise-kit
-Version: 0.5.0
+Version: 0.5.1
 Summary: rn: frame-exact probing, contact sheets, EDL assembly, finishing and QA for AI video edits
 License-Expression: MIT
 Project-URL: Homepage, https://renoise.ai
@@ -58,17 +58,20 @@ machine.
 
 `rn` drives the system's ffmpeg and ffprobe, version 7 or later, built with libass for burned-in subtitles.
 
-| | macOS | Debian / Ubuntu |
-|---|---|---|
-| ffmpeg with libass | `brew install ffmpeg-full`, then `export PATH="$(brew --prefix ffmpeg-full)/bin:$PATH"` (keg-only; Homebrew's plain `ffmpeg` has no libass) | `apt install ffmpeg fontconfig` |
-| OCR for `rn text` (optional) | comes with `ffmpeg-full` | `apt install tesseract-ocr tesseract-ocr-chi-sim` |
+| | macOS | Debian / Ubuntu | Windows 10 / 11 |
+|---|---|---|---|
+| ffmpeg with libass | `brew install ffmpeg-full`, then `export PATH="$(brew --prefix ffmpeg-full)/bin:$PATH"` (keg-only; Homebrew's plain `ffmpeg` has no libass) | `apt install ffmpeg fontconfig` | `winget install Gyan.FFmpeg` (the full build, libass and fontconfig included), then open a new terminal |
+| OCR for `rn text` (optional) | comes with `ffmpeg-full` | `apt install tesseract-ocr tesseract-ocr-chi-sim` | `winget install UB-Mannheim.TesseractOCR`, then add `C:\Program Files\Tesseract-OCR` to `PATH` |
 
 ```sh
-uv tool install 'renoise-kit[all]'     # or: pipx install 'renoise-kit[all]'
+uv tool install "renoise-kit[all]"     # or: pipx install "renoise-kit[all]"
 playwright install chromium            # only for rn capture
 rn --version
 ```
 
+The same commands work in a POSIX shell, PowerShell and cmd (hence the double quotes). On Windows `rn fonts` reads the
+system font folders directly when fontconfig is not installed, and `rn` writes UTF-8 when its output goes to a pipe.
+
 The core install is Pillow and numpy. Heavier dependencies come as extras, so you can install only what you use. When
 a command needs a missing extra, its error line names the extra to install.
 
@@ -286,8 +289,9 @@ In the Renoise media sandbox, `rn` comes preinstalled at the version pinned by t
 uv run pytest -q    # Python 3.12 (as the sandbox); needs ffmpeg with libass; playwright install chromium for capture tests
 ```
 
-CI runs on Debian trixie with ffmpeg 7.1 (the same base as the sandbox) and on macOS with ffmpeg-full. The tests
-generate their own media.
+CI runs on Debian trixie with ffmpeg 7.1 (the same base as the sandbox), on macOS with ffmpeg-full and on Windows with
+gyan.dev's full build (with Windows' own default encoding, so a file read without `encoding='utf-8'` fails there). The
+tests generate their own media.
 
 To release:
 

{renoise_kit-0.5.0 → renoise_kit-0.5.1}/README.md

@@ -22,17 +22,20 @@ machine.
 
 `rn` drives the system's ffmpeg and ffprobe, version 7 or later, built with libass for burned-in subtitles.
 
-| | macOS | Debian / Ubuntu |
-|---|---|---|
-| ffmpeg with libass | `brew install ffmpeg-full`, then `export PATH="$(brew --prefix ffmpeg-full)/bin:$PATH"` (keg-only; Homebrew's plain `ffmpeg` has no libass) | `apt install ffmpeg fontconfig` |
-| OCR for `rn text` (optional) | comes with `ffmpeg-full` | `apt install tesseract-ocr tesseract-ocr-chi-sim` |
+| | macOS | Debian / Ubuntu | Windows 10 / 11 |
+|---|---|---|---|
+| ffmpeg with libass | `brew install ffmpeg-full`, then `export PATH="$(brew --prefix ffmpeg-full)/bin:$PATH"` (keg-only; Homebrew's plain `ffmpeg` has no libass) | `apt install ffmpeg fontconfig` | `winget install Gyan.FFmpeg` (the full build, libass and fontconfig included), then open a new terminal |
+| OCR for `rn text` (optional) | comes with `ffmpeg-full` | `apt install tesseract-ocr tesseract-ocr-chi-sim` | `winget install UB-Mannheim.TesseractOCR`, then add `C:\Program Files\Tesseract-OCR` to `PATH` |
 
 ```sh
-uv tool install 'renoise-kit[all]'     # or: pipx install 'renoise-kit[all]'
+uv tool install "renoise-kit[all]"     # or: pipx install "renoise-kit[all]"
 playwright install chromium            # only for rn capture
 rn --version
 ```
 
+The same commands work in a POSIX shell, PowerShell and cmd (hence the double quotes). On Windows `rn fonts` reads the
+system font folders directly when fontconfig is not installed, and `rn` writes UTF-8 when its output goes to a pipe.
