# Tdarr Local Flow Plugin Suite

A collection of [Tdarr](https://tdarr.io/) Flow plugins for automated media transcoding with **x265 2-pass encoding**, **audio track selection**, **subtitle extraction** and **OCR fallback**.

> **Server**: Docker (Linux)  **Workers**: Windows (mapped drives)

---

## ⚡ Quick Install via curl

> **How it works**: Gitea serves a real ZIP of any branch at `/archive/<branch>.zip`.
> This is a built-in git archive — **no release workflow required**, always up to date.

### Full Suite (all 7 plugins)

```bash
PLUGIN_DIR=/path/to/tdarr/plugins/FlowPlugins

# Download a live ZIP of the main branch directly from Gitea
curl -fsSL https://git.cdtech.freeddns.org/PickleRick/Tdarr-plugin/archive/main.zip \
  -o /tmp/tdarr-plugins.zip

# Extract and copy into your plugin directory
unzip -q -o /tmp/tdarr-plugins.zip -d /tmp/tdarr-extract
cp -r /tmp/tdarr-extract/Tdarr-plugin/LocalFlowPlugins "${PLUGIN_DIR}/"
cp -r /tmp/tdarr-extract/Tdarr-plugin/FlowHelpers     "${PLUGIN_DIR}/"
rm -rf /tmp/tdarr-plugins.zip /tmp/tdarr-extract
```

After extracting, restart the Tdarr server so it picks up the new plugins:

```bash
docker restart tdarr-server
```

### One-liner (copy-paste for Docker hosts)

```bash
PLUGIN_DIR=/your/tdarr/plugins/FlowPlugins && \
curl -fsSL https://git.cdtech.freeddns.org/PickleRick/Tdarr-plugin/archive/main.zip -o /tmp/tp.zip && \
unzip -q -o /tmp/tp.zip -d /tmp/tpx && \
cp -r /tmp/tpx/Tdarr-plugin/LocalFlowPlugins "${PLUGIN_DIR}/" && \
cp -r /tmp/tpx/Tdarr-plugin/FlowHelpers     "${PLUGIN_DIR}/" && \
rm -rf /tmp/tp.zip /tmp/tpx && \
docker restart tdarr-server && echo "Done."
```

### Individual Plugin (single raw file)

Each plugin is a single `index.js` file. Download just what you need:

```bash
# Example: transcode2PassX265
BASE=https://git.cdtech.freeddns.org/PickleRick/Tdarr-plugin/raw/branch/main
PLUGIN=transcode2PassX265
CATEGORY=video
DEST=/path/to/tdarr/plugins/FlowPlugins/LocalFlowPlugins/${CATEGORY}/${PLUGIN}/1.0.0

mkdir -p "${DEST}"
curl -fsSL "${BASE}/LocalFlowPlugins/${CATEGORY}/${PLUGIN}/1.0.0/index.js" \
  -o "${DEST}/index.js"
```

| Plugin | Category | Path segment |
|---|---|---|
| `mediaAnalyzerGOP` | `utility` | `LocalFlowPlugins/utility/mediaAnalyzerGOP/1.0.0/index.js` |
| `setAnimationFlag` | `utility` | `LocalFlowPlugins/utility/setAnimationFlag/1.0.0/index.js` |
| `extractChaptersXml` | `utility` | `LocalFlowPlugins/utility/extractChaptersXml/1.0.0/index.js` |
| `audioSelectorClamping` | `audio` | `LocalFlowPlugins/audio/audioSelectorClamping/1.0.0/index.js` |
| `backupSourceAudio` | `audio` | `LocalFlowPlugins/audio/backupSourceAudio/1.0.0/index.js` |
| `extractSubtitlesAndOcr` | `subtitles` | `LocalFlowPlugins/subtitles/extractSubtitlesAndOcr/1.0.0/index.js` |
| `transcode2PassX265` | `video` | `LocalFlowPlugins/video/transcode2PassX265/1.0.0/index.js` |

### Tagged Releases (optional, requires CI push)

Once a release is published by the CI workflow, versioned ZIPs are also available:

```bash
VERSION=v1.0.1
curl -L https://git.cdtech.freeddns.org/PickleRick/Tdarr-plugin/releases/download/${VERSION}/Tdarr-plugins-${VERSION}.zip \
  -o tdarr-plugins.zip
unzip -o tdarr-plugins.zip -d "${PLUGIN_DIR}"
rm tdarr-plugins.zip
```

---

## 📁 Directory Layout

```
LocalFlowPlugins/
├── utility/
│   ├── mediaAnalyzerGOP/1.0.0/index.js
│   ├── setAnimationFlag/1.0.0/index.js
│   └── extractChaptersXml/1.0.0/index.js
├── audio/
│   ├── audioSelectorClamping/1.0.0/index.js
│   └── backupSourceAudio/1.0.0/index.js
├── subtitles/
│   └── extractSubtitlesAndOcr/1.0.0/index.js
└── video/
    └── transcode2PassX265/1.0.0/index.js
```

