meeting-noter 0.3.0__py3-none-any.whl

meeting_noter/transcription/engine.py

@@ -0,0 +1,208 @@
+"""Transcription engine using faster-whisper."""
+
+from __future__ import annotations
+
+import click
+from pathlib import Path
+from datetime import datetime
+from typing import Optional
+
+
+def get_model(model_size: str = "tiny.en"):
+    """Load the Whisper model.
+
+    Models are cached after first download (~75MB for tiny.en).
+    """
+    from faster_whisper import WhisperModel
+
+    click.echo(f"Loading model '{model_size}'...")
+
+    # Use INT8 for CPU efficiency
+    model = WhisperModel(
+        model_size,
+        device="cpu",
+        compute_type="int8",
+    )
+
+    return model
+
+
+def format_timestamp(seconds: float) -> str:
+    """Format seconds as [HH:MM:SS] or [MM:SS]."""
+    hours = int(seconds // 3600)
+    minutes = int((seconds % 3600) // 60)
+    secs = int(seconds % 60)
+
+    if hours > 0:
+        return f"[{hours:02d}:{minutes:02d}:{secs:02d}]"
+    return f"[{minutes:02d}:{secs:02d}]"
+
+
+def transcribe_audio(audio_path: Path, model_size: str = "tiny.en") -> str:
+    """Transcribe an audio file.
+
+    Args:
+        audio_path: Path to MP3 or WAV file
+        model_size: Whisper model to use
+
+    Returns:
+        Formatted transcript string
+    """
+    model = get_model(model_size)
+
+    click.echo(f"Transcribing {audio_path.name}...")
+
+    segments, info = model.transcribe(
+        str(audio_path),
+        beam_size=5,
+        vad_filter=True,  # Voice Activity Detection
+        vad_parameters=dict(
+            min_silence_duration_ms=500,
+        ),
+    )
+
+    # Build transcript
+    lines = []
+    lines.append("Meeting Transcription")
+    lines.append(f"File: {audio_path.name}")
+    lines.append(f"Transcribed: {datetime.now().strftime('%Y-%m-%d %H:%M:%S')}")
+    lines.append(f"Language: {info.language} (probability: {info.language_probability:.2f})")
+    lines.append(f"Duration: {info.duration:.1f} seconds")
+    lines.append("-" * 40)
+    lines.append("")
+
+    for segment in segments:
+        timestamp = format_timestamp(segment.start)
+        text = segment.text.strip()
+        if text:
+            lines.append(f"{timestamp} {text}")
+
+    return "\n".join(lines)
+
+
+def transcribe_file(
+    file: Optional[str],
+    output_dir: Path,
+    model_size: str = "tiny.en",
+    transcripts_dir: Optional[Path] = None,
+):
+    """Transcribe a recording file.
+
+    If no file specified, transcribes the most recent recording.
+    Saves transcript to transcripts_dir (defaults to same as audio file).
+    """
+    # Find the file to transcribe
+    if file:
+        audio_path = Path(file)
+        if not audio_path.exists():
+            # Try in output dir
+            audio_path = output_dir / file
+        if not audio_path.exists():
+            click.echo(click.style(f"File not found: {file}", fg="red"))
+            return
+    else:
+        # Find most recent recording
+        mp3_files = sorted(output_dir.glob("*.mp3"), key=lambda p: p.stat().st_mtime)
+        if not mp3_files:
+            click.echo(click.style(
+                f"No recordings found in {output_dir}",
+                fg="yellow"
+            ))
+            return
+        audio_path = mp3_files[-1]
+        click.echo(f"Using most recent recording: {audio_path.name}")
+
+    # Determine transcript path
+    if transcripts_dir:
+        transcripts_dir.mkdir(parents=True, exist_ok=True)
+        transcript_path = transcripts_dir / audio_path.with_suffix(".txt").name
+    else:
+        transcript_path = audio_path.with_suffix(".txt")
+    if transcript_path.exists():
+        click.echo(click.style(
+            f"Transcript already exists: {transcript_path.name}",
+            fg="yellow"
+        ))
+        if not click.confirm("Overwrite?"):
+            return
+
+    # Transcribe
+    try:
+        transcript = transcribe_audio(audio_path, model_size)
+
+        # Save transcript
+        transcript_path.write_text(transcript)
+        click.echo(click.style(
+            f"\nTranscript saved: {transcript_path}",
+            fg="green"
+        ))
+
+        # Show preview
+        click.echo("\n--- Preview ---")
+        lines = transcript.split("\n")
+        preview_lines = lines[:15]
+        for line in preview_lines:
+            click.echo(line)
+        if len(lines) > 15:
+            click.echo(f"... ({len(lines) - 15} more lines)")
+
+    except Exception as e:
+        click.echo(click.style(f"Transcription failed: {e}", fg="red"))
+        raise
+
+
+def transcribe_live(output_dir: Path, model_size: str = "tiny.en"):
+    """Real-time transcription of ongoing recording.
+
+    Note: This is more CPU intensive and may have latency.
+    """
+    click.echo(click.style(
+        "Live transcription is experimental and CPU-intensive.",
+        fg="yellow"
+    ))
+
+    # Find active recording (most recent, still being written)
+    mp3_files = sorted(output_dir.glob("*.mp3"), key=lambda p: p.stat().st_mtime)
+    if not mp3_files:
+        click.echo("No recordings found. Is the daemon running?")
+        return
+
+    latest = mp3_files[-1]
+    click.echo(f"Monitoring: {latest.name}")
+    click.echo("Press Ctrl+C to stop.\n")
+
+    model = get_model(model_size)
+    last_position = 0
+
+    try:
+        import time
+        while True:
+            # Check file size
+            current_size = latest.stat().st_size
+            if current_size > last_position:
+                # New data available - transcribe the whole file
+                # (faster-whisper doesn't support streaming well)
+                try:
+                    segments, info = model.transcribe(
+                        str(latest),
+                        beam_size=3,  # Faster for live
+                        vad_filter=True,
+                    )
+
+                    # Print new segments
+                    for segment in segments:
+                        if segment.end > last_position / 1000:  # Rough estimate
+                            timestamp = format_timestamp(segment.start)
+                            text = segment.text.strip()
+                            if text:
+                                click.echo(f"{timestamp} {text}")
+
+                    last_position = current_size
+                except Exception as e:
+                    # File might be locked, retry
+                    pass
+
+            time.sleep(5)  # Check every 5 seconds
+
+    except KeyboardInterrupt:
+        click.echo("\nStopped live transcription.")

