VCD

KO | EN

gitlite — search

#javascript #hacktoberfest #typescript #go #android #rust #deep-learning #docker #macos #windows #security #ios #mcp #open-source

(★ 24)

Vadana Class Downloader

VCD 최신버젼 다운로드

최종 버전 다운로드 (.zip)

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

<h1 align="center">
 
 <a href="https://github.com/IAUCourseExp/VCD">
 <img src="https://raw.githubusercontent.com/IAUCourseExp/VCD/main/vcd0.4.png" alt="VCD – Vadana Class Downloader" width="900">
 </a>
 
 VCD – Vadana Class Downloader
 
</h1>

Download, sync, and render Adobe Connect (Vadana) class recordings into a single MP4. 
 Screenshare + audio &nbsp;·&nbsp; GPU-accelerated &nbsp;·&nbsp; Audio-only export &nbsp;·&nbsp; Batch queue

<a href="#overview">Overview</a> ·
 <a href="#architecture">Architecture</a> ·
 <a href="#features">Features</a> ·
 <a href="#requirements">Requirements</a> ·
 <a href="#installation">Installation</a> ·
 <a href="#usage">Usage</a> ·
 <a href="#how-it-works">How It Works</a> ·
 <a href="#output">Output</a> ·
 <a href="#troubleshooting">Troubleshooting</a> ·
 <a href="#changelog">Changelog</a>

<a href="https://t.me/IAUCourseExp">@IAUCourseExp</a>
 &nbsp;·&nbsp;
 <a href="https://t.me/JozveIAU">@JozveIAU</a>
 &nbsp;·&nbsp;
 <a href="https://github.com/IAUCourseExp/VCD">⭐ Star this project ⭐</a>

---

## Overview

IAU Azad University's Vadana platform is built on Adobe Connect. Class recordings are stored as a ZIP archive of FLV video segments and XML timing files — not a single watchable video. Reconstructing those pieces into one correctly-synchronised MP4, with screenshare overlaid on audio and all timing offsets precisely applied, is what VCD does.

The tool covers the full pipeline: authentication, download with resume support, millisecond-accurate timeline extraction from Adobe Connect's internal `pacingTick` format, and FFmpeg rendering with hardware GPU acceleration and an audio-only fast path.

**Two files, one job:**

| File | Role |
|------|------|
| `vcd_core.py` | Download and render engine — runs standalone as a CLI |
| `vcd_gui.py` | PySide6 desktop GUI — wraps core without modifying its logic |

---

## Architecture

The codebase has one hard rule: **the GUI never modifies core's logic.** `vcd_gui.py` intercepts `sys.stdout`/`sys.stderr` to capture `core.log()` output and `tqdm` progress bars, converting them into Qt signals that drive UI updates. Core runs in a background `QThread` and knows nothing about the GUI.

### System diagram

```
┌───────────────────────────────────────────────────────────────────┐
│ vcd_gui.py — PySide6 Desktop GUI │
│ │
│ MainWindow │
│ ├─ BatchQueue sequential URL processing │
│ ├─ JobHistoryDB ~/.vcd/history.json │
│ ├─ StatsWidget speed graph · ETA · bytes │
│ ├─ TrayManager system tray + notifications │
│ └─ Worker ──────────────────────────────────────────→ QThread │
│ │ │
│ └──→ _StreamRouter │
│ hijacks sys.stdout / sys.stderr │
│ parses log() lines + tqdm bars into signals: │
│ sig_log · sig_progress · sig_speed │
│ sig_eta · sig_bytes · sig_done │
└──────────────────────────┬────────────────────────────────────────┘
 │ calls (never modifies)
┌──────────────────────────▼────────────────────────────────────────┐
│ vcd_core.py — Download & Render Engine │
│ │
│ URL │
│ └─→ acquire_authenticated_session() │
│ ?session= token → manual cookie → interactive prompt │
│ └─→ download_and_extract() │
│ HTTPS stream · HTTP Range resume · ZIP extraction │
│ └─→ collect_media_intervals() │
│ parse pacingTick XML → millisecond-accurate clip offsets │
│ └─→ write_timeline_xml() │
│ serialise for inspection / reuse │
│ │
│ ┌──────────────────────────────┐ │
│ │ render_video_from_timeline │──→ Class-<id>.mp4 │
│ │ FilterGraphBuilder │ GPU: NVENC/AMF/QSV │
│ │ _video_encoder_args() │ CPU: libx264 │
│ └──────────────────────────────┘ │
│ ┌──────────────────────────────┐ │
│ │ export_audio() │──→ Class-<id>.m4a │
│ │ amix · silenceremove │ audio-only fast path │
│ └──────────────────────────────┘ │
└───────────────────────────────────────────────────────────────────┘
```