+
 The core install is Pillow and numpy. Heavier dependencies come as extras, so you can install only what you use. When
 a command needs a missing extra, its error line names the extra to install.
 
@@ -250,8 +253,9 @@ In the Renoise media sandbox, `rn` comes preinstalled at the version pinned by t
 uv run pytest -q    # Python 3.12 (as the sandbox); needs ffmpeg with libass; playwright install chromium for capture tests
 ```
 
-CI runs on Debian trixie with ffmpeg 7.1 (the same base as the sandbox) and on macOS with ffmpeg-full. The tests
-generate their own media.
+CI runs on Debian trixie with ffmpeg 7.1 (the same base as the sandbox), on macOS with ffmpeg-full and on Windows with
+gyan.dev's full build (with Windows' own default encoding, so a file read without `encoding='utf-8'` fails there). The
+tests generate their own media.
 
 To release:
 

{renoise_kit-0.5.0 → renoise_kit-0.5.1/renoise_kit.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: renoise-kit
-Version: 0.5.0
+Version: 0.5.1
 Summary: rn: frame-exact probing, contact sheets, EDL assembly, finishing and QA for AI video edits
 License-Expression: MIT
 Project-URL: Homepage, https://renoise.ai
@@ -58,17 +58,20 @@ machine.
 
 `rn` drives the system's ffmpeg and ffprobe, version 7 or later, built with libass for burned-in subtitles.
 
-| | macOS | Debian / Ubuntu |
-|---|---|---|
-| ffmpeg with libass | `brew install ffmpeg-full`, then `export PATH="$(brew --prefix ffmpeg-full)/bin:$PATH"` (keg-only; Homebrew's plain `ffmpeg` has no libass) | `apt install ffmpeg fontconfig` |
-| OCR for `rn text` (optional) | comes with `ffmpeg-full` | `apt install tesseract-ocr tesseract-ocr-chi-sim` |
+| | macOS | Debian / Ubuntu | Windows 10 / 11 |
+|---|---|---|---|
+| ffmpeg with libass | `brew install ffmpeg-full`, then `export PATH="$(brew --prefix ffmpeg-full)/bin:$PATH"` (keg-only; Homebrew's plain `ffmpeg` has no libass) | `apt install ffmpeg fontconfig` | `winget install Gyan.FFmpeg` (the full build, libass and fontconfig included), then open a new terminal |
+| OCR for `rn text` (optional) | comes with `ffmpeg-full` | `apt install tesseract-ocr tesseract-ocr-chi-sim` | `winget install UB-Mannheim.TesseractOCR`, then add `C:\Program Files\Tesseract-OCR` to `PATH` |
 
 ```sh
-uv tool install 'renoise-kit[all]'     # or: pipx install 'renoise-kit[all]'
+uv tool install "renoise-kit[all]"     # or: pipx install "renoise-kit[all]"
 playwright install chromium            # only for rn capture
 rn --version
 ```
 
+The same commands work in a POSIX shell, PowerShell and cmd (hence the double quotes). On Windows `rn fonts` reads the
+system font folders directly when fontconfig is not installed, and `rn` writes UTF-8 when its output goes to a pipe.
+
 The core install is Pillow and numpy. Heavier dependencies come as extras, so you can install only what you use. When
 a command needs a missing extra, its error line names the extra to install.
 
@@ -286,8 +289,9 @@ In the Renoise media sandbox, `rn` comes preinstalled at the version pinned by t
 uv run pytest -q    # Python 3.12 (as the sandbox); needs ffmpeg with libass; playwright install chromium for capture tests
 ```
 
-CI runs on Debian trixie with ffmpeg 7.1 (the same base as the sandbox) and on macOS with ffmpeg-full. The tests
-generate their own media.
+CI runs on Debian trixie with ffmpeg 7.1 (the same base as the sandbox), on macOS with ffmpeg-full and on Windows with
+gyan.dev's full build (with Windows' own default encoding, so a file read without `encoding='utf-8'` fails there). The
+tests generate their own media.
 
 To release:
 

{renoise_kit-0.5.0 → renoise_kit-0.5.1}/renoise_kit.egg-info/SOURCES.txt

@@ -40,6 +40,7 @@ tests/test_edl_v3.py
 tests/test_faces.py
 tests/test_fx.py
 tests/test_listen.py
+tests/test_portability.py
 tests/test_rn.py
 tests/test_v3_commands.py
 tests/test_v4_commands.py

{renoise_kit-0.5.0 → renoise_kit-0.5.1}/rn/__init__.py

@@ -1,2 +1,2 @@
 """Renoise kit (`rn`, PyPI `renoise-kit`). Released on its own; the media sandbox template pins a version."""
