erster docker release

.gitignore

@@ -1,6 +1,8 @@
-# Benutzerspezifische Konfiguration (aus .dist erstellt)
+# Persistente Daten (DB, Config, Recordings)
-config/stream-recorder.conf
+data/
 
-# Job-Dateien (nur Vorlagen werden versioniert)
+# Python
-jobs/*.job
+__pycache__/
-!jobs/templates/
+*.pyc
+*.pyo
+.venv/

Dockerfile

@@ -0,0 +1,27 @@
+FROM python:3.12-slim
+
+RUN apt-get update && \
+    apt-get install -y --no-install-recommends ffmpeg curl && \
+    rm -rf /var/lib/apt/lists/*
+
+WORKDIR /app
+
+COPY bin/backend/requirements.txt .
+RUN pip install --no-cache-dir -r requirements.txt
+
+COPY bin/backend/ /app/
+COPY bin/frontend/ /app/frontend/
+COPY config/config.yml.dist /app/config.yml.dist
+
+RUN mkdir -p /app/data/recordings
+
+ENV DATA_DIR=/app/data
+ENV STATIC_DIR=/app/frontend/static
+ENV PYTHONUNBUFFERED=1
+
+EXPOSE 8484
+
+HEALTHCHECK --interval=30s --timeout=5s --start-period=10s --retries=3 \
+    CMD curl -f http://localhost:8484/api/health || exit 1
+
+CMD ["python", "app.py"]

README.md

@@ -7,7 +7,7 @@
 <h1 align="center">stream-recorder</h1>
 
 <h4 align="center">
-  Generischer Stream-Recorder für RTMP, HLS, HTTP-Audio und weitere Formate
+  Generischer Stream-Recorder für RTMP, HLS, HTTP-Audio und weitere Formate — mit Web-UI und Docker
 </h4>
 
 <h6 align="center">
@@ -23,88 +23,73 @@
 
 ## Funktionen
 
+- **Web-Interface** — Jobs planen, Aufnahmen steuern und verlängern im Browser
+- **Docker-basiert** — Ein Container, läuft auf ARM und x86/x64
 - **Mehrere Stream-Typen** — RTMP, RTMPS, HLS (m3u8), MP3, AAC, OGG, FLAC und HTTP-Streams
-- **Job-basiert** — Ein Stream = eine Job-Datei, inklusive Zeitplanung
+- **Aufnahmen verlängern** — Laufende Aufnahmen spontan verlängern oder auf unbegrenzt setzen
-- **Segment-Splitting** — Lange Aufnahmen automatisch in Teile zerlegen (z.B. stündlich)
+- **Zeitplanung** — Automatische Aufnahmezeiten direkt im Web-UI konfigurieren
-- **Plik-Upload** — Aufnahmen und Segmente automatisch auf Plik hochladen
+- **Segment-Splitting** — Lange Aufnahmen automatisch in Teile zerlegen
-- **Zeitplanung** — Aufnahmezeiten direkt in der Job-Datei konfigurieren (kein Cron nötig)
 - **Auto-Reconnect** — Automatische Wiederverbindung bei Verbindungsabbrüchen
-- **Live-Status** — Laufende Aufnahmen mit Laufzeit, Dateigröße und Segment-Anzahl anzeigen
+- **Schnellaufnahme** — URL eingeben, sofort aufnehmen
-- **NTFY-Benachrichtigungen** — Push-Nachrichten bei Start, Stop, Fehler und Upload
+- **NTFY-Benachrichtigungen** — Push-Nachrichten bei Start, Stop, Fehler und Upload (pro Job schaltbar)
-- **Schnellaufnahme** — Interaktiver Modus für spontane Aufnahmen
+- **Plik-Upload** — Aufnahmen automatisch hochladen, optional danach löschen (pro Job schaltbar)
-
+- **Metadata-Monitor** — Automatische Show-Erkennung via Stream-Metadaten (Plugin)
-## Voraussetzungen
+- **REST-API** — Alle Funktionen auch per API steuerbar
-
-- **Linux** (getestet unter Ubuntu/Debian)
-- **Bash** 4.0+
-- **ffmpeg** (`sudo apt install ffmpeg`)
-- **curl** (optional, für NTFY und Plik)
 
 ## Schnellstart
 
 ```bash
-# Klonen und einrichten
+# Repository klonen
-git clone https://git.techniverse.net/scriptos/stream-recorder.git /opt/stream-recorder
+git clone https://git.techniverse.net/scriptos/stream-recorder.git
-cd /opt/stream-recorder
+cd stream-recorder
-cp config/stream-recorder.conf.dist config/stream-recorder.conf
-nano config/stream-recorder.conf
 
-# Scheduler installieren (optional)
+# Container starten
-sudo install/install.sh
+docker compose up -d
-
-# Ersten Job erstellen
-./stream-recorder.sh create
-
-# Oder direkt loslegen
-./stream-recorder.sh quick
 ```
 
-## Befehle
+Das Web-UI ist unter `http://localhost:8484` erreichbar. Die Konfiguration wird beim ersten Start automatisch in `data/config.yml` erstellt.
+
+## Docker Compose
+
+```yaml
+services:
+  stream-recorder:
+    image: stream-recorder:latest
+    container_name: stream-recorder
+    restart: unless-stopped
+    ports:
+      - "8484:8484"
+    volumes:
+      - ./data:/app/data
+    environment:
+      TZ: Europe/Berlin
+```
+
+## Daten
+
+Alle persistenten Daten liegen in einem einzigen Volume (`./data`):
+
+| Pfad | Beschreibung |
+|------|-------------|
+| `data/config.yml` | Konfiguration (wird beim ersten Start automatisch erstellt) |
+| `data/stream-recorder.db` | SQLite-Datenbank |
+| `data/recordings/` | Aufnahmen (pro Job ein Unterordner) |
+
+## Multi-Arch Build
 
 ```bash
-./stream-recorder.sh record [job]    # Aufnahme starten (ohne Angabe: Auswahl)
+docker buildx build --platform linux/amd64,linux/arm64,linux/arm/v7 -t stream-recorder:latest .
-./stream-recorder.sh stop [job]      # Aufnahme stoppen (ohne Angabe: Auswahl)
-./stream-recorder.sh status          # Laufende Aufnahmen anzeigen
-./stream-recorder.sh list            # Alle Jobs anzeigen
-./stream-recorder.sh create          # Neuen Job erstellen
-./stream-recorder.sh quick           # Schnellaufnahme
-./stream-recorder.sh logs            # Logs live verfolgen
 ```
 
 ## Dokumentation
 
 | Thema | Beschreibung |
 |-------|-------------|
-| [Installation](docs/installation.md) | Installation, Updates, Deinstallation |
+| [Jobs](docs/jobs.md) | Jobs erstellen, Zeitplanung, Schnellaufnahme |
-| [Konfiguration](docs/konfiguration.md) | Alle Konfigurationsparameter |
+| [Aufnahmen](docs/aufnahmen.md) | Verlängern, Stoppen, Reconnect, Segment-Splitting |
-| [Job-Dateien](docs/jobs.md) | Jobs erstellen und verwalten |
+| [Plugins](docs/plugins.md) | NTFY, Plik-Upload, Metadata-Monitor |
-| [Zeitplanung](docs/zeitplanung.md) | Automatische Aufnahmezeiten |
+| [Konfiguration](docs/konfiguration.md) | config.yml, alle Parameter |
-| [Segment-Splitting](docs/segment-splitting.md) | Aufnahmen in Teile zerlegen |
+| [API](docs/api.md) | REST-API Endpunkte |
-| [Plik-Upload](docs/plik-upload.md) | Automatischer Upload auf Plik |
-| [Benachrichtigungen](docs/benachrichtigungen.md) | NTFY Push-Nachrichten |
-| [Befehle](docs/befehle.md) | Alle Befehle im Detail |
-| [Beispiele](docs/beispiele.md) | Praxisbeispiele und Workflows |
-
-## Verzeichnisstruktur
-
-```
-stream-recorder/
-├── stream-recorder.sh              # Hauptscript
-├── config/
-│   └── stream-recorder.conf.dist   # Konfigurations-Vorlage
-├── jobs/                            # Job-Definitionen
-│   └── templates/                   # Vorlagen
-│       ├── example-rtmp.job.dist
-│       ├── example-hls.job.dist
-│       └── example-mp3.job.dist
-├── docs/                            # Dokumentation
-├── install/
-│   ├── install.sh                   # Scheduler installieren
-│   └── uninstall.sh                 # Scheduler deinstallieren
-├── README.md
-└── LICENSE
-```
-
 