meeting_noter-0.3.0.dist-info/METADATA

@@ -0,0 +1,261 @@
+Metadata-Version: 2.4
+Name: meeting-noter
+Version: 0.3.0
+Summary: Offline meeting transcription for macOS with automatic meeting detection
+Author: Victor
+License: MIT
+Project-URL: Homepage, https://github.com/tech4vision/meeting-noter
+Project-URL: Repository, https://github.com/tech4vision/meeting-noter
+Project-URL: Issues, https://github.com/tech4vision/meeting-noter/issues
+Keywords: meeting,transcription,whisper,offline,speech-to-text,audio
+Classifier: Development Status :: 3 - Alpha
+Classifier: Environment :: Console
+Classifier: Intended Audience :: End Users/Desktop
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Operating System :: MacOS
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.9
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Topic :: Multimedia :: Sound/Audio :: Speech
+Requires-Python: >=3.9
+Description-Content-Type: text/markdown
+Requires-Dist: click>=8.0
+Requires-Dist: sounddevice>=0.4.6
+Requires-Dist: numpy>=1.21
+Requires-Dist: faster-whisper>=1.0.0
+Requires-Dist: lameenc>=1.5.0
+Requires-Dist: rumps>=0.4.0
+Requires-Dist: PyQt6>=6.5.0
+Requires-Dist: pyobjc-framework-Cocoa>=9.0; sys_platform == "darwin"
+Requires-Dist: pyobjc-framework-Quartz>=9.0; sys_platform == "darwin"
+Requires-Dist: pyobjc-framework-ScreenCaptureKit>=9.0; sys_platform == "darwin"
+Requires-Dist: pyobjc-framework-AVFoundation>=9.0; sys_platform == "darwin"
+Requires-Dist: pyobjc-framework-CoreMedia>=9.0; sys_platform == "darwin"
+Requires-Dist: pyobjc-framework-libdispatch>=9.0; sys_platform == "darwin"
+Provides-Extra: dev
+Requires-Dist: pytest>=7.0; extra == "dev"
+Requires-Dist: pytest-cov; extra == "dev"
+Requires-Dist: black; extra == "dev"
+Requires-Dist: ruff; extra == "dev"
+Requires-Dist: mypy; extra == "dev"
+
+# Meeting Noter
+
+Offline meeting transcription tool for macOS. Captures both your voice and meeting participants' audio, saves to MP3, and transcribes locally using Whisper.
+
+## Features
+
+- **No virtual audio devices needed** - Uses ScreenCaptureKit (like Notion, Discord)
+- **Captures both sides** - Your mic + system audio (meeting participants)
+- **Offline transcription** - Uses Whisper locally, no API calls
+- **Auto-detection** - Detects active meetings (Zoom, Teams, Meet, Slack)
+- **Multiple interfaces** - Menu bar app, desktop GUI, or CLI
+- **Auto-segmentation** - One file per meeting (detects silence)
+
+## Installation
+
+### Option 1: Using pipx (Recommended)
+
+```bash
+# Install pipx if you don't have it
+brew install pipx
+pipx ensurepath
+
+# Install meeting-noter
+cd /path/to/meeting-noter
+pipx install -e .
+```
+
+### Option 2: Using a virtual environment
+
+```bash
+cd /path/to/meeting-noter
+python3 -m venv .venv
+.venv/bin/pip install --upgrade pip
+.venv/bin/pip install -e .
+```
+
+Then add an alias in your `~/.zshrc`:
+
+```bash
+echo 'alias meeting-noter="/path/to/meeting-noter/.venv/bin/meeting-noter"' >> ~/.zshrc
+source ~/.zshrc
+```
+
+## Quick Start
+
+### 1. One-time Setup
+
+```bash
+meeting-noter setup
+```
+
+This will request Screen Recording permission, which is needed to capture meeting participants' audio.
+
+### 2. Launch the App
+
+**Menu Bar App** (recommended):
+```bash
+meeting-noter menubar
+```
+
+**Desktop GUI**:
+```bash
+meeting-noter gui
+```
+
+**CLI Recording**:
+```bash
+meeting-noter start "Weekly Standup"
+```
+
+### 3. Record a Meeting
+
+- The menu bar app auto-detects meetings and prompts to record
+- Or manually start recording via the GUI/CLI
+- Press Ctrl+C (CLI) or click Stop to end recording
+
+### 4. Transcribe
+
+Recordings are auto-transcribed by default. Or manually:
+
+```bash
+# Transcribe the most recent recording
+meeting-noter transcribe