-__version__ = '0.5.0'
+__version__ = '0.5.1'

{renoise_kit-0.5.0 → renoise_kit-0.5.1}/rn/assemble.py

@@ -211,7 +211,9 @@ def _join(edit: Edit, slots: list[Path], work: Path) -> Path:
              '-map', '[v]', *INTERMEDIATE, '-r', str(edit.fps), '-y', str(piece)])
         pieces.append(piece)
     listing = work / 'main.txt'
-    listing.write_text(''.join(f"file '{piece.resolve()}'\n" for piece in pieces))
+    # Forward slashes work on every OS (a backslash is the demuxer's escape character); a quote closes and reopens.
+    quoted = (piece.resolve().as_posix().replace("'", "'\\''") for piece in pieces)
+    listing.write_text(''.join(f"file '{path}'\n" for path in quoted), encoding='utf-8')
     return listing
 
 
@@ -467,7 +469,7 @@ def _cached(cache: Path | None, key: str, suffix: str, work: Path, render) -> Pa
 def _fingerprint(path: Path, placeholders: bool) -> str:
     """A nested edit's render depends on its text and, recursively, on every file it reads."""
     edit = load(path, placeholders=placeholders)
-    parts = [__version__, path.read_text(), placeholders]
+    parts = [__version__, path.read_text(encoding='utf-8'), placeholders]
     for clip in edit.base + edit.overlays:
         parts.append(_fingerprint(clip.source, placeholders) if clip.kind == 'edit' else _files(clip.source, clip.kind == 'sequence'))
     parts += [_files(item.source) for item in edit.audio] + [_files(edit.subtitles)]

{renoise_kit-0.5.0 → renoise_kit-0.5.1}/rn/assets.py

@@ -56,7 +56,7 @@ def fetch_checked(source: str, digest: str, cached: Path, *, name: str) -> Path:
     if hasher.hexdigest() != digest:
         partial.unlink()
         raise RnError(f'{name}: checksum mismatch, refusing the download')
-    partial.rename(cached)
+    partial.replace(cached)  # rename() refuses to overwrite on Windows
     return cached
 
 

{renoise_kit-0.5.0 → renoise_kit-0.5.1}/rn/audio.py

@@ -20,7 +20,7 @@ MAX_LIMITER_PASSES = 6
 def measure(path: str | Path, *, stream: str = 'a:0') -> dict:
     """Integrated loudness (LUFS), loudness range and true peak (dBTP) via ebur128."""
     result = subprocess.run(['ffmpeg', '-hide_banner', '-nostdin', '-i', str(path), '-map', f'0:{stream}',
-                             '-af', 'ebur128=peak=true:framelog=quiet', '-f', 'null', '-'], capture_output=True, text=True)
+                             '-af', 'ebur128=peak=true:framelog=quiet', '-f', 'null', '-'], capture_output=True, text=True, encoding='utf-8', errors='replace')
     if result.returncode != 0:
         raise RnError(f'loudness measurement failed: {result.stderr.strip().splitlines()[-1:]}')
     summary = result.stderr[result.stderr.rfind('Summary:'):]

{renoise_kit-0.5.0 → renoise_kit-0.5.1}/rn/captions.py

@@ -7,6 +7,7 @@ from __future__ import annotations
 
 import json
 import re
+import sys
 import unicodedata
 from difflib import SequenceMatcher
 from pathlib import Path
@@ -27,7 +28,9 @@ OPENERS = set('「『（([“‘《〈【')
 SENTENCE_END = set('。！？.!?…‼⁇')
 SOFT_BREAK = set('，、；：,;:—')
 CLOSERS = set('」』）)]”’》〉】"\'')
-FONTS = {'zh': 'Noto Sans CJK SC', 'yue': 'Noto Sans CJK SC', 'ja': 'Noto Sans CJK JP', 'ko': 'Noto Sans CJK KR'}
+FONTS = {'zh': 'Noto Sans CJK SC', 'yue': 'Noto Sans CJK SC', 'ja': 'Noto Sans CJK JP', 'ko': 'Noto Sans CJK KR', '': 'Noto Sans'}
+# Windows ships no Noto face and libass would fall back silently, so name the faces every Windows has.