 <br><br>
 <p align="center">

bin/backend/app.py

@@ -0,0 +1,369 @@
+import base64
+import contextlib
+import datetime
+import os
+import secrets
+import sys
+
+from fastapi import FastAPI, HTTPException, Request, Query
+from fastapi.middleware.cors import CORSMiddleware
+from fastapi.responses import HTMLResponse, JSONResponse
+from fastapi.staticfiles import StaticFiles
+from pydantic import BaseModel
+from typing import Optional
+
+from config import Config
+from models import init_db, get_session, Job, Recording as RecordingModel, LogEntry
+from recorder import RecordingManager, format_elapsed
+from scheduler import StreamScheduler
+
+DATA_DIR = os.environ.get("DATA_DIR", "/app/data")
+STATIC_DIR = os.environ.get("STATIC_DIR", "/app/frontend/static")
+
+os.makedirs(DATA_DIR, exist_ok=True)
+
+config = Config(os.path.join(DATA_DIR, "config.yml"))
+init_db(os.path.join(DATA_DIR, "stream-recorder.db"))
+
+recorder = RecordingManager(config)
+scheduler = StreamScheduler(recorder)
+
+
+@contextlib.asynccontextmanager
+async def lifespan(app):
+    recorder.cleanup_stale()
+    scheduler.start()
+    yield
+    scheduler.stop()
+
+
+app = FastAPI(title="stream-recorder", version="3.0.0", lifespan=lifespan)
+
+app.add_middleware(
+    CORSMiddleware,
+    allow_origins=["*"],
+    allow_methods=["*"],
+    allow_headers=["*"],
+)
+
+
+@app.middleware("http")
+async def auth_middleware(request: Request, call_next):
+    if not config.get("auth", "enabled", default=False):
+        return await call_next(request)
+
+    if request.url.path.startswith("/api/health"):
+        return await call_next(request)
+
+    auth_header = request.headers.get("Authorization", "")
+    if auth_header.startswith("Basic "):
+        try:
+            decoded = base64.b64decode(auth_header[6:]).decode("utf-8")
+            username, password = decoded.split(":", 1)
+            expected_user = config.get("auth", "username", default="admin")
+            expected_pass = config.get("auth", "password", default="stream-recorder")
+            if secrets.compare_digest(username, expected_user) and secrets.compare_digest(password, expected_pass):
+                return await call_next(request)
+        except Exception:
+            pass
+
+    return JSONResponse(
+        status_code=401,
+        content={"detail": "Unauthorized"},
+        headers={"WWW-Authenticate": 'Basic realm="stream-recorder"'},
+    )
+
+
+# --- Health ---
+
+@app.get("/api/health")
+async def health():
+    return {"status": "ok", "version": "3.0.0"}
+
+
+# --- Jobs CRUD ---
+
+class JobCreate(BaseModel):
+    name: str
+    stream_url: str
+    stream_type: str = "auto"
+    output_format: str = ""
+    max_duration: Optional[int] = None
+    extra_ffmpeg_args: str = ""
+    segment_duration: Optional[int] = None
+    schedule_enabled: bool = False
+    schedule_once: bool = False
+    schedule_date: str = ""
+    schedule_days: str = "*"
+    schedule_start: str = ""
+    schedule_stop: str = ""
+    ntfy_enabled: bool = False
+    plik_enabled: bool = False
+    delete_after_upload: bool = False
+    metadata_monitor_enabled: bool = False
+    metadata_pattern: str = ""
+    metadata_grace_period: int = 300
+    metadata_poll_interval: int = 30
+
+
+class JobUpdate(JobCreate):
+    pass
+
+
+@app.get("/api/jobs")
+async def list_jobs():
+    session = get_session()
+    try:
+        jobs = session.query(Job).order_by(Job.id).all()
+        result = []
+        for j in jobs:
+            is_recording = session.query(
+                session.query(RecordingModel).filter_by(job_id=j.id, status="running").exists()
+            ).scalar()
+
+            result.append({
+                "id": j.id,
+                "name": j.name,
+                "stream_url": j.stream_url,
+                "stream_type": j.stream_type,
+                "output_format": j.output_format,
+                "max_duration": j.max_duration,
+                "extra_ffmpeg_args": j.extra_ffmpeg_args,
+                "segment_duration": j.segment_duration,
+                "schedule_enabled": j.schedule_enabled,
+                "schedule_once": j.schedule_once,
+                "schedule_date": j.schedule_date,
+                "schedule_days": j.schedule_days,
+                "schedule_start": j.schedule_start,
+                "schedule_stop": j.schedule_stop,
+                "ntfy_enabled": j.ntfy_enabled,
+                "plik_enabled": j.plik_enabled,
+                "delete_after_upload": j.delete_after_upload,
+                "metadata_monitor_enabled": j.metadata_monitor_enabled,
+                "metadata_pattern": j.metadata_pattern,
+                "metadata_grace_period": j.metadata_grace_period,
+                "metadata_poll_interval": j.metadata_poll_interval,
+                "is_recording": is_recording,
+                "created_at": j.created_at.isoformat() if j.created_at else None,
+            })
+        return result
+    finally:
+        session.close()
+
+
+@app.post("/api/jobs")
+async def create_job(data: JobCreate):
+    session = get_session()
+    try:
+        job = Job(**data.model_dump())
+        session.add(job)
+        session.commit()
+        return {"id": job.id, "message": "Job erstellt"}
+    finally:
+        session.close()
+
+
+@app.get("/api/jobs/{job_id}")
+async def get_job(job_id: int):
+    session = get_session()
+    try:
+        job = session.query(Job).filter_by(id=job_id).first()
+        if not job:
+            raise HTTPException(404, "Job nicht gefunden")
+        return {
+            "id": job.id, "name": job.name, "stream_url": job.stream_url,
+            "stream_type": job.stream_type, "output_format": job.output_format,
+            "max_duration": job.max_duration, "extra_ffmpeg_args": job.extra_ffmpeg_args,
+            "segment_duration": job.segment_duration,
+            "schedule_enabled": job.schedule_enabled, "schedule_once": job.schedule_once,
+            "schedule_date": job.schedule_date, "schedule_days": job.schedule_days,
+            "schedule_start": job.schedule_start, "schedule_stop": job.schedule_stop,
+            "ntfy_enabled": job.ntfy_enabled,
+            "plik_enabled": job.plik_enabled, "delete_after_upload": job.delete_after_upload,
+            "metadata_monitor_enabled": job.metadata_monitor_enabled,
+            "metadata_pattern": job.metadata_pattern,
+            "metadata_grace_period": job.metadata_grace_period,
+            "metadata_poll_interval": job.metadata_poll_interval,
+        }
+    finally:
+        session.close()
+
+
+@app.put("/api/jobs/{job_id}")
+async def update_job(job_id: int, data: JobUpdate):
+    session = get_session()
+    try:
+        job = session.query(Job).filter_by(id=job_id).first()
+        if not job:
+            raise HTTPException(404, "Job nicht gefunden")
+        for k, v in data.model_dump().items():
+            setattr(job, k, v)
+        job.updated_at = datetime.datetime.utcnow()
+        session.commit()
+        return {"message": "Job aktualisiert"}
+    finally:
+        session.close()
+
+
+@app.delete("/api/jobs/{job_id}")
+async def delete_job(job_id: int):
+    session = get_session()
+    try:
+        job = session.query(Job).filter_by(id=job_id).first()
+        if not job:
+            raise HTTPException(404, "Job nicht gefunden")
+        session.delete(job)
+        session.commit()
+        return {"message": "Job gelöscht"}
+    finally:
+        session.close()
+
+
+@app.post("/api/jobs/{job_id}/record")
+async def start_job_recording(job_id: int):
+    session = get_session()
+    try:
+        job = session.query(Job).filter_by(id=job_id).first()
+        if not job:
+            raise HTTPException(404, "Job nicht gefunden")
+        job_data = {
+            "name": job.name, "stream_url": job.stream_url,
+            "stream_type": job.stream_type, "output_format": job.output_format,
+            "max_duration": job.max_duration, "extra_ffmpeg_args": job.extra_ffmpeg_args,
+            "segment_duration": job.segment_duration,
+            "metadata_monitor_enabled": job.metadata_monitor_enabled,
+            "metadata_pattern": job.metadata_pattern,
+            "metadata_grace_period": job.metadata_grace_period or 300,
+            "metadata_poll_interval": job.metadata_poll_interval or 30,
+            "ntfy_enabled": job.ntfy_enabled,
+            "plik_enabled": job.plik_enabled,
+            "delete_after_upload": job.delete_after_upload,
+        }
+    finally:
+        session.close()
+
+    result = recorder.start_recording(job_id=job_id, **job_data)
+    return result
+
+
+# --- Quick Recording ---
+
+class QuickRecordRequest(BaseModel):
+    url: str
+    name: str = ""
+    stream_type: str = "auto"
+    max_duration: Optional[int] = None
+
+
+@app.post("/api/quick-record")
+async def quick_record(data: QuickRecordRequest):
+    name = data.name or f"Schnellaufnahme_{datetime.datetime.now().strftime('%H%M%S')}"
+    result = recorder.start_recording(
+        job_id=None,
+        name=name,
+        stream_url=data.url,
+        stream_type=data.stream_type,
+        max_duration=data.max_duration,
+    )
+    return result
+
+
+# --- Recordings ---
+
+@app.get("/api/recordings")
+async def list_recordings(active_only: bool = False, limit: int = 50):
+    if active_only:
+        return recorder.get_all_active()
+    return recorder.get_all_recordings(limit=limit)
+
+
+@app.get("/api/recordings/{rec_id}")
+async def get_recording(rec_id: int):
+    status = recorder.get_status(rec_id)
+    if not status:
+        raise HTTPException(404, "Aufnahme nicht gefunden")
+    return status
+
+
+@app.post("/api/recordings/{rec_id}/stop")
+async def stop_recording(rec_id: int):
+    recorder.stop_recording(rec_id)
+    return {"message": "Aufnahme gestoppt"}
+
+
+@app.post("/api/recordings/{rec_id}/extend")
+async def extend_recording(rec_id: int, minutes: int = Query(default=30)):
+    ok = recorder.extend_recording(rec_id, minutes * 60)
+    if not ok:
+        raise HTTPException(400, "Aufnahme kann nicht verlängert werden")
+    return {"message": f"Aufnahme um {minutes} Minuten verlängert"}
+
+
+@app.post("/api/recordings/{rec_id}/unlimited")
+async def remove_limit(rec_id: int):
+    ok = recorder.remove_duration_limit(rec_id)
+    if not ok:
+        raise HTTPException(400, "Zeitlimit kann nicht entfernt werden")
+    return {"message": "Zeitlimit entfernt"}
+
+
+# --- Logs ---
+
+@app.get("/api/logs")
+async def get_logs(limit: int = 100, level: str = ""):
+    session = get_session()
+    try:
+        q = session.query(LogEntry).order_by(LogEntry.id.desc())
+        if level:
+            q = q.filter(LogEntry.level == level.upper())
+        logs = q.limit(limit).all()
+        return [
+            {
+                "id": l.id,
+                "timestamp": l.timestamp.isoformat() if l.timestamp else None,
+                "level": l.level,
+                "message": l.message,
+                "job_name": l.job_name,
+            }
+            for l in logs
+        ]
+    finally:
+        session.close()
+
+
+# --- Config ---
+
+@app.get("/api/config")
+async def get_config():
+    safe_config = {**config.data}
+    if "auth" in safe_config:
+        safe_auth = {**safe_config["auth"]}
+        if safe_auth.get("password"):
+            safe_auth["password"] = "***"
+        safe_config["auth"] = safe_auth
+    return safe_config
+
+
+class ConfigUpdate(BaseModel):
+    data: dict
+
+
+@app.put("/api/config")
+async def update_config(payload: ConfigUpdate):
+    from config import _deep_merge
+    _deep_merge(config.data, payload.data)
+    config.save()
+    return {"message": "Konfiguration gespeichert"}
+
+
+# --- Static files (Frontend) ---
+
+if os.path.isdir(STATIC_DIR):
+    app.mount("/", StaticFiles(directory=STATIC_DIR, html=True), name="static")
+
+
+if __name__ == "__main__":
+    import uvicorn
+    host = config.get("server", "host", default="0.0.0.0")
+    port = config.get("server", "port", default=8484)
+    uvicorn.run("app:app", host=host, port=port, reload=False)