+
+# Transcribe a specific file
+meeting-noter transcribe recording.mp3
+
+# List all recordings
+meeting-noter list
+```
+
+## Commands
+
+| Command | Description |
+|---------|-------------|
+| `meeting-noter setup` | One-time setup (Screen Recording permission) |
+| `meeting-noter menubar` | Launch menu bar app |
+| `meeting-noter gui` | Launch desktop GUI |
+| `meeting-noter start [name]` | Interactive CLI recording |
+| `meeting-noter daemon` | Start background audio capture |
+| `meeting-noter status` | Check daemon status |
+| `meeting-noter stop` | Stop the daemon |
+| `meeting-noter list` | List recent recordings |
+| `meeting-noter transcribe` | Transcribe a recording |
+| `meeting-noter devices` | List audio devices |
+
+## Options
+
+### `start`
+- First argument: Meeting name (optional, auto-generates timestamp if omitted)
+
+### `daemon`
+- `-o, --output-dir`: Where to save recordings (default: `~/meetings`)
+- `-f, --foreground`: Run in foreground instead of background
+- `-n, --name`: Meeting name for the recording
+
+### `transcribe`
+- `-m, --model`: Whisper model size (tiny.en, base.en, small.en, medium.en, large-v3)
+- `-l, --live`: Real-time transcription (experimental)
+- `-o, --output-dir`: Directory with recordings
+
+### `list`
+- `-n, --limit`: Number of recordings to show
+- `-o, --output-dir`: Directory with recordings
+
+## How It Works
+
+```
+┌─────────────────────────────────────┐
+│         Your Meeting App            │
+│      (Zoom/Teams/Meet/Slack)        │
+└──────────────────┬──────────────────┘
+                   │
+     ┌─────────────┴─────────────┐
+     ▼                           ▼
+┌─────────┐               ┌─────────────┐
+│   Mic   │               │ System Audio│
+│(default)│               │(ScreenCaptureKit)
+└────┬────┘               └──────┬──────┘
+     │                           │
+     └───────────┬───────────────┘
+                 ▼
+         ┌─────────────┐
+         │Meeting Noter│
+         │  (capture)  │
+         └──────┬──────┘
+                │
+                ▼
+    ~/meetings/2024-01-15_Weekly_Standup.mp3
+                │
+                ▼ (auto or on-demand)
+         ┌─────────────┐
+         │   Whisper   │ (local)
+         └──────┬──────┘
+                │
+                ▼
+    ~/meetings/2024-01-15_Weekly_Standup.txt
+```
+
+## Permissions Required
+
+1. **Microphone** - For capturing your voice
+2. **Screen Recording** - For capturing system audio (meeting participants)
+   - Grant in: System Settings > Privacy & Security > Screen Recording
+   - This uses ScreenCaptureKit, the same API used by Notion, Discord, etc.
+
+## Troubleshooting
+
+### No system audio captured (only my voice)
+
+Screen Recording permission not granted. Go to:
+System Settings > Privacy & Security > Screen Recording
+
+Enable the toggle for Terminal (or your IDE/app).
+
+### Meeting not auto-detected
+
+Meeting detection works for: Zoom, Microsoft Teams, Google Meet, Slack.
+The meeting window must be open (not minimized).
+
+### Transcription is slow
+
+Use a smaller model:
+
+```bash
+meeting-noter transcribe --model tiny.en
+```
+
+Model sizes:
+- `tiny.en` (~75MB) - Fastest, good for most cases
+- `base.en` (~150MB) - Better accuracy
+- `small.en` (~500MB) - High accuracy
+- `medium.en` (~1.5GB) - Very high accuracy
+- `large-v3` (~3GB) - Best accuracy
+
+## Configuration
+
+Config file: `~/.config/meeting-noter/config.json`
+
+```json
+{
+    "recordings_dir": "~/meetings",
+    "transcripts_dir": "~/meetings",
+    "whisper_model": "tiny.en",
+    "auto_transcribe": true,
+    "silence_timeout": 5,
+    "capture_system_audio": true
+}
+```
+
+## Requirements
+
+- macOS 12.3+ (for ScreenCaptureKit)
+- Python 3.9+
+
+## License
+
+MIT