+WINDOWS_FONTS = {'zh': 'Microsoft YaHei', 'yue': 'Microsoft YaHei', 'ja': 'Yu Gothic', 'ko': 'Malgun Gothic', '': 'Segoe UI'}
 PAUSE_BREAK = 0.5  # seconds of silence that ends a caption
 
 
@@ -220,7 +223,7 @@ def captions_command(words_json: str, *, script: str | None = None, faces_json:
     if width <= 0 or height <= 0 or max_lines < 1 or min_duration > max_duration:
         raise RnError('captions: width/height must be positive, max_lines >= 1, min_duration <= max_duration')
     try:
-        listen = json.loads(Path(words_json).read_text(encoding='utf-8'))
+        listen = json.loads(Path(words_json).read_text(encoding='utf-8-sig'))
     except (OSError, ValueError) as error:
         raise RnError(f'cannot read words json {words_json}: {error}') from error
     words = [w for w in listen.get('words', []) if str(w.get('word', '')).strip()]
@@ -234,7 +237,7 @@ def captions_command(words_json: str, *, script: str | None = None, faces_json:
     alignment = None
     if script is not None:
         try:
-            script_text = Path(script).read_text(encoding='utf-8')
+            script_text = Path(script).read_text(encoding='utf-8-sig')
         except OSError as error:
             raise RnError(f'cannot read script {script}: {error}') from error
         script_units = _script_units(script_text, cjk)
@@ -251,7 +254,8 @@ def captions_command(words_json: str, *, script: str | None = None, faces_json:
     max_chars = max_chars or max(4, round(BASE_CHARS['cjk' if cjk else 'latin'] * width / 1080))
     em_per_char = 1.0 if cjk else LATIN_EM
     font_size = round(min(height * (0.03 if height > width else 0.05), safe_width / (max_chars * em_per_char)))
-    font_name = FONTS.get(language.split('-')[0].lower(), 'Noto Sans') if language else ('Noto Sans CJK SC' if cjk else 'Noto Sans')
+    fonts = WINDOWS_FONTS if sys.platform == 'win32' else FONTS
+    font_name = fonts.get(language.split('-')[0].lower(), fonts['']) if language else fonts['zh' if cjk else '']
 
     chunks = _chunk(units, max_chars=max_chars, max_lines=max_lines, min_duration=min_duration, max_duration=max_duration, cjk=cjk)
     for chunk in chunks:
@@ -263,7 +267,7 @@ def captions_command(words_json: str, *, script: str | None = None, faces_json:
     scale = (1.0, 1.0)
     if faces_json:
         try:
-            faces = json.loads(Path(faces_json).read_text(encoding='utf-8'))
+            faces = json.loads(Path(faces_json).read_text(encoding='utf-8-sig'))
         except (OSError, ValueError) as error:
             raise RnError(f'cannot read faces json {faces_json}: {error}') from error
         frames = faces.get('frames', [])

{renoise_kit-0.5.0 → renoise_kit-0.5.1}/rn/capture.py

@@ -22,7 +22,7 @@ FRAME_TIMEOUT_MS = 20_000
 def _init_script(fps: float, frames: int, width: int, height: int, shard: int) -> str:
     from .fx import js_path
     meta = json.dumps({'fps': fps, 'frames': frames, 'width': width, 'height': height, 'shard': shard})
-    return f'window.RN = {meta};\n' + js_path().read_text()
+    return f'window.RN = {meta};\n' + js_path().read_text(encoding='utf-8')
 
 
 def _shard(args: tuple) -> dict:
@@ -64,7 +64,7 @@ def capture_command(page: str, *, out: str, frames: int, width: int, height: int
     try:
         import playwright  # noqa: F401
     except ImportError:
-        raise RnError("rn capture needs playwright (extra \"capture\"): uv tool install 'renoise-kit[all]' && playwright install chromium") from None
+        raise RnError('rn capture needs playwright (extra "capture"): uv tool install "renoise-kit[all]" && playwright install chromium') from None
     out_dir = Path(out)
     out_dir.mkdir(parents=True, exist_ok=True)
     for stale in out_dir.glob('*.png'):

{renoise_kit-0.5.0 → renoise_kit-0.5.1}/rn/cli.py

@@ -9,6 +9,7 @@ from __future__ import annotations
 
 import argparse
 import json
+import os
 import shutil
 import sys
 from pathlib import Path
@@ -492,6 +493,11 @@ def _run_summarized(args) -> tuple[dict, list[str]]:
 
 
 def main(argv: list[str] | None = None) -> int:
+    if sys.platform == 'win32' and not os.environ.get('PYTHONIOENCODING'):
+        # A pipe (an agent reading rn) would get the ANSI code page, which cannot hold every file name or caption.