### Core components

| Component | Responsibility |
|-----------|---------------|
| `acquire_authenticated_session()` | Auth chain: URL `?session=` → CLI cookie → interactive prompt |
| `download_and_extract()` | HTTPS streaming with `Range` header resume; ZIP validation and extraction |
| `collect_media_intervals()` | Reads `pacingTick` from each XML, computes `start_ms` / `end_ms` per clip against a global base tick |
| `write_timeline_xml()` | Serialises computed timeline with exact offsets for every segment |
| `render_video_from_timeline()` | Builds FFmpeg `-filter_complex`: black canvas + video overlays + `amix` |
| `export_audio()` | Audio-only: `amix normalize=0` + optional `silenceremove` → AAC output |
| `FilterGraphBuilder` | Constructs filter graph strings segment by segment — no manual string concatenation |
| `_video_encoder_args()` | Returns encoder-specific CLI args for CPU / NVIDIA NVENC / AMD AMF / Intel QSV |
| `_current_proc` / `_current_response` | Module-level refs for cross-thread kill: `proc.terminate()` + `resp.close()` |

### GUI threading model

```
UI Thread                         Worker Thread
─────────────────────             ─────────────────────────────────
MainWindow._start()
  → Worker.run() ───────────────→ sys.stdout = _StreamRouter
                                  core.init_tools()
                                  core.download_and_extract()
                                  core.process_recording()
                                  sys.stdout = old_stdout

← sig_log(level, msg)           ← _StreamRouter.write() ← core.log()
← sig_progress(mode, pct)       ← tqdm bar pattern matched
← sig_speed("2.4 MB/s")         ← speed parsed from tqdm line
← sig_eta("01:23")              ← remaining time parsed
← sig_bytes("142 MB", "380 MB") ← bytes parsed from tqdm line
← sig_done(ok, msg, out_path)   ← emitted at end of Worker.run()
```

### Stop mechanism

Stopping immediately requires interrupting work across two layers:

| Stage | Method | Effect |
|-------|--------|--------|
| Download | `_current_response.close()` | Closes the socket; `iter_content()` raises immediately |
| FFmpeg render | `_current_proc.terminate()` | SIGTERM to the running ffmpeg process |

Both calls happen in `Worker.cancel()`, triggered by `_stop()` on the UI thread.

---

## Features

### Current — v0.4

#### Download

- HTTP `Range` header resume — continues a partial download from the byte offset on disk
- Instant stop — closes socket and kills FFmpeg without waiting for the current stage to finish
- Auto-retry on failure, up to 3× with increasing delay
- Disk space check before starting

#### Render

- GPU encoding: **NVIDIA NVENC**, **AMD AMF**, **Intel QSV** — detected at launch, auto-selected, validated against your FFmpeg build
- Header chip shows detected GPU hardware; selecting an unavailable encoder shows an error and reverts the choice
- CPU fallback (`libx264`) always available
- Volume fix: `amix normalize=0` (prior versions halved audio volume when mixed with the silence padding track)

#### Audio export

- Audio-only mode — exports `.m4a` without any video processing, 10–20× faster
- Silence trimming — removes gaps longer than 1.5 s via FFmpeg's `silenceremove` filter

#### GUI

- Batch queue — add multiple URLs, process sequentially with automatic continuation
- Job history — every job logged to `~/.vcd/history.json`, viewable in the History tab
- Output files tab — thumbnail preview, play, open folder, delete
- Real-time speed graph with ETA and bytes downloaded
- System tray — minimize to background, desktop notification on finish
- Log panel — per-level filter (ALL / INFO / STEP / WARN / ERROR) + free-text search
- Cookie profiles — save and reload multiple BREEZESESSION sessions
- URL autocomplete — last 30 entries
- Clipboard URL detection — auto-fills on launch if a class link is copied
- Custom FFmpeg path — bypass PATH lookup with a specific binary
- Auto-open output when done
- Resizable splitter between controls panel and log

#### Keyboard shortcuts

| Key | Action |
|-----|--------|
| `Ctrl+Enter` | Start job |
| `Esc` | Stop job |
| `Ctrl+Shift+Q` | Add current URL to batch queue |
| `Ctrl+L` | Log tab |
| `Ctrl+H` | History tab |
| `Ctrl+F` | Focus log search |
| `Ctrl+S` | Save log to file |
| `Ctrl+Q` | Quit |

---

## Requirements

- **Python** 3.10 or newer
- **FFmpeg + FFprobe** in PATH → [ffmpeg.org](https://ffmpeg.org/download.html)
- **PySide6** — GUI only

**GPU encoding** (optional — 3–8× faster rendering):

| GPU | Encoder | Platform |
|-----|---------|---------|
| NVIDIA | `h264_nvenc` | Windows, Linux |
| AMD | `h264_amf` | Windows |
| Intel | `h264_qsv` | Windows, Linux |
| CPU | `libx264` | All (default) |

> For GPU support, use the `win64-gpl` build from [BtbN/FFmpeg-Builds](https://github.com/BtbN/FFmpeg-Builds/releases) — it ships with all three hardware encoders.