meeting_noter-0.3.0.dist-info/RECORD

@@ -0,0 +1,38 @@
+meeting_noter/__init__.py,sha256=bLOErRC3sfnQ4a4RyZUzUljZEikXy7zOiYYUz5GytPg,103
+meeting_noter/__main__.py,sha256=6sSOqH1o3jvgvkVzsVKmF6-xVGcUAbNVQkRl2CrygdE,120
+meeting_noter/cli.py,sha256=w4cBrvg58inkgbApK_z90csBfb4HWBuXveLdD3J1bc0,8540
+meeting_noter/config.py,sha256=41LFBNp5o0IojYS5Hf0FJVIr7GNn7B5O1TJDE8SQkkk,5977
+meeting_noter/daemon.py,sha256=o7U11WmdoKG58SLs70-vzS64kkjMTY5qN0F0bs0eApk,16239
+meeting_noter/meeting_detector.py,sha256=I8zzSdSSmbfd3yyCOyzPL8AS-xSHttFCagrDE35qcho,9412
+meeting_noter/menubar.py,sha256=Wmwaw-_f2Zky6-0DLC5Ql2o6_VBG1-sWacL85dNeGwU,15366
+meeting_noter/audio/__init__.py,sha256=O7PU8CxHSHxMeHbc9Jdwt9kePLQzsPh81GQU7VHCtBY,44
+meeting_noter/audio/capture.py,sha256=fDrT5oXfva8vdFlht9cv60NviKbksw2QeJ8eOtI19uE,6469
+meeting_noter/audio/encoder.py,sha256=UT9yLbURsRfVafRbT-DRianSCx1fOEwWPLP-LULYqVo,5443
+meeting_noter/audio/system_audio.py,sha256=jbHGjNCerI19weXap0a90Ik17lVTCT1hCEgRKYke-p8,13016
+meeting_noter/gui/__init__.py,sha256=z5GxxaeXyjqyEa9ox0dQxuL5u_BART0bi7cI6rfntEI,103
+meeting_noter/gui/__main__.py,sha256=A2HWdYod0bTgjQQIi21O7XpmgxLH36e_X0aygEUZLls,146
+meeting_noter/gui/app.py,sha256=COUAWu_dR5HriNYxbE86CVGS1eGYqyteH2oUFN_YtYQ,1370
+meeting_noter/gui/main_window.py,sha256=vSvNO86CHMgJf9Pem8AOdrqKyTV9ITp3W4nCoqIuAmI,1667
+meeting_noter/gui/meetings_tab.py,sha256=pqXqMv5YvCj8H6yR_TF3SMzsIDMAxyLe2otbENbM2SY,12315
+meeting_noter/gui/recording_tab.py,sha256=UlrPkUiOmkGgOKqVrfAVp4EyiY5sq86W49Y-YAhYexY,12521
+meeting_noter/gui/settings_tab.py,sha256=NUQVKDdSpyNp_MVxPLw2dB93wxaD5VeBKiDtGS4CyoU,8446
+meeting_noter/install/__init__.py,sha256=SX5vLFMrV8aBDEGW18jhaqBqJqnRXaeo0Ct7QVGDgvE,38
+meeting_noter/install/macos.py,sha256=dO-86zbNKRtt0l4D8naVn7kFWjzI8TufWLWE3FRLHQ8,3400
+meeting_noter/output/__init__.py,sha256=F7xPlOrqweZcPbZtDrhved1stBI59vnWnLYfGwdu6oY,31