+        for stream in (sys.stdout, sys.stderr):
+            if not stream.isatty() and hasattr(stream, 'reconfigure'):
+                stream.reconfigure(encoding='utf-8', errors='replace')
     args = _parser().parse_args(argv)
     for tool in ('ffmpeg', 'ffprobe'):
         if not shutil.which(tool):
@@ -536,7 +542,7 @@ def main(argv: list[str] | None = None) -> int:
     except ModuleNotFoundError as error:
         extra = EXTRAS.get(args.command, 'all')
         print(f"rn: error: rn {args.command} needs the Python module {error.name} (extra \"{extra}\"): "
-              f"uv tool install 'renoise-kit[all]'", file=sys.stderr)
+              'uv tool install "renoise-kit[all]"', file=sys.stderr)
         return 1
     except Exception as error:  # a kit bug: keep the traceback for the report, end on the usual one-liner
         import traceback
@@ -545,7 +551,7 @@ def main(argv: list[str] | None = None) -> int:
         return 1
     if args.json:
         Path(args.json).parent.mkdir(parents=True, exist_ok=True)
-        Path(args.json).write_text(json.dumps(result, indent=1, ensure_ascii=False, default=str))
+        Path(args.json).write_text(json.dumps(result, indent=1, ensure_ascii=False, default=str), encoding='utf-8')
     if args.command in SUMMARIZED:
         print('\n'.join(lines))
     else:

{renoise_kit-0.5.0 → renoise_kit-0.5.1}/rn/editops.py

@@ -17,7 +17,7 @@ from .ff import RnError
 
 def _read(path: Path) -> dict:
     try:
-        return json.loads(path.read_text())
+        return json.loads(path.read_text(encoding='utf-8-sig'))
     except FileNotFoundError:
         raise RnError(f'no such file: {path}') from None
     except json.JSONDecodeError as error:
@@ -27,7 +27,7 @@ def _read(path: Path) -> dict:
 def _write_checked(path: Path, raw: dict) -> dict:
     """Lays the edited JSON out next to the original (relative paths resolve the same), then replaces it."""
     trial = path.with_name(f'.{path.stem}.rn-edit.json')
-    trial.write_text(json.dumps(raw, indent=2, ensure_ascii=False) + '\n')
+    trial.write_text(json.dumps(raw, indent=2, ensure_ascii=False) + '\n', encoding='utf-8')
     try:
         summary = load(trial).summary()
     except RnError:
@@ -209,7 +209,7 @@ def _destination(raw: dict, before: str | None, after: str | None, track: str |
 
 def insert_op(path: Path, clip_json: str, before: str | None, after: str | None, track: str | None) -> dict:
     """Adds a clip given as JSON (or @file.json)."""
-    text = Path(clip_json[1:]).read_text() if clip_json.startswith('@') else clip_json
+    text = Path(clip_json[1:]).read_text(encoding='utf-8-sig') if clip_json.startswith('@') else clip_json
     try:
         clip = json.loads(text)
     except json.JSONDecodeError as error:

{renoise_kit-0.5.0 → renoise_kit-0.5.1}/rn/edl.py

@@ -242,7 +242,7 @@ class Timeline:
         path = _resolve(self.edit_dir, value, where)
         if str(path) not in self._listen:
             try:
-                self._listen[str(path)] = json.loads(path.read_text())
+                self._listen[str(path)] = json.loads(path.read_text(encoding='utf-8-sig'))
             except json.JSONDecodeError as error:
                 raise RnError(f'{where}: {path.name} is not JSON ({error}); write it with rn listen --json') from None
         return self._listen[str(path)]
@@ -388,21 +388,29 @@ def _ident(raw: dict, where: str) -> tuple[str | None, object]:
     return ident, raw.get('meta')
 
 
-_SANS_CJK = ('Noto Sans CJK', 'NotoSansCJK', 'PingFang', 'Source Han Sans', 'SourceHanSans', 'Hiragino Sans', 'Heiti', 'WenQuanYi')
+_SANS_CJK = ('Noto Sans CJK', 'NotoSansCJK', 'NotoSansSC', 'PingFang', 'Source Han Sans', 'SourceHanSans', 'Hiragino Sans', 'Heiti',
+             'WenQuanYi', 'wqy', 'msyh', 'Microsoft YaHei', 'Deng', 'simhei')
+_CJK_FILES = _SANS_CJK + ('simsun', 'STHeiti', 'Songti', 'NotoSerifCJK')  # without fontconfig: file names to try, sans first
 
 
 def _font(size: int):
-    """A face that covers Chinese when fontconfig lists one (placeholder text is often a prompt), else Pillow's default.
-    fc-match ':lang=zh' is not enough: it falls back to any font; fc-list only lists fonts that cover the language."""
+    """A face that covers Chinese when one is installed (placeholder text is often a prompt), else Pillow's default.
+    fc-match ':lang=zh' is not enough: it falls back to any font; fc-list only lists fonts that cover the language.