---

## Installation

```bash
git clone https://github.com/IAUCourseExp/VCD
cd VCD
pip install -r requirements.txt
```

`requirements.txt`:

```
requests
urllib3
tqdm
colorama
PySide6              # GUI only — safe to omit for CLI
pyfiglet             # optional — ASCII banner in CLI
beautifulsoup4       # optional — HTML form auth fallback
browser-cookie3      # optional — auto-extract browser cookies
```

Optional packages degrade gracefully when missing.

---

## Usage

### GUI

```bash
python vcd_gui.py
```

On launch: FFmpeg and FFprobe are located, the best available GPU encoder is auto-detected and selected, and the URL field is auto-filled if a class link is in the clipboard.

Paste a class URL and click **Download & Render**.  
For batch processing: **+ Add to Queue** → build a list → **Run Queue**.

### CLI

```bash
# Video — default
python vcd_core.py "https://vadavc32.ec.iau.ir/<id>/?session=TOKEN&proto=true"

# Custom output filename
python vcd_core.py --output lecture_week3.mp4 "https://..."

# Audio only
python vcd_core.py --audio-only "https://..."

# Audio only, silence removed
python vcd_core.py --audio-only --trim-silence "https://..."

# Timeline XML only, skip render
python vcd_core.py --xml-only "https://..."
```

**All CLI flags:**

```
python vcd_core.py [OPTIONS] URL

--output FILE Output filename (default: Class-<id>.mp4 or .m4a)
 --cookie VALUE BREEZESESSION value or full cookie string
 --xml-only Write timeline.xml only, skip render
 --audio-only Export .m4a without video processing
 --trim-silence Remove silent gaps (requires --audio-only)
 --crf INT Quality — lower = better (default: 30)
 --fps INT Frame rate (default: 30)
```

### Authentication

Tried in order — the first working method is used:

1. **`?session=` in URL** — paste the full recording link; the tool extracts the token automatically
2. **`--cookie` flag / Cookie field** — your `BREEZESESSION` value (browser DevTools → Application → Cookies → your server domain)
3. **Interactive prompt** — CLI only; guides you through copying the cookie manually when no other method works

---

## How It Works

1. **Authentication** — establishes a session via the chain above
2. **Download** — fetches `/<id>/output/<id>.zip` over HTTPS; sends `Range: bytes=<offset>-` if a partial file exists on disk from a prior attempt
3. **Extraction** — validates and unzips FLV and XML files into `<id>/`
4. **Timing extraction** — reads `pacingTick` events from each XML file to compute a shared millisecond timebase; each clip gets an exact `start_ms` and `end_ms` relative to the recording start
5. **Segment building** — determines which screenshare and audio clip is active at each point in time, handling gaps and overlaps
6. **Timeline serialisation** — writes `<id>/timeline.xml` with `start`, `end`, and `offset` for every media segment; reusable for debugging or re-render without re-downloading
7. **Render** — builds FFmpeg filter graph:
 - Black canvas for the full recording duration
 - Each screenshare scaled and overlaid with a PTS offset (`setpts` + `overlay`)
 - Each audio clip delayed to its start time (`adelay`) then mixed with `amix normalize=0`
 - Encoded with the selected GPU encoder or CPU `libx264`
8. **Output** — `Class-<id>.mp4`; or `.m4a` for audio-only (skips steps 5–7 for video)

---

## Output

| File | Description |
|------|-------------|
| `Class-<id>.mp4` | Final synced video |
| `Class-<id>.m4a` | Audio-only export |
| `<id>/timeline.xml` | Serialised timeline — inspect or reuse for re-render |
| `<id>/` | Extracted raw files — safe to delete after a successful render |