bin/backend/config.py

@@ -0,0 +1,78 @@
+import os
+import yaml
+
+DEFAULT_CONFIG = {
+    "recording": {
+        "download_path": "/app/data/recordings",
+        "max_retries": 5,
+        "retry_delay": 5,
+    },
+    "server": {
+        "host": "0.0.0.0",
+        "port": 8484,
+    },
+    "auth": {
+        "enabled": False,
+        "username": "admin",
+        "password": "stream-recorder",
+    },
+    "ntfy": {
+        "url": "",
+        "token": "",
+        "events": "error",
+    },
+    "plik": {
+        "url": "",
+        "api_key": "",
+        "ttl": "30d",
+    },
+    "logging": {
+        "level": "INFO",
+    },
+}
+
+
+class Config:
+    def __init__(self, config_path: str = "/app/data/config.yml"):
+        self.config_path = config_path
+        self.data = {}
+        self.load()
+
+    def load(self):
+        self.data = _deep_copy(DEFAULT_CONFIG)
+        if os.path.exists(self.config_path):
+            with open(self.config_path, "r") as f:
+                user_config = yaml.safe_load(f) or {}
+            _deep_merge(self.data, user_config)
+        else:
+            self.save()
+
+    def save(self):
+        os.makedirs(os.path.dirname(self.config_path), exist_ok=True)
+        with open(self.config_path, "w") as f:
+            yaml.dump(self.data, f, default_flow_style=False, allow_unicode=True, sort_keys=False)
+
+    def get(self, *keys, default=None):
+        val = self.data
+        for key in keys:
+            if isinstance(val, dict) and key in val:
+                val = val[key]
+            else:
+                return default
+        return val
+
+
+def _deep_copy(d):
+    if isinstance(d, dict):
+        return {k: _deep_copy(v) for k, v in d.items()}
+    if isinstance(d, list):
+        return [_deep_copy(v) for v in d]
+    return d
+
+
+def _deep_merge(base, override):
+    for key, value in override.items():
+        if key in base and isinstance(base[key], dict) and isinstance(value, dict):
+            _deep_merge(base[key], value)
+        else:
+            base[key] = value

bin/backend/models.py

@@ -0,0 +1,83 @@
+import datetime
+from sqlalchemy import create_engine, Column, Integer, String, Boolean, DateTime, Text
+from sqlalchemy.orm import declarative_base, sessionmaker
+
+Base = declarative_base()
+
+
+class Job(Base):
+    __tablename__ = "jobs"
+
+    id = Column(Integer, primary_key=True, autoincrement=True)
+    name = Column(String(200), nullable=False)
+    stream_url = Column(String(1000), nullable=False)
+    stream_type = Column(String(20), default="auto")
+    output_format = Column(String(10), default="")
+    max_duration = Column(Integer, nullable=True)
+    extra_ffmpeg_args = Column(String(500), default="")
+    segment_duration = Column(Integer, nullable=True)
+
+    schedule_enabled = Column(Boolean, default=False)
+    schedule_once = Column(Boolean, default=False)
+    schedule_date = Column(String(10), default="")
+    schedule_days = Column(String(50), default="*")
+    schedule_start = Column(String(5), default="")
+    schedule_stop = Column(String(5), default="")
+
+    plik_enabled = Column(Boolean, default=False)
+    delete_after_upload = Column(Boolean, default=False)
+
+    ntfy_enabled = Column(Boolean, default=False)
+
+    metadata_monitor_enabled = Column(Boolean, default=False)
+    metadata_pattern = Column(String(200), default="")
+    metadata_grace_period = Column(Integer, default=300)
+    metadata_poll_interval = Column(Integer, default=30)
+
+    created_at = Column(DateTime, default=datetime.datetime.utcnow)
+    updated_at = Column(DateTime, default=datetime.datetime.utcnow, onupdate=datetime.datetime.utcnow)
+
+
+class Recording(Base):
+    __tablename__ = "recordings"
+
+    id = Column(Integer, primary_key=True, autoincrement=True)
+    job_id = Column(Integer, nullable=True)
+    job_name = Column(String(200), nullable=False)
+    stream_url = Column(String(1000), nullable=False)
+    stream_type = Column(String(20), default="")
+    output_dir = Column(String(500), default="")
+    output_file = Column(String(500), default="")
+    pid = Column(Integer, nullable=True)
+    status = Column(String(20), default="running")  # running, stopped, completed, error
+    started_at = Column(DateTime, default=datetime.datetime.utcnow)
+    stopped_at = Column(DateTime, nullable=True)
+    scheduled_stop = Column(DateTime, nullable=True)
+    segment_mode = Column(Boolean, default=False)
+    is_scheduled = Column(Boolean, default=False)
+    error_message = Column(Text, nullable=True)
+
+
+class LogEntry(Base):
+    __tablename__ = "logs"
+
+    id = Column(Integer, primary_key=True, autoincrement=True)
+    timestamp = Column(DateTime, default=datetime.datetime.utcnow)
+    level = Column(String(10), default="INFO")
+    message = Column(Text, nullable=False)
+    job_name = Column(String(200), nullable=True)
+
+
+_engine = None
+_SessionLocal = None
+
+
+def init_db(db_path: str):
+    global _engine, _SessionLocal
+    _engine = create_engine(f"sqlite:///{db_path}", connect_args={"check_same_thread": False})
+    Base.metadata.create_all(_engine)
+    _SessionLocal = sessionmaker(bind=_engine)
+
+
+def get_session():
+    return _SessionLocal()