+    Without fontconfig (Windows) the system font folders are searched for well-known Chinese faces by file name."""
     from PIL import ImageFont
+    from .fonts import FONT_SUFFIXES, font_files
     if shutil.which('fc-list'):
-        files = subprocess.run(['fc-list', ':lang=zh-cn', '--format', '%{file}\n'], capture_output=True, text=True).stdout.split('\n')
-        files = [f for f in files if Path(f).suffix.lower() in ('.ttf', '.otf', '.ttc')]
-        for path in sorted(files, key=lambda f: not any(name in f for name in _SANS_CJK)):
-            try:
-                return ImageFont.truetype(path, size)
-            except OSError:
-                continue
+        files = subprocess.run(['fc-list', ':lang=zh-cn', '--format', '%{file}\n'], capture_output=True, text=True, encoding='utf-8', errors='replace').stdout.split('\n')
+        files = [f for f in files if Path(f).suffix.lower() in FONT_SUFFIXES]
+        files = sorted(files, key=lambda f: not any(name in f for name in _SANS_CJK))
+    else:
+        rank = lambda f: next((i for i, name in enumerate(_CJK_FILES) if name.lower() in f.name.lower()), None)  # noqa: E731
+        files = [str(f) for f in sorted((f for f in font_files() if rank(f) is not None), key=rank)]
+    for path in files:
+        try:
+            return ImageFont.truetype(path, size)
+        except OSError:
+            continue
     return ImageFont.load_default(size=size)
 
 
@@ -765,7 +773,7 @@ def load(path: str | Path, *, placeholders: bool = False, _stack: tuple[Path, ..
     path = Path(path).resolve()
     stack = (*_stack, path)
     try:
-        raw = json.loads(path.read_text())
+        raw = json.loads(path.read_text(encoding='utf-8-sig'))
     except FileNotFoundError:
         raise RnError(f'no such file: {path}') from None
     except json.JSONDecodeError as error:

{renoise_kit-0.5.0 → renoise_kit-0.5.1}/rn/ff.py

@@ -5,6 +5,7 @@ import json
 import re
 import struct
 import subprocess
+import tempfile
 from fractions import Fraction
 from functools import lru_cache
 from pathlib import Path
@@ -14,11 +15,28 @@ class RnError(Exception):
     """A failure the agent can act on; the CLI prints it as one line."""
 
 
+# Windows caps a whole command line at 32767 characters, Linux one argument at 128 KiB: a long filter graph (many
+# overlays, keyframes) goes to ffmpeg as a file instead, through ffmpeg 7's "-/option file".
+LONG_FILTER = 8000
+FILTER_OPTIONS = ('-filter_complex', '-lavfi', '-vf', '-af', '-filter:v', '-filter:a')
+
+
+def _filters_from_files(args: list[str], folder: Path) -> list[str]:
+    out = list(args)
+    for i in range(len(out) - 1):
+        if out[i] in FILTER_OPTIONS and len(out[i + 1]) > LONG_FILTER:
+            script = folder / f'filter{i}.txt'
+            script.write_text(out[i + 1], encoding='utf-8')
+            out[i], out[i + 1] = '-/' + out[i][1:], str(script)
+    return out
+
+
 def run(args: list[str], *, quiet: bool = True, cwd: str | Path | None = None) -> subprocess.CompletedProcess:
     """Runs ffmpeg (or ffprobe); raises RnError with the tail of stderr on failure."""
     flags = ['-nostdin', *(['-v', 'error'] if quiet else [])] if args[0] == 'ffmpeg' else []
-    command = [args[0], '-hide_banner', *flags, *args[1:]]
-    result = subprocess.run(command, capture_output=True, text=True, cwd=cwd)
+    with tempfile.TemporaryDirectory(prefix='rn-filters-') as tmp:
+        command = [args[0], '-hide_banner', *flags, *_filters_from_files(args[1:], Path(tmp))]
+        result = subprocess.run(command, capture_output=True, text=True, encoding='utf-8', errors='replace', cwd=cwd)
     if result.returncode != 0:
         lines = result.stderr.strip().splitlines()
         # The cause is usually an earlier line than the generic 'Task finished with error' tail.
@@ -31,7 +49,7 @@ def run(args: list[str], *, quiet: bool = True, cwd: str | Path | None = None) -
 @lru_cache(maxsize=None)
 def has_filter(name: str) -> bool:
     """Whether this ffmpeg build has a filter (Homebrew's plain ffmpeg, for one, has no libass `subtitles`)."""
-    filters = subprocess.run(['ffmpeg', '-hide_banner', '-filters'], capture_output=True, text=True).stdout
+    filters = subprocess.run(['ffmpeg', '-hide_banner', '-filters'], capture_output=True, text=True, encoding='utf-8', errors='replace').stdout
     return any(line.split()[1:2] == [name] for line in filters.splitlines())
 
 

renoise_kit-0.5.1/rn/fonts.py

@@ -0,0 +1,154 @@
+"""rn fonts: what is actually installed, from fontconfig, or where there is none (Windows) from the system font folders.