---

## Troubleshooting

| Problem | Cause & fix |
|---------|-------------|
| `AuthenticationError` — "Could not authenticate" | Session expired. Get a fresh class link or copy a fresh `BREEZESESSION` from browser DevTools → Application → Cookies. |
| `DownloadError` — HTTP 403 / 404 | Recording may no longer exist or requires a different account. Open the ZIP URL in your browser while logged in to verify access. |
| `ToolNotFoundError` — ffmpeg/ffprobe not found | Add FFmpeg's `bin` folder to PATH, or set a custom path in GUI → Advanced → FFmpeg path. |
| `MediaProcessingError` — "No media files with a valid pacingTick" | Extracted folder may be empty or corrupt from an earlier interrupted download. Delete `<id>/` and retry. |
| GPU option grayed out | That encoder is not in your FFmpeg build. Use the `win64-gpl` build from [BtbN/FFmpeg-Builds](https://github.com/BtbN/FFmpeg-Builds/releases). |
| Audio is very quiet | Affects v0.3 and earlier — `amix` was halving volume. Update to v0.4. |
| Download restarts from 0 after interruption | Affects v0.3 and earlier — no resume support. Update to v0.4. |
| No screenshare in output | The class had no screenshare stream, only audio. Use `--audio-only`. |

---

## Changelog

> **For maintainers:** to add a new release, paste a new `### vX.Y — YYYY-MM-DD` block
> at the **top** of this section. Keep the structure: one-line summary, then bullet
> groups by category (New / Fixed / Changed). Dates go in the header, not in bullets.

---

### v0.4

*Audio export · download resume · GPU validation · six bug fixes.*

**New**
- `--audio-only` + GUI checkbox: export `.m4a` without any video processing — 10–20× faster, ~95% smaller file size
- `--trim-silence` + GUI checkbox: remove long silent gaps using FFmpeg's `silenceremove` filter
- HTTP `Range` resume: an interrupted download continues from the existing partial file on disk
- GPU auto-detection: header chip shows detected hardware; selecting an unavailable encoder shows an error dialog and reverts
- Auto-open output file when done, custom FFmpeg path in Advanced, clipboard URL auto-fill on launch

**Fixed**
- `amix normalize=0` — audio was 6 dB too quiet in all prior versions (`amix` default divides output by input count)
- `@retry` on `_stream_to_file` — manual Stop during download was delayed 5–8 s by retry backoff; decorator removed
- Double `_auto_retry()` call in `_on_done` — "Max retries reached" message appeared twice in the log
- Auto-retry fired after manual Stop — `_stop_requested` flag was not checked before scheduling retry
- `_stop_requested` not initialised in `__init__` — could raise `AttributeError` before first job
- Dead imports (`getpass`, `random`, `Callable`, `urlencode`) and duplicate CLI code removed

---

### v0.3

*First GUI release.*

**New**
- `vcd_gui.py` — PySide6 desktop GUI wrapping the unchanged v0.2 core
- Quality presets: Ultra (1080p) / High (720p) / Balanced (720p) / Compact (480p) / Custom
- GPU encoder selection: NVIDIA NVENC · AMD AMF · Intel QSV · CPU
- Batch queue, job history (`~/.vcd/history.json`), output files tab with thumbnail extraction
- Real-time speed graph with ETA and bytes downloaded
- System tray with notifications, cookie profiles, URL autocomplete
- Log panel with per-level filter and free-text search
- Auto-retry on failure, disk space check, resizable splitter

---

### v0.2

*Authentication and network resilience.*

**New**
- Auth chain: `?session=` URL token → `--cookie` flag → browser cookies (`browser_cookie3`)
- Automatic retry with exponential backoff for unreliable connections
- `argparse` CLI: `--cookie`, `--output`, `--crf`, `--fps`, `--xml-only`
- `ToolManager`, `RenderConfig`, `DownloadConfig` dataclasses
- Custom exceptions: `AuthenticationError`, `DownloadError`, `MediaProcessingError`, `ToolNotFoundError`

---

### v0.1 — Initial release

- ZIP download and extraction (unauthenticated)
- `pacingTick` alignment for screenshare + audio synchronisation
- `timeline.xml` generation and single-pass FFmpeg render

---

## Contributing

Pull requests and issues are welcome.  
Please open an issue before implementing large changes to align on approach first.

---

## License

MIT — see [LICENSE](LICENSE) for details.

---

<a href="https://t.me/IAUCourseExp">@IAUCourseExp</a>
 &nbsp;·&nbsp;
 <a href="https://t.me/JozveIAU">@JozveIAU</a>
 &nbsp;·&nbsp;
 <a href="https://github.com/IAUCourseExp/VCD">⭐ Star on GitHub</a>

VCD

VCD 최신버젼 다운로드

오늘의 추천 프로젝트

Was this content helpful?

Recent Feedback

VCD

VCD 최신버젼 다운로드

오늘의 추천 프로젝트

Was this content helpful?

Recent Feedback

Download README