# Tdarr - Auto-Requeue Script

Automatically refills the Tdarr **staging section** whenever the number of
actively processed files drops below a configurable threshold.

## Features

| Feature                     | What it does                                                                                       |
| --------------------------- | -------------------------------------------------------------------------------------------------- |
| **Staging guard**           | Checks how many files are currently being processed.                                               |
| **Smart picker**            | Pulls up to `BATCH_SIZE` items from Status ▪ **Processed** (`table2`) whose **New Size** is `"-"`. |
| **GUI-identical requeue**   | Re-queues the files with **one** `bulk-update-files` call (same payload the web UI sends).         |
| **Retry + back-off**        | Network errors or timeouts are retried with exponential back-off.                                  |
| **Timeout-resilient**       | If `bulk-update-files` times out, the script checks if the operation succeeded anyway.             |
| **Internal scheduler**      | No cron needed – the script runs continuously and checks every `TDARR_INTERVAL_MIN`.               |
| **JSON / pretty logging**   | Toggle pretty output with `LOG_PRETTY=1`.                                                          |
| **All settings via `.env`** | No hard-coded values; perfect for CI or Docker.                                                    |

## File Selection Criteria

The script selects files based on the following logic:

* It queries the internal `table2` (equivalent to the **Status ▪ Processed** tab in the Tdarr UI).
* It filters for files that have the **“New Size” field set to `"-"`**, meaning:

  * The file was either skipped by a plugin decision or
  * The file was marked as “Transcode Success / Not Required”
  * And **no new output file** was produced (Tdarr left the file untouched).
* The top `BATCH_SIZE` of these files (sorted by Tdarr’s internal logic) are requeued.

This selection ensures that:

* You don’t accidentally requeue already optimized/transcoded files
* Only skipped or pass-through candidates get another chance (e.g., after plugin changes)

## Requirements

* Tdarr Server ≥ 2.40 (the `bulk-update-files` endpoint exists since 2024)
* Node 18+ (ESM support)

## Installation

```bash
git clone https://github.com/your-org/tdarr-auto-requeue.git
cd tdarr-auto-requeue
npm i axios, pino, minimist, dotenv
```

## Configuration

Create a `.env` file in the project root (or set the variables in your CI /
container manager):

```dotenv
# Tdarr connection
TDARR_URL=https://encode.computerliebe.org/api/v2
TDARR_API_KEY=tapi_XXXXXXXXXXXX

# Behaviour
TDARR_STAGING_LIMIT=50        # refill if staging count < 50
TDARR_BATCH_SIZE=50           # requeue at most 50 files per cycle
TDARR_INTERVAL_MIN=60         # how often the script runs (minutes)
TDARR_RETRIES=4               # API retries
TDARR_BACKOFF_MS=2000         # initial back-off in ms
BULK_TIMEOUT_MS=120000        # max wait time for bulk-update-files in ms

# Logging
LOG_LEVEL=info                # or debug / warn / error
LOG_PRETTY=1                  # pretty-printed logs
```

> **Tip:** create a dedicated API key in *Tdarr ▪ Tools ▪ API keys* with **Server
> write** permissions only.

## Usage

```bash
node tdarr_requeue.mjs            # runs once, or in interval mode if TDARR_INTERVAL_MIN is set
```

The script will automatically re-run every `TDARR_INTERVAL_MIN` minutes.
No cron, systemd, or external timers needed.

## Troubleshooting

| Symptom                         | Fix                                                                                             |
| ------------------------------- | ----------------------------------------------------------------------------------------------- |
| `401 Unauthorized`              | API key wrong or missing.                                                                       |
| `FST_ERR_VALIDATION`            | Your server expects a different payload – update Tdarr or open an issue with the error JSON.    |
| Files not visible after requeue | Refresh the **Status ▪ Queued / Staging** view; workers may take a few seconds to pick them up. |
| `ECONNABORTED` timeout          | This is handled gracefully. If the staging count increased, the script proceeds.                |
| “Too many open files”           | Increase `ulimit -n` on the host; Tdarr can be I/O intensive.                                   |

---

**Happy transcoding!**