+meeting_noter/output/writer.py,sha256=zO8y6FAFUAp0EEtALY-M5e2Ja5P-hgV38JjcKW7c-bA,3017
+meeting_noter/resources/__init__.py,sha256=yzHNxgypkuVDFZWv6xZjUygOVB_Equ9NNX_HGRvN7VM,43
+meeting_noter/resources/icon.icns,sha256=zMWqXCq7pI5acS0tbekFgFDvLt66EKUBP5-5IgztwPM,35146
+meeting_noter/resources/icon.png,sha256=nK0hM6CPMor_xXRFYURpm0muA2O7yzotyydUQx-ukyg,2390
+meeting_noter/resources/icon_128.png,sha256=Wg27dIpFfMzv15HG6eQpSqg72jrbGBcqLkGjOfAptO4,1169
+meeting_noter/resources/icon_16.png,sha256=GONIIZCP9FUMb_MWZWMz1_iWxsPCPOwW_XcrPF7W8KY,215
+meeting_noter/resources/icon_256.png,sha256=nK0hM6CPMor_xXRFYURpm0muA2O7yzotyydUQx-ukyg,2390
+meeting_noter/resources/icon_32.png,sha256=Bw6hJXqkd-d1OfDpFv2om7Stex7daedxnpXt32xR8Sg,344
+meeting_noter/resources/icon_512.png,sha256=o7X3ngYcppcIAAk9AcfPx94MUmrsPRp0qBTpb9SzfX8,5369
+meeting_noter/resources/icon_64.png,sha256=TqG7Awx3kK8YdiX1e_z1odZonosZyQI2trlkNZCzUoI,607
+meeting_noter/transcription/__init__.py,sha256=7GY9diP06DzFyoli41wddbrPv5bVDzH35bmnWlIJev4,29
+meeting_noter/transcription/engine.py,sha256=HK2J2QOBNIDm1MXW-gkagXP8C8cqUfK_WylHQD_LqOI,6320
+meeting_noter-0.3.0.dist-info/METADATA,sha256=FhIlKw7Wg1tRRLsFBmrXkKLa8zf3L2csAtO1bQADbcQ,7952
+meeting_noter-0.3.0.dist-info/WHEEL,sha256=wUyA8OaulRlbfwMtmQsvNngGrxQHAvkKcvRmdizlJi0,92
+meeting_noter-0.3.0.dist-info/entry_points.txt,sha256=rKNhzjSF5-e3bLRr8LVe22FeiwcacXabCvNpoEXfu4I,56
+meeting_noter-0.3.0.dist-info/top_level.txt,sha256=9Tuq04_0SXM0OXOHVbOHkHkB5tG3fqkrMrfzCMpbLpY,14
+meeting_noter-0.3.0.dist-info/RECORD,,

meeting_noter-0.3.0.dist-info/WHEEL

@@ -0,0 +1,5 @@
+Wheel-Version: 1.0
+Generator: setuptools (80.10.2)
+Root-Is-Purelib: true
+Tag: py3-none-any
+

meeting_noter-0.3.0.dist-info/entry_points.txt

@@ -0,0 +1,2 @@
+[console_scripts]
+meeting-noter = meeting_noter.cli:cli

meeting_noter-0.3.0.dist-info/top_level.txt

@@ -0,0 +1 @@
+meeting_noter