bin/backend/plugins/euer_radio.py

@@ -0,0 +1,138 @@
+"""
+Euer-Radio Metadata-Monitor Plugin.
+
+Polls RTMP stream metadata via ffprobe to detect whether a specific show
+is currently live. When the show ends (pattern disappears from metadata),
+the recording is stopped after a configurable grace period.
+"""
+
+import json
+import subprocess
+import threading
+import time
+from typing import Optional, Callable
+
+
+def probe_stream_title(stream_url: str, timeout: int = 10) -> Optional[str]:
+    """Query the current title tag from a live stream via ffprobe."""
+    cmd = [
+        "ffprobe", "-v", "quiet",
+        "-print_format", "json",
+        "-show_entries", "format_tags",
+        stream_url,
+    ]
+    try:
+        result = subprocess.run(
+            cmd,
+            capture_output=True,
+            text=True,
+            timeout=timeout,
+            stdin=subprocess.DEVNULL,
+        )
+        if result.returncode != 0:
+            return None
+        data = json.loads(result.stdout)
+        return data.get("format", {}).get("tags", {}).get("title")
+    except (subprocess.TimeoutExpired, json.JSONDecodeError, Exception):
+        return None
+
+
+def title_matches_pattern(title: str, pattern: str) -> bool:
+    """
+    Check whether a stream title matches the configured show pattern.
+
+    TODO: Finalize once we've seen real live-show metadata from euer.tv.
+    Current assumption: the title contains the pattern string as a prefix
+    or substring, e.g. "Donnerwolke live - Artist - Song".
+    """
+    if not title or not pattern:
+        return False
+    return pattern.lower() in title.lower()
+
+
+class MetadataMonitor:
+    """Monitors a stream's metadata and stops the recording when a show ends."""
+
+    def __init__(self, log_fn: Callable, stop_fn: Callable):
+        self._log = log_fn
+        self._stop_recording = stop_fn
+        self._monitors: dict[int, threading.Thread] = {}
+        self._cancel_events: dict[int, threading.Event] = {}
+
+    def start_monitoring(
+        self,
+        rec_id: int,
+        stream_url: str,
+        pattern: str,
+        grace_period: int = 300,
+        poll_interval: int = 30,
+        job_name: str = "",
+    ):
+        if rec_id in self._monitors:
+            return
+
+        cancel = threading.Event()
+        self._cancel_events[rec_id] = cancel
+
+        thread = threading.Thread(
+            target=self._monitor_loop,
+            args=(rec_id, stream_url, pattern, grace_period, poll_interval, cancel, job_name),
+            daemon=True,
+        )
+        self._monitors[rec_id] = thread
+        thread.start()
+        self._log("INFO", f"Metadata-Monitor gestartet für '{job_name}' (Pattern: '{pattern}')", job_name)
+
+    def stop_monitoring(self, rec_id: int):
+        cancel = self._cancel_events.pop(rec_id, None)
+        if cancel:
+            cancel.set()
+        self._monitors.pop(rec_id, None)
+
+    def is_monitoring(self, rec_id: int) -> bool:
+        return rec_id in self._monitors
+
+    def _monitor_loop(
+        self,
+        rec_id: int,
+        stream_url: str,
+        pattern: str,
+        grace_period: int,
+        poll_interval: int,
+        cancel: threading.Event,
+        job_name: str,
+    ):
+        grace_start = None
+
+        while not cancel.is_set():
+            title = probe_stream_title(stream_url)
+
+            if title is None:
+                if cancel.wait(timeout=poll_interval):
+                    break
+                continue
+
+            if title_matches_pattern(title, pattern):
+                if grace_start is not None:
+                    self._log("INFO",
+                              f"Show '{pattern}' wieder erkannt, Karenz abgebrochen [{job_name}]",
+                              job_name)
+                    grace_start = None
+            else:
+                if grace_start is None:
+                    grace_start = time.time()
+                    self._log("INFO",
+                              f"Show '{pattern}' nicht mehr erkannt, Karenzzeit läuft ({grace_period}s) [{job_name}]",
+                              job_name)
+                elif time.time() - grace_start >= grace_period:
+                    self._log("INFO",
+                              f"Karenzzeit abgelaufen, Aufnahme wird gestoppt [{job_name}]",
+                              job_name)
+                    self._stop_recording(rec_id)
+                    break
+
+            if cancel.wait(timeout=poll_interval):
+                break
+
+        self._monitors.pop(rec_id, None)
+        self._cancel_events.pop(rec_id, None)

bin/backend/plugins/ntfy.py

@@ -0,0 +1,65 @@
+"""NTFY push notification plugin."""
+
+import subprocess
+import threading
+
+EVENT_TAGS = {
+    "start": "red_circle",
+    "stop": "white_check_mark",
+    "error": "rotating_light",
+    "upload": "outbox_tray",
+}
+
+
+def send_notification(
+    url: str,
+    title: str,
+    message: str,
+    token: str = "",
+    priority: str = "default",
+    tags: str = "",
+):
+    if not url:
+        return
+
+    cmd = [
+        "curl", "-s", "-o", "/dev/null", "--max-time", "10",
+        "-H", f"Title: {title}",
+        "-H", f"Priority: {priority}",
+    ]
+    if tags:
+        cmd += ["-H", f"Tags: {tags}"]
+    cmd += ["-d", message]
+    if token:
+        cmd += ["-H", f"Authorization: Bearer {token}"]
+    cmd.append(url)
+
+    def _send():
+        try:
+            subprocess.run(cmd, capture_output=True, timeout=15, stdin=subprocess.DEVNULL)
+        except Exception:
+            pass
+
+    threading.Thread(target=_send, daemon=True).start()
+
+
+class NtfyNotifier:
+    def __init__(self, config):
+        self.config = config
+
+    def _should_notify(self, event: str, job_ntfy_enabled: bool) -> bool:
+        if not job_ntfy_enabled:
+            return False
+        url = self.config.get("ntfy", "url", default="")
+        if not url:
+            return False
+        events = self.config.get("ntfy", "events", default="error")
+        return event in [e.strip() for e in events.split(",")]
+
+    def notify(self, event: str, title: str, message: str, job_ntfy_enabled: bool = False, priority: str = "default"):
+        if not self._should_notify(event, job_ntfy_enabled):
+            return
+        url = self.config.get("ntfy", "url", default="")
+        token = self.config.get("ntfy", "token", default="")
+        tag = EVENT_TAGS.get(event, "")
+        send_notification(url, title, message, token=token, priority=priority, tags=tag)