+A missing weight makes browsers and libass fall back silently (Katana's sandbox only had some families in ExtraBold), so
+check the weight and script here before using a face."""
+from __future__ import annotations
+
+import os
+import shutil
+import subprocess
+import sys
+from pathlib import Path
+
+from .ff import RnError
+
+# fontconfig weight -> CSS weight (fontconfig.h FC_WEIGHT_*).
+_CSS = [(0, 100), (40, 200), (50, 300), (80, 400), (100, 500), (180, 600), (200, 700), (205, 800), (210, 900)]
+_SCRIPTS = {'latin': ('en',), 'zh-hans': ('zh-cn', 'zh-sg'), 'zh-hant': ('zh-tw', 'zh-hk'), 'ja': ('ja',), 'ko': ('ko',),
+            'th': ('th',), 'ar': ('ar',), 'he': ('he',), 'hi': ('hi',), 'ru': ('ru',)}
+
+
+FONT_SUFFIXES = ('.ttf', '.otf', '.ttc', '.otc')
+# Without fontconfig a font covers a script when its cmap has all of these characters.
+_SAMPLES = {'latin': 'AZaz', 'zh-hans': '这们说', 'zh-hant': '這們說', 'ja': 'あアの', 'ko': '한국어', 'th': 'กขค', 'ar': 'بتث',
+            'he': 'אבג', 'hi': 'कखग', 'ru': 'ЖЯЫ', 'emoji': '\U0001F600\U0001F389\U0001F44D'}
+
+
+def font_dirs() -> list[Path]:
+    """Where the OS keeps installed fonts, system-wide and per user."""
+    home = Path.home()
+    if sys.platform == 'win32':
+        local = Path(os.environ.get('LOCALAPPDATA') or home / 'AppData' / 'Local')
+        dirs = [Path(os.environ.get('WINDIR') or 'C:/Windows') / 'Fonts', local / 'Microsoft' / 'Windows' / 'Fonts']
+    elif sys.platform == 'darwin':
+        dirs = [Path('/System/Library/Fonts'), Path('/Library/Fonts'), home / 'Library' / 'Fonts']
+    else:
+        data = Path(os.environ.get('XDG_DATA_HOME') or home / '.local' / 'share')
+        dirs = [Path('/usr/share/fonts'), Path('/usr/local/share/fonts'), data / 'fonts', home / '.fonts']
+    return [d for d in dirs if d.is_dir()]
+
+
+def font_files() -> list[Path]:
+    return sorted({f for d in font_dirs() for f in d.rglob('*') if f.suffix.lower() in FONT_SUFFIXES and f.is_file()})
+
+
+def css_weight(fc_weight: float) -> int:
+    return min(_CSS, key=lambda pair: abs(pair[0] - fc_weight))[1]
+
+
+def _license(path: str) -> str | None:
+    try:
+        from fontTools.ttLib import TTFont
+    except ImportError:
+        return None
+    try:
+        font = TTFont(path, fontNumber=0, lazy=True)
+        names = font['name']
+        text = names.getDebugName(13) or names.getDebugName(14)
+    except Exception:  # unreadable or exotic font: no license info rather than a failure
+        return None
+    return ' '.join(text.split())[:120] if text else None
+
+
+def _entry(families: dict, family: str) -> dict:
+    return families.setdefault(family, {'family': family, 'weights': set(), 'styles': set(), 'variable': False,
+                                        'scripts': set(), 'files': set()})
+
+
+def installed() -> dict[str, dict]:
+    if not shutil.which('fc-list'):
+        return _scanned()
+    output = subprocess.run(['fc-list', '--format', '%{family[0]}\t%{style[0]}\t%{weight}\t%{variable}\t%{file}\t%{lang}\n'],
+                            capture_output=True, text=True, encoding='utf-8', errors='replace', check=True).stdout
+    families: dict[str, dict] = {}
+    for line in output.splitlines():
+        parts = line.split('\t')
+        if len(parts) != 6 or not parts[0] or parts[0].startswith('.'):
+            continue
+        family, style, weight, variable, path, langs = parts
+        entry = _entry(families, family)
+        try:
+            entry['weights'].add(css_weight(float(weight)))
+        except ValueError:
+            pass
+        entry['styles'].add(style)
+        entry['variable'] |= variable == 'True'
+        entry['files'].add(path)
+        codes = set(langs.split('|'))
+        entry['scripts'] |= {name for name, keys in _SCRIPTS.items() if codes & set(keys)}
+        if 'emoji' in family.lower():
+            entry['scripts'].add('emoji')
+    return families
+
+
+def _scanned() -> dict[str, dict]:
+    """The same facts read from the font files themselves: name table, OS/2 weight, fvar axes, cmap coverage."""