---

## 🔌 Plugin Reference

### 1. `mediaAnalyzerGOP` — Media Analyzer & GOP Calculator

**Category**: `utility` · **Step**: 1st in flow

Reads video stream metadata and populates `args.variables` for all downstream plugins.

| Input | Default | Description |
|---|---|---|
| `moviePathPatterns` | `/Radarr` | Pipe-separated path fragments that identify a Film library (e.g. `/Radarr\|/films`) |
| `animeSeries` | _(empty)_ | Pipe-separated anime series names to auto-detect (e.g. `Naruto\|One Piece`) |
| `filmBitrate` | `3000` | Target kbps for live-action films |
| `seriesBitrate` | `3000` | Target kbps for TV series |
| `animeBitrateHd` | `2000` | Target kbps for anime ≥720p |
| `animeBitrateSd` | `1000` | Target kbps for anime <720p |

**Outputs**: `1` → continue

**Variables set**: `gopKeyint`, `gopMinKeyint`, `isAnimation`, `isMovie`, `vBitrate`, `vMaxRate`, `vBufSize`, `x265Tune`, `videoHeight`, `videoFps`

---

### 2. `setAnimationFlag` — Set Animation Flag

**Category**: `utility` · **Step**: optional, after `mediaAnalyzerGOP`

Manually forces `isAnimation=true` and `x265Tune=animation` when a file was not auto-detected as anime. Optionally recalculates bitrate targets.

| Input | Default | Description |
|---|---|---|
| `animeBitrateHd` | `2000` | Bitrate (kbps) for anime ≥720p |
| `animeBitrateSd` | `1000` | Bitrate (kbps) for anime <720p |

**Outputs**: `1` → continue

---

### 3. `extractChaptersXml` — Extract Chapters to XML

**Category**: `utility` · **Step**: before transcode

Runs `mkvextract` to dump chapter markers to an XML file in `workDir`. Stores the path in `variables.chaptersXmlPath` for downstream use. Safe to run even when the source has no chapters (mkvextract exit 1 is treated as "no chapters").

**No inputs required.**

**Outputs**: `1` → continue

> **Requires**: MKVToolNix (`mkvextract`) — automatically located from the `mkvpropeditPath` node setting.

---

### 4. `audioSelectorClamping` — Audio Selector & Opus Bitrate Clamping

**Category**: `audio` · **Step**: after `mediaAnalyzerGOP`

Selects the best ITA and ENG audio tracks by codec+bitrate scoring, then calculates clamped Opus target bitrates. No upcoding — output bitrate never exceeds source.

**Codec scoring** (higher = better): TrueHD 100 · DTS-HD MA 95 · FLAC 90 · EAC3 75 · DTS 60 · AC3 45 · AAC 35 · Opus 30

| Input | Default | Description |
|---|---|---|
| `filmMultiChannelBitrate` | `640` | Max kbps for 5.1+ film audio |
| `seriesMultiChannelThreshold` | `384` | kbps threshold: ≥threshold → 384k, else → 256k |
| `stereoBitrateMax` | `128` | Max kbps for stereo tracks |

**Outputs**: `1` → ITA or ENG track found · `2` → no target tracks found

**Variables set**: `audioHasIta`, `audioItaStreamIndex`, `audioItaOpusBitrate`, `audioItaChannels`, `audioItaCodecName`, `audioHasEng`, `audioEngStreamIndex`, `audioEngOpusBitrate`, `audioEngCodecName`

---

### 5. `backupSourceAudio` — Backup Source Audio (DTS / TrueHD ITA)

**Category**: `audio` · **Step**: after `audioSelectorClamping`

Extracts the ITA audio track as a lossless bitstream copy (`-c:a copy`) **only when** the codec is DTS or TrueHD. Output is saved next to the source file. Silently skips if no backup is needed.

**No inputs required.**

**Outputs**: `1` → continue

**Variables set**: `audioBackupPath` (path of the extracted file, if created)

---

### 6. `extractSubtitlesAndOcr` — Extract Subtitles & OCR Fallback

**Category**: `subtitles` · **Step**: after `audioSelectorClamping`

Extracts ITA and ENG subtitle streams to SRT files **next to the source file**.

- **Text-based** codecs (`subrip`, `srt`, `ass`, `ssa`, `webvtt`, `mov_text`): direct FFmpeg extraction
- **Bitmap-based** codecs (`hdmv_pgs_subtitle`, `dvd_subtitle`, `pgssub`, `dvbsub`): FFmpeg `.sup` extract → SubtitleEdit CLI OCR → `.srt`

OCR failures are logged as **warnings** without stopping the pipeline.