bin/backend/plugins/plik.py

@@ -0,0 +1,168 @@
+"""Plik file upload plugin."""
+
+import json
+import os
+import re
+import subprocess
+import threading
+from typing import Optional, Callable
+
+
+def _parse_ttl_seconds(ttl: str) -> Optional[int]:
+    if not ttl:
+        return None
+    m = re.match(r"^(\d+)d$", ttl)
+    if m:
+        return int(m.group(1)) * 86400
+    m = re.match(r"^(\d+)h$", ttl)
+    if m:
+        return int(m.group(1)) * 3600
+    m = re.match(r"^(\d+)$", ttl)
+    if m:
+        return int(m.group(1))
+    return None
+
+
+def upload_file(
+    file_path: str,
+    plik_url: str,
+    api_key: str = "",
+    ttl: str = "30d",
+) -> Optional[dict]:
+    """Upload a file to Plik. Returns dict with 'browser_url' and 'download_url', or None on failure."""
+    if not plik_url or not os.path.isfile(file_path):
+        return None
+    if os.path.getsize(file_path) == 0:
+        return None
+
+    plik_base = plik_url.rstrip("/")
+    filename = os.path.basename(file_path)
+
+    cmd = ["curl", "-sS", "--max-time", "7200"]
+    if api_key:
+        cmd += ["-H", f"X-PlikToken: {api_key}"]
+
+    cmd += ["-F", f"file=@{file_path};filename={filename}"]
+
+    ttl_secs = _parse_ttl_seconds(ttl)
+    if ttl_secs is not None:
+        cmd += ["-F", f"ttl={ttl_secs}"]
+
+    cmd.append(plik_base)
+
+    try:
+        result = subprocess.run(cmd, capture_output=True, text=True, timeout=7200, stdin=subprocess.DEVNULL)
+        if result.returncode != 0:
+            return None
+
+        try:
+            data = json.loads(result.stdout)
+            upload_id = data.get("id", "")
+            if not upload_id:
+                return None
+
+            browser_url = f"{plik_base}/#/?id={upload_id}"
+
+            download_url = None
+            files = data.get("files", [])
+            if files:
+                file_id = files[0].get("id", "")
+                file_name = files[0].get("fileName", filename)
+                if file_id:
+                    download_url = f"{plik_base}/file/{upload_id}/{file_id}/{file_name}"
+
+            return {"browser_url": browser_url, "download_url": download_url}
+        except (json.JSONDecodeError, KeyError):
+            urls = re.findall(r'https?://\S+', result.stdout)
+            if urls:
+                download_url = urls[-1]
+                prefix = f"{plik_base}/file/"
+                if download_url.startswith(prefix):
+                    rest = download_url[len(prefix):]
+                    uid = rest.split("/")[0]
+                    if uid:
+                        return {"browser_url": f"{plik_base}/#/?id={uid}", "download_url": download_url}
+                return {"browser_url": download_url, "download_url": download_url}
+            return None
+    except Exception:
+        return None
+
+
+class PlikUploader:
+    def __init__(self, config, log_fn: Callable, notify_fn: Callable = None):
+        self.config = config
+        self._log = log_fn
+        self._notify = notify_fn
+
+    def upload_recording(self, output_dir: str, output_file: str, job_name: str,
+                         segment_mode: bool, delete_after: bool, ntfy_enabled: bool):
+        plik_url = self.config.get("plik", "url", default="")
+        if not plik_url:
+            self._log("WARN", f"Plik-Upload übersprungen: Keine URL konfiguriert [{job_name}]", job_name)
+            return
+
+        api_key = self.config.get("plik", "api_key", default="")
+        ttl = self.config.get("plik", "ttl", default="30d")
+
+        def _do_upload():
+            try:
+                if segment_mode:
+                    self._upload_segments(output_dir, job_name, plik_url, api_key, ttl, delete_after, ntfy_enabled)
+                else:
+                    self._upload_single(output_file, job_name, plik_url, api_key, ttl, delete_after, ntfy_enabled)
+            except Exception as e:
+                self._log("ERROR", f"Plik-Upload Fehler: {e} [{job_name}]", job_name)
+
+        threading.Thread(target=_do_upload, daemon=True).start()
+
+    def _upload_single(self, file_path, job_name, plik_url, api_key, ttl, delete_after, ntfy_enabled):
+        if not os.path.isfile(file_path) or os.path.getsize(file_path) == 0:
+            self._log("WARN", f"Plik-Upload übersprungen: Datei fehlt oder leer [{job_name}]", job_name)
+            return
+
+        self._log("INFO", f"Plik-Upload: {os.path.basename(file_path)} [{job_name}]", job_name)
+        result = upload_file(file_path, plik_url, api_key, ttl)
+
+        if result:
+            self._log("INFO", f"Plik-Upload abgeschlossen: {result['browser_url']} [{job_name}]", job_name)
+            if result.get("download_url"):
+                self._log("INFO", f"Plik-Download: {result['download_url']} [{job_name}]", job_name)
+            if self._notify:
+                self._notify("upload", "Plik-Upload", f"{job_name}: {result['browser_url']}", ntfy_enabled)
+            if delete_after:
+                os.remove(file_path)
+                self._log("INFO", f"Lokale Datei gelöscht: {file_path} [{job_name}]", job_name)
+        else:
+            self._log("ERROR", f"Plik-Upload fehlgeschlagen [{job_name}]", job_name)
+
+    def _upload_segments(self, output_dir, job_name, plik_url, api_key, ttl, delete_after, ntfy_enabled):
+        if not os.path.isdir(output_dir):
+            return
+
+        files = sorted(f for f in os.listdir(output_dir) if not f.startswith(".") and os.path.isfile(os.path.join(output_dir, f)))
+        if not files:
+            return
+
+        uploaded = 0
+        last_browser_url = None
+        for fname in files:
+            fpath = os.path.join(output_dir, fname)
+            if os.path.getsize(fpath) == 0:
+                continue
+            self._log("INFO", f"Plik Segment-Upload: {fname} [{job_name}]", job_name)
+            result = upload_file(fpath, plik_url, api_key, ttl)
+            if result:
+                last_browser_url = result["browser_url"]
+                self._log("INFO", f"Plik Segment hochgeladen: {last_browser_url} [{job_name}]", job_name)
+                uploaded += 1
+                if delete_after:
+                    os.remove(fpath)
+                    self._log("INFO", f"Segment gelöscht: {fname} [{job_name}]", job_name)
+            else:
+                self._log("ERROR", f"Plik Segment-Upload fehlgeschlagen: {fname} [{job_name}]", job_name)
+
+        if uploaded > 0 and self._notify:
+            msg = f"{job_name}: {uploaded} Segment(e) hochgeladen"
+            if last_browser_url:
+                msg += f"\n{last_browser_url}"
+            self._notify("upload", "Plik-Upload", msg, ntfy_enabled)

bin/backend/recorder.py

@@ -0,0 +1,501 @@
+import asyncio
+import datetime
+import os
+import re
+import signal
+import subprocess
+import threading
+import time
+from pathlib import Path
+from typing import Optional
+
+from models import Recording, get_session, LogEntry
+
+
+def detect_stream_type(url: str) -> str:
+    url_lower = url.lower().split("?")[0]
+    if url_lower.startswith(("rtmp://", "rtmps://")):
+        return "rtmp"
+    if url_lower.endswith(".m3u8"):
+        return "hls"
+    if url_lower.endswith(".mp3"):
+        return "mp3"
+    if url_lower.endswith(".aac"):
+        return "aac"
+    if url_lower.endswith(".ogg"):
+        return "ogg"
+    if url_lower.endswith(".flac"):
+        return "flac"
+    return "http"
+
+
+def default_extension(stream_type: str) -> str:
+    return {"mp3": "mp3", "aac": "aac", "ogg": "ogg", "flac": "flac"}.get(stream_type, "mp4")
+
+
+def safe_name(name: str) -> str:
+    return re.sub(r"[^A-Za-z0-9_-]", "_", name).strip("_")
+
+
+def format_elapsed(seconds: int) -> str:
+    d, rem = divmod(seconds, 86400)
+    h, rem = divmod(rem, 3600)
+    m, s = divmod(rem, 60)
+    if d > 0:
+        return f"{d}d {h}h {m}m"
+    if h > 0:
+        return f"{h}h {m}m {s}s"
+    return f"{m}m {s}s"
+
+
+def format_size(size_bytes: int) -> str:
+    if size_bytes >= 1073741824:
+        return f"{size_bytes / 1073741824:.1f} GB"
+    if size_bytes >= 1048576:
+        return f"{size_bytes / 1048576:.1f} MB"
+    if size_bytes >= 1024:
+        return f"{size_bytes / 1024:.1f} KB"
+    return f"{size_bytes} B"
+
+
+class RecordingManager:
+    def __init__(self, config):
+        self.config = config
+        self._processes: dict[int, subprocess.Popen] = {}
+        self._stop_events: dict[int, threading.Event] = {}
+        self._threads: dict[int, threading.Thread] = {}
+        self._metadata_monitor = None
+        self._ntfy = None
+        self._plik = None
+        self._rec_options: dict[int, dict] = {}
+
+    def _get_metadata_monitor(self):
+        if self._metadata_monitor is None:
+            from plugins.euer_radio import MetadataMonitor
+            self._metadata_monitor = MetadataMonitor(
+                log_fn=self._log,
+                stop_fn=self.stop_recording,
+            )
+        return self._metadata_monitor
+
+    def _get_ntfy(self):
+        if self._ntfy is None:
+            from plugins.ntfy import NtfyNotifier
+            self._ntfy = NtfyNotifier(self.config)
+        return self._ntfy
+
+    def _get_plik(self):
+        if self._plik is None:
+            from plugins.plik import PlikUploader
+            self._plik = PlikUploader(
+                self.config,
+                log_fn=self._log,
+                notify_fn=lambda event, title, msg, ntfy_en: self._get_ntfy().notify(event, title, msg, ntfy_en),
+            )
+        return self._plik
+
+    def _notify(self, event: str, title: str, message: str, rec_id: int = None, priority: str = "default"):
+        ntfy_enabled = False
+        if rec_id and rec_id in self._rec_options:
+            ntfy_enabled = self._rec_options[rec_id].get("ntfy_enabled", False)
+        self._get_ntfy().notify(event, title, message, job_ntfy_enabled=ntfy_enabled, priority=priority)
+
+    def _log(self, level: str, message: str, job_name: str = None):
+        session = get_session()
+        try:
+            entry = LogEntry(level=level, message=message, job_name=job_name)
+            session.add(entry)
+            session.commit()
+        finally:
+            session.close()
+
+    def start_recording(
+        self,
+        job_id: Optional[int],
+        name: str,
+        stream_url: str,
+        stream_type: str = "auto",
+        output_format: str = "",
+        max_duration: Optional[int] = None,
+        extra_ffmpeg_args: str = "",
+        segment_duration: Optional[int] = None,
+        is_scheduled: bool = False,
+        metadata_monitor_enabled: bool = False,
+        metadata_pattern: str = "",
+        metadata_grace_period: int = 300,
+        metadata_poll_interval: int = 30,
+        ntfy_enabled: bool = False,
+        plik_enabled: bool = False,
+        delete_after_upload: bool = False,
+    ) -> dict:
+        if stream_type == "auto":
+            stream_type = detect_stream_type(stream_url)
+
+        ext = output_format if output_format and output_format.lower() != "auto" else default_extension(stream_type)
+        sname = safe_name(name)
+        timestamp = datetime.datetime.now().strftime("%Y%m%d_%H%M%S")
+        download_path = self.config.get("recording", "download_path", default="/app/data/recordings")
+        output_dir = os.path.join(download_path, sname)
+        os.makedirs(output_dir, exist_ok=True)
+
+        segment_mode = segment_duration is not None and segment_duration > 0
+
+        if segment_mode:
+            output_file = os.path.join(output_dir, f"{sname}_{timestamp}_seg%03d.{ext}")
+        else:
+            output_file = os.path.join(output_dir, f"{sname}_{timestamp}.{ext}")
+
+        scheduled_stop = None
+        if max_duration and max_duration > 0:
+            scheduled_stop = datetime.datetime.utcnow() + datetime.timedelta(seconds=max_duration)
+
+        session = get_session()
+        try:
+            recording = Recording(
+                job_id=job_id,
+                job_name=name,
+                stream_url=stream_url,
+                stream_type=stream_type,
+                output_dir=output_dir,
+                output_file=output_file,
+                status="starting",
+                scheduled_stop=scheduled_stop,
+                segment_mode=segment_mode,
+                is_scheduled=is_scheduled,
+            )
+            session.add(recording)
+            session.commit()
+            rec_id = recording.id
+        finally:
+            session.close()
+
+        stop_event = threading.Event()
+        self._stop_events[rec_id] = stop_event
+
+        thread = threading.Thread(
+            target=self._recording_loop,
+            args=(rec_id, stream_url, stream_type, output_file, output_dir,
+                  extra_ffmpeg_args, segment_mode, segment_duration, stop_event, name),
+            daemon=True,
+        )
+        self._threads[rec_id] = thread
+        thread.start()
+
+        self._rec_options[rec_id] = {
+            "ntfy_enabled": ntfy_enabled,
+            "plik_enabled": plik_enabled,
+            "delete_after_upload": delete_after_upload,
+            "segment_mode": segment_mode,
+        }
+
+        self._log("INFO", f"Aufnahme gestartet: {name} [{stream_type}]", name)
+        self._notify("start", "Aufnahme gestartet", f"{name} [{stream_type}]", rec_id)
+
+        if metadata_monitor_enabled and metadata_pattern:
+            self._get_metadata_monitor().start_monitoring(
+                rec_id=rec_id,
+                stream_url=stream_url,
+                pattern=metadata_pattern,
+                grace_period=metadata_grace_period,
+                poll_interval=metadata_poll_interval,
+                job_name=name,
+            )
+
+        return {"recording_id": rec_id, "output_dir": output_dir, "output_file": output_file}
+
+    def _recording_loop(
+        self,
+        rec_id: int,
+        stream_url: str,
+        stream_type: str,
+        output_file: str,
+        output_dir: str,
+        extra_ffmpeg_args: str,
+        segment_mode: bool,
+        segment_duration: Optional[int],
+        stop_event: threading.Event,
+        job_name: str,
+    ):
+        max_retries = self.config.get("recording", "max_retries", default=5)
+        retry_delay = self.config.get("recording", "retry_delay", default=5)
+        retry = 0
+
+        while not stop_event.is_set():
+            cmd = self._build_ffmpeg_cmd(
+                stream_url, stream_type, output_file,
+                extra_ffmpeg_args, segment_mode, segment_duration,
+            )
+
+            try:
+                log_path = os.path.join(output_dir, f".ffmpeg_{rec_id}.log")
+                log_fh = open(log_path, "a")
+                process = subprocess.Popen(
+                    cmd,
+                    stdout=log_fh,
+                    stderr=log_fh,
+                    stdin=subprocess.DEVNULL,
+                )
+            except Exception as e:
+                self._update_recording(rec_id, status="error", error_message=str(e))
+                self._log("ERROR", f"ffmpeg konnte nicht gestartet werden: {e}", job_name)
+                self._notify("error", "Aufnahme fehlgeschlagen", f"{job_name}: {e}", rec_id, priority="high")
+                return
+
+            self._processes[rec_id] = process
+            self._update_recording(rec_id, status="running", pid=process.pid)
+
+            process.wait()
+            retcode = process.returncode
+            log_fh.close()
+            self._processes.pop(rec_id, None)
+
+            if stop_event.is_set() or retcode == 0:
+                status = "completed" if retcode == 0 else "stopped"
+                self._update_recording(rec_id, status=status)
+                self._log("INFO", f"Aufnahme {status}: {job_name}", job_name)
+                self._notify("stop", f"Aufnahme {status}", job_name, rec_id)
+                self._handle_post_recording(rec_id, output_dir, output_file, job_name, segment_mode)
+                return
+
+            if retcode in (-2, 130, 143, 255):
+                self._update_recording(rec_id, status="stopped")
+                self._log("INFO", f"Aufnahme gestoppt: {job_name}", job_name)
+                self._notify("stop", "Aufnahme gestoppt", job_name, rec_id)
+                self._handle_post_recording(rec_id, output_dir, output_file, job_name, segment_mode)
+                return
+
+            retry += 1
+            if max_retries > 0 and retry >= max_retries:
+                self._update_recording(rec_id, status="error",
+                                       error_message=f"Max retries ({max_retries}) reached, exit code {retcode}")
+                self._log("ERROR", f"Max. Versuche ({max_retries}) erreicht: {job_name}", job_name)
+                self._notify("error", "Aufnahme fehlgeschlagen", f"{job_name}: Max. Versuche erreicht", rec_id, priority="high")
+                return
+
+            self._log("WARN", f"Verbindung verloren (Code: {retcode}), Reconnect in {retry_delay}s... [{job_name}]", job_name)
+            self._notify("error", "Stream abgerissen", f"{job_name}: Reconnect in {retry_delay}s...", rec_id)
+            if stop_event.wait(timeout=retry_delay):
+                self._update_recording(rec_id, status="stopped")
+                return
+
+            sname = safe_name(job_name)
+            ts = datetime.datetime.now().strftime("%Y%m%d_%H%M%S")
+            ext = Path(output_file).suffix.lstrip(".")
+            if segment_mode:
+                output_file = os.path.join(output_dir, f"{sname}_{ts}_seg%03d.{ext}")
+            else:
+                output_file = os.path.join(output_dir, f"{sname}_{ts}.{ext}")
+
+        self._update_recording(rec_id, status="stopped")
+
+    def _build_ffmpeg_cmd(self, url, stream_type, output_file, extra_args, segment_mode, segment_duration):
+        cmd = ["ffmpeg", "-nostdin", "-y", "-hide_banner", "-loglevel", "warning"]
+
+        if stream_type == "hls":
+            cmd += ["-i", url, "-c", "copy", "-bsf:a", "aac_adtstoasc"]
+        else:
+            cmd += ["-i", url, "-c", "copy"]
+
+        if segment_mode and segment_duration:
+            cmd += ["-f", "segment", "-segment_time", str(segment_duration), "-reset_timestamps", "1"]
+
+        if extra_args:
+            cmd += extra_args.split()
+
+        cmd.append(output_file)
+        return cmd
+
+    def stop_recording(self, rec_id: int) -> bool:
+        if self._metadata_monitor:
+            self._metadata_monitor.stop_monitoring(rec_id)
+
+        stop_event = self._stop_events.get(rec_id)
+        if stop_event:
+            stop_event.set()
+
+        process = self._processes.get(rec_id)
+        if process:
+            try:
+                process.send_signal(signal.SIGINT)
+                try:
+                    process.wait(timeout=10)
+                except subprocess.TimeoutExpired:
+                    process.kill()
+            except (ProcessLookupError, OSError):
+                pass
+
+        self._update_recording(rec_id, status="stopped")
+        self._log("INFO", f"Aufnahme gestoppt (ID: {rec_id})")
+        self._cleanup(rec_id)
+        return True
+
+    def extend_recording(self, rec_id: int, extra_seconds: int) -> bool:
+        session = get_session()
+        try:
+            rec = session.query(Recording).filter_by(id=rec_id).first()
+            if not rec or rec.status != "running":
+                return False
+
+            if rec.scheduled_stop:
+                rec.scheduled_stop = rec.scheduled_stop + datetime.timedelta(seconds=extra_seconds)
+            else:
+                rec.scheduled_stop = datetime.datetime.utcnow() + datetime.timedelta(seconds=extra_seconds)
+
+            session.commit()
+            self._log("INFO", f"Aufnahme verlängert um {format_elapsed(extra_seconds)}: {rec.job_name}", rec.job_name)
+            return True
+        finally:
+            session.close()
+
+    def remove_duration_limit(self, rec_id: int) -> bool:
+        session = get_session()
+        try:
+            rec = session.query(Recording).filter_by(id=rec_id).first()
+            if not rec or rec.status != "running":
+                return False
+            rec.scheduled_stop = None
+            session.commit()
+            self._log("INFO", f"Zeitlimit entfernt: {rec.job_name}", rec.job_name)
+            return True
+        finally:
+            session.close()
+
+    def get_status(self, rec_id: int) -> Optional[dict]:
+        session = get_session()
+        try:
+            rec = session.query(Recording).filter_by(id=rec_id).first()
+            if not rec:
+                return None
+            return self._recording_to_dict(rec)
+        finally:
+            session.close()
+
+    def get_all_active(self) -> list[dict]:
+        session = get_session()
+        try:
+            recs = session.query(Recording).filter(Recording.status.in_(["running", "starting"])).all()
+            return [self._recording_to_dict(r) for r in recs]
+        finally:
+            session.close()
+
+    def get_all_recordings(self, limit: int = 50) -> list[dict]:
+        session = get_session()
+        try:
+            recs = session.query(Recording).order_by(Recording.id.desc()).limit(limit).all()
+            return [self._recording_to_dict(r) for r in recs]
+        finally:
+            session.close()
+
+    def check_scheduled_stops(self):
+        session = get_session()
+        try:
+            now = datetime.datetime.utcnow()
+            recs = session.query(Recording).filter(
+                Recording.status == "running",
+                Recording.scheduled_stop.isnot(None),
+                Recording.scheduled_stop <= now,
+            ).all()
+            rec_ids = [r.id for r in recs]
+        finally:
+            session.close()
+
+        for rec_id in rec_ids:
+            self._log("INFO", f"Geplanter Stopp erreicht (ID: {rec_id})")
+            self.stop_recording(rec_id)
+
+    def cleanup_stale(self):
+        session = get_session()
+        try:
+            recs = session.query(Recording).filter(Recording.status.in_(["running", "starting"])).all()
+            for rec in recs:
+                if rec.pid:
+                    try:
+                        os.kill(rec.pid, 0)
+                    except (ProcessLookupError, OSError):
+                        rec.status = "error"
+                        rec.error_message = "Process vanished"
+                        rec.stopped_at = datetime.datetime.utcnow()
+            session.commit()
+        finally:
+            session.close()
+
+    def _recording_to_dict(self, rec: Recording) -> dict:
+        elapsed = 0
+        if rec.started_at:
+            end = rec.stopped_at or datetime.datetime.utcnow()
+            elapsed = int((end - rec.started_at).total_seconds())
+
+        total_size = 0
+        segment_count = 0
+        if rec.output_dir and os.path.isdir(rec.output_dir):
+            for f in os.listdir(rec.output_dir):
+                if f.startswith("."):
+                    continue
+                fp = os.path.join(rec.output_dir, f)
+                if os.path.isfile(fp):
+                    total_size += os.path.getsize(fp)
+                    segment_count += 1
+        elif rec.output_file and os.path.isfile(rec.output_file):
+            total_size = os.path.getsize(rec.output_file)
+
+        remaining = None
+        if rec.scheduled_stop and rec.status == "running":
+            diff = (rec.scheduled_stop - datetime.datetime.utcnow()).total_seconds()
+            remaining = max(0, int(diff))
+
+        return {
+            "id": rec.id,
+            "job_id": rec.job_id,
+            "job_name": rec.job_name,
+            "stream_url": rec.stream_url,
+            "stream_type": rec.stream_type,
+            "output_dir": rec.output_dir,
+            "output_file": rec.output_file,
+            "pid": rec.pid,
+            "status": rec.status,
+            "started_at": rec.started_at.isoformat() if rec.started_at else None,
+            "stopped_at": rec.stopped_at.isoformat() if rec.stopped_at else None,
+            "scheduled_stop": rec.scheduled_stop.isoformat() if rec.scheduled_stop else None,
+            "segment_mode": rec.segment_mode,
+            "is_scheduled": rec.is_scheduled,
+            "error_message": rec.error_message,
+            "elapsed": elapsed,
+            "elapsed_display": format_elapsed(elapsed),
+            "total_size": total_size,
+            "total_size_display": format_size(total_size),
+            "segment_count": segment_count if rec.segment_mode else None,
+            "remaining": remaining,
+            "remaining_display": format_elapsed(remaining) if remaining is not None else None,
+        }
+
+    def _update_recording(self, rec_id: int, **kwargs):
+        session = get_session()
+        try:
+            rec = session.query(Recording).filter_by(id=rec_id).first()
+            if rec:
+                for k, v in kwargs.items():
+                    setattr(rec, k, v)
+                if kwargs.get("status") in ("stopped", "completed", "error"):
+                    rec.stopped_at = datetime.datetime.utcnow()
+                session.commit()
+        finally:
+            session.close()
+
+    def _handle_post_recording(self, rec_id: int, output_dir: str, output_file: str,
+                               job_name: str, segment_mode: bool):
+        opts = self._rec_options.get(rec_id, {})
+        if opts.get("plik_enabled"):
+            self._get_plik().upload_recording(
+                output_dir=output_dir,
+                output_file=output_file,
+                job_name=job_name,
+                segment_mode=segment_mode,
+                delete_after=opts.get("delete_after_upload", False),
+                ntfy_enabled=opts.get("ntfy_enabled", False),
+            )
+
+    def _cleanup(self, rec_id: int):
+        self._processes.pop(rec_id, None)
+        self._stop_events.pop(rec_id, None)
+        self._threads.pop(rec_id, None)
+        self._rec_options.pop(rec_id, None)