+    try:
+        from fontTools.ttLib import TTCollection, TTFont
+    except ImportError:
+        raise RnError('fc-list not found: install fontconfig, or fontTools (extra "fonts") to read the system font '
+                      'folders directly: uv tool install "renoise-kit[all]"') from None
+    families: dict[str, dict] = {}
+    for path in font_files():
+        try:
+            faces = TTCollection(str(path), lazy=True).fonts if path.suffix.lower() in ('.ttc', '.otc') else [TTFont(str(path), lazy=True)]
+            for font in faces:
+                names = font['name']
+                family = names.getDebugName(16) or names.getDebugName(1)
+                if not family or family.startswith('.'):
+                    continue
+                style = names.getDebugName(17) or names.getDebugName(2) or 'Regular'
+                weight = font['OS/2'].usWeightClass if 'OS/2' in font else 400
+                weight = weight * 100 if weight < 10 else weight  # a few old fonts store 1-9
+                weights = {min(900, max(100, round(weight / 100) * 100))}
+                axes = {axis.axisTag: axis for axis in font['fvar'].axes} if 'fvar' in font else {}
+                if 'wght' in axes:
+                    weights |= {w for w in range(100, 1000, 100) if axes['wght'].minValue <= w <= axes['wght'].maxValue}
+                cmap = font.getBestCmap() or {}
+                entry = _entry(families, family)
+                entry['styles'].add(style)
+                entry['weights'] |= weights
+                entry['variable'] |= bool(axes)
+                entry['scripts'] |= {name for name, chars in _SAMPLES.items() if all(ord(c) in cmap for c in chars)}
+                entry['files'].add(str(path))
+        except Exception:  # unreadable or exotic font file: skip it rather than fail the listing
+            continue
+    return families
+
+
+def fonts_command(*, family: str | None, script: str | None, licenses: bool) -> dict:
+    if script and script not in (*_SCRIPTS, 'emoji'):
+        raise RnError(f'--script: one of {", ".join([*_SCRIPTS, "emoji"])}')
+    rows = []
+    for entry in sorted(installed().values(), key=lambda e: e['family'].lower()):
+        if family and family.lower() not in entry['family'].lower():
+            continue
+        if script and script not in entry['scripts']:
+            continue
+        files = sorted(entry['files'])
+        row = {'family': entry['family'], 'weights': sorted(entry['weights']), 'styles': sorted(entry['styles']),
+               'variable': entry['variable'], 'scripts': sorted(entry['scripts']), 'files': files}
+        if licenses:
+            row['license'] = _license(files[0])
+        rows.append(row)
+    if family and not rows:
+        raise RnError(f'no installed family matches "{family}"; run rn fonts to list them, or download one into your project')
+    return {'count': len(rows), 'fonts': rows}
+
+
+def summary_lines(result: dict) -> list[str]:
+    lines = []
+    for row in result['fonts']:
+        weights = ','.join(str(w) for w in row['weights']) + (' (variable)' if row['variable'] else '')
+        lines.append(f'{row["family"]}: {weights}  [{" ".join(row["scripts"]) or "-"}]' + (f'  {row["license"]}' if row.get('license') else ''))
+    lines.append(f'{result["count"]} families')
+    return lines

{renoise_kit-0.5.0 → renoise_kit-0.5.1}/rn/listen.py

@@ -22,7 +22,7 @@ BEAT_CLEAR = 0.15  # autocorrelation of the onset envelope at the beat period, 0
 def _short_term_curve(wav: Path) -> list[dict]:
     """Short-term (3 s) loudness every CURVE_STEP seconds from ebur128's frame log (it logs every 0.1 s)."""
     result = subprocess.run(['ffmpeg', '-hide_banner', '-nostdin', '-i', str(wav), '-af', 'ebur128',
-                             '-f', 'null', '-'], capture_output=True, text=True)
+                             '-f', 'null', '-'], capture_output=True, text=True, encoding='utf-8', errors='replace')
     curve, next_t = [], 0.0
     for match in re.finditer(r't:\s*([\d.]+)\s.*?M:\s*(-?[\d.]+|-inf)\s+S:\s*(-?[\d.]+|-inf)', result.stderr):
         t = float(match.group(1))
@@ -82,7 +82,7 @@ def _transcribe(wav: Path, *, language: str | None, model: str) -> dict:
     try:
         from faster_whisper import WhisperModel
     except ImportError:
-        raise RnError("--transcribe needs faster-whisper (extra \"transcribe\"): uv tool install 'renoise-kit[all]'")
+        raise RnError('--transcribe needs faster-whisper (extra "transcribe"): uv tool install "renoise-kit[all]"')
     import numpy as np
     # Decode here and pass samples: faster-whisper's own file decoding goes through PyAV, whose API it does not pin
     # (av 16 dropped the `metadata_errors` argument it passes).