**Output naming**: `<basename>.<lang>[.forced][.default][.sdh].srt`

| Input | Default | Description |
|---|---|---|
| `subtitleEditPath` | `SubtitleEdit` | Name or full path of SubtitleEdit CLI. Must be in PATH or provide absolute path. |
| `languages` | `ita,eng` | Comma-separated ISO 639-2 language codes to extract |

**Outputs**: `1` → continue (warnings in log if OCR failed)

> **Requires**: SubtitleEdit CLI for PGS/VOBSUB OCR (Windows workers only).

---

### 7. `transcode2PassX265` — 2-Pass x265 Transcode

**Category**: `video` · **Step**: last

Performs 2-pass libx265 encoding using VBV parameters from `args.variables`.

- **Video**: bound-1080p Lanczos scaling (no letterbox, even dimensions)
- **Audio**: Opus re-encode with per-track indexed bitrate (ITA default, ENG secondary)
- **Subtitles**: stripped (`-sn`)
- **Chapters**: preserved (`-map_chapters 0`)
- **2-pass logs**: automatically cleaned up after encode

| Input | Default | Description |
|---|---|---|
| `x265Preset` | `slow` | x265 preset (`ultrafast` … `veryslow`). Slower = better quality. |
| `outputContainer` | `mkv` | Output file extension |

**Outputs**: `1` → transcode complete

**Reads from variables**: `gopKeyint`, `gopMinKeyint`, `vBitrate`, `vMaxRate`, `vBufSize`, `x265Tune`, `audioHasIta`, `audioItaStreamIndex`, `audioItaOpusBitrate`, `audioHasEng`, `audioEngStreamIndex`, `audioEngOpusBitrate`

---

## 🔗 Flow Variable Map

```
mediaAnalyzerGOP ──► gopKeyint, gopMinKeyint, vBitrate, vMaxRate, vBufSize
                 ──► isAnimation, isMovie, x265Tune, videoHeight, videoFps
                          │
              (optional) setAnimationFlag ──► overrides vBitrate, vMaxRate, vBufSize, x265Tune
                          │
audioSelectorClamping ──► audioHasIta, audioItaStreamIndex, audioItaOpusBitrate, audioItaCodecName
                     ──► audioHasEng, audioEngStreamIndex, audioEngOpusBitrate
                          │
backupSourceAudio ──► audioBackupPath
                          │
transcode2PassX265 ◄── all of the above
```

---

## 🛠 Requirements

### Tdarr Server (Docker/Linux)
- Tdarr ≥ 2.11.01
- Plugins folder mounted at the path configured in Tdarr settings

### Tdarr Worker (Windows)
| Tool | Purpose | Download |
|---|---|---|
| FFmpeg (with libopus) | Video/audio/subtitle processing | https://ffmpeg.org/download.html |
| MKVToolNix | Chapter extraction (`mkvextract`) | https://mkvtoolnix.download/ |
| SubtitleEdit CLI | PGS/VOBSUB OCR to SRT | https://www.nikse.dk/subtitleedit |

Add to Windows system PATH: `ffmpeg\bin`, `MKVToolNix`, `SubtitleEdit`

Configure in Tdarr node settings:
- **FFmpeg Path**: `C:\Tools\ffmpeg\bin\ffmpeg.exe`
- **FFprobe Path**: `C:\Tools\ffmpeg\bin\ffprobe.exe`
- **MKVPropedit Path**: `C:\Program Files\MKVToolNix\mkvpropedit.exe`

---

## 🧪 Local Testing

```bash
# Prerequisites: Node.js, FFmpeg in PATH
cd /path/to/Tdarr-plugin

# 1. Generate sample test files (requires FFmpeg)
node tests/generate-test-files.js

# 2. Run logic tests (no real encode)
node tests/quick-logic-test.js

# 3. Run full integration tests
node tests/run-tests.js

# 4. Run integration tests skipping real x265 encode
node tests/run-tests.js --skip-transcode
```

---

## 📦 Release & Update

Releases are created automatically via GitHub Actions on every push to `main` or on git tag `v*`.

To trigger a manual release from the Actions tab, use **"Run workflow"** and specify the version (e.g. `v1.0.1`).

To update plugins on the server to the latest release:

```bash
VERSION=latest
PLUGIN_DIR=/path/to/tdarr/plugins/FlowPlugins

curl -L https://git.cdtech.freeddns.org/PickleRick/Tdarr-plugin/releases/download/${VERSION}/Tdarr-plugins-${VERSION}.zip \
  -o tdarr-plugins.zip
unzip -o tdarr-plugins.zip -d "${PLUGIN_DIR}"
rm tdarr-plugins.zip
docker restart tdarr-server
```