bin/backend/requirements.txt

@@ -0,0 +1,8 @@
+fastapi==0.115.12
+uvicorn[standard]==0.34.2
+sqlalchemy==2.0.40
+pydantic==2.11.3
+apscheduler==3.11.0
+pyyaml==6.0.2
+python-multipart==0.0.20
+httpx==0.28.1

bin/backend/scheduler.py

@@ -0,0 +1,187 @@
+import datetime
+from apscheduler.schedulers.background import BackgroundScheduler
+from models import Job, Recording, get_session
+
+DAY_MAP = {
+    "mo": 1, "mon": 1, "montag": 1,
+    "di": 2, "tue": 2, "dienstag": 2,
+    "mi": 3, "wed": 3, "mittwoch": 3,
+    "do": 4, "thu": 4, "donnerstag": 4,
+    "fr": 5, "fri": 5, "freitag": 5,
+    "sa": 6, "sat": 6, "samstag": 6,
+    "so": 7, "sun": 7, "sonntag": 7,
+}
+
+
+def _day_to_num(day: str) -> int:
+    return DAY_MAP.get(day.lower().strip(), 0)
+
+
+def _matches_day(today_num: int, schedule: str) -> bool:
+    if schedule.strip() == "*":
+        return True
+    for entry in schedule.split(","):
+        entry = entry.strip()
+        if "-" in entry:
+            parts = entry.split("-", 1)
+            start = _day_to_num(parts[0])
+            end = _day_to_num(parts[1])
+            if start <= end:
+                if start <= today_num <= end:
+                    return True
+            else:
+                if today_num >= start or today_num <= end:
+                    return True
+        else:
+            if _day_to_num(entry) == today_num:
+                return True
+    return False
+
+
+def _time_to_minutes(t: str) -> int:
+    parts = t.split(":")
+    return int(parts[0]) * 60 + int(parts[1])
+
+
+def _in_time_window(now_time: str, start: str, stop: str) -> bool:
+    now_min = _time_to_minutes(now_time)
+    start_min = _time_to_minutes(start)
+    stop_min = _time_to_minutes(stop)
+    if start_min <= stop_min:
+        return start_min <= now_min < stop_min
+    else:
+        return now_min >= start_min or now_min < stop_min
+
+
+class StreamScheduler:
+    def __init__(self, recorder):
+        self.recorder = recorder
+        self._scheduler = BackgroundScheduler()
+        self._started_keys: set[str] = set()
+
+    def start(self):
+        self._scheduler.add_job(self._tick, "interval", seconds=30, id="scheduler_tick",
+                                replace_existing=True, max_instances=1)
+        self._scheduler.add_job(self.recorder.check_scheduled_stops, "interval", seconds=10,
+                                id="check_stops", replace_existing=True, max_instances=1)
+        self._scheduler.start()
+
+    def stop(self):
+        self._scheduler.shutdown(wait=False)
+
+    def _occurrence_key(self, job: Job) -> str:
+        today = datetime.date.today().isoformat()
+        if job.schedule_once:
+            today = job.schedule_date or today
+        return f"{job.id}|{job.schedule_once}|{today}|{job.schedule_start}"
+
+    def _is_job_recording(self, job_id: int) -> bool:
+        session = get_session()
+        try:
+            return session.query(Recording).filter(
+                Recording.job_id == job_id,
+                Recording.status.in_(["running", "starting"]),
+            ).first() is not None
+        finally:
+            session.close()
+
+    def _get_active_recording_for_job(self, job_id: int):
+        session = get_session()
+        try:
+            rec = session.query(Recording).filter(
+                Recording.job_id == job_id,
+                Recording.status.in_(["running", "starting"]),
+            ).first()
+            return rec.id if rec else None
+        finally:
+            session.close()
+
+    def _tick(self):
+        now = datetime.datetime.now()
+        today_num = now.isoweekday()
+        today_date = now.strftime("%Y-%m-%d")
+        now_time = now.strftime("%H:%M")
+
+        session = get_session()
+        try:
+            jobs = session.query(Job).filter_by(schedule_enabled=True).all()
+            job_list = []
+            for j in jobs:
+                job_list.append({
+                    "id": j.id, "name": j.name, "stream_url": j.stream_url,
+                    "stream_type": j.stream_type, "output_format": j.output_format,
+                    "max_duration": j.max_duration, "extra_ffmpeg_args": j.extra_ffmpeg_args,
+                    "segment_duration": j.segment_duration,
+                    "schedule_once": j.schedule_once, "schedule_date": j.schedule_date,
+                    "schedule_days": j.schedule_days, "schedule_start": j.schedule_start,
+                    "schedule_stop": j.schedule_stop,
+                    "metadata_monitor_enabled": j.metadata_monitor_enabled,
+                    "metadata_pattern": j.metadata_pattern,
+                    "metadata_grace_period": j.metadata_grace_period,
+                    "metadata_poll_interval": j.metadata_poll_interval,
+                    "ntfy_enabled": j.ntfy_enabled,
+                    "plik_enabled": j.plik_enabled,
+                    "delete_after_upload": j.delete_after_upload,
+                })
+        finally:
+            session.close()
+
+        for job in job_list:
+            if not job["schedule_start"]:
+                continue
+
+            is_today = False
+            if job["schedule_once"]:
+                if job["schedule_date"] == today_date:
+                    is_today = True
+            else:
+                if _matches_day(today_num, job["schedule_days"] or "*"):
+                    is_today = True
+
+            in_window = False
+            if is_today:
+                if job["schedule_stop"]:
+                    in_window = _in_time_window(now_time, job["schedule_start"], job["schedule_stop"])
+                else:
+                    now_min = _time_to_minutes(now_time)
+                    start_min = _time_to_minutes(job["schedule_start"])
+                    in_window = now_min >= start_min
+
+            occ_key = f"{job['id']}|{job['schedule_once']}|{today_date if not job['schedule_once'] else job['schedule_date']}|{job['schedule_start']}"
+
+            if in_window:
+                if occ_key not in self._started_keys and not self._is_job_recording(job["id"]):
+                    duration = job["max_duration"]
+                    if not duration and job["schedule_stop"]:
+                        start_min = _time_to_minutes(job["schedule_start"])
+                        stop_min = _time_to_minutes(job["schedule_stop"])
+                        if stop_min > start_min:
+                            duration = (stop_min - start_min) * 60
+                        else:
+                            duration = (1440 - start_min + stop_min) * 60
+
+                    self.recorder.start_recording(
+                        job_id=job["id"],
+                        name=job["name"],
+                        stream_url=job["stream_url"],
+                        stream_type=job["stream_type"],
+                        output_format=job["output_format"] or "",
+                        max_duration=duration,
+                        extra_ffmpeg_args=job["extra_ffmpeg_args"] or "",
+                        segment_duration=job["segment_duration"],
+                        is_scheduled=True,
+                        metadata_monitor_enabled=job.get("metadata_monitor_enabled", False),
+                        metadata_pattern=job.get("metadata_pattern", ""),
+                        metadata_grace_period=job.get("metadata_grace_period") or 300,
+                        metadata_poll_interval=job.get("metadata_poll_interval") or 30,
+                        ntfy_enabled=job.get("ntfy_enabled", False),
+                        plik_enabled=job.get("plik_enabled", False),
+                        delete_after_upload=job.get("delete_after_upload", False),
+                    )
+                    self._started_keys.add(occ_key)
+            else:
+                if job["schedule_stop"] and self._is_job_recording(job["id"]):
+                    rec_id = self._get_active_recording_for_job(job["id"])
+                    if rec_id:
+                        self.recorder.stop_recording(rec_id)
+                        self._started_keys.discard(occ_key)

config/config.yml.dist

@@ -0,0 +1,42 @@
+###############################################
+# stream-recorder v3 - Konfiguration
+#
+# Diese Datei ist eine Vorlage.
+# Die aktive Konfiguration liegt in ./data/config.yml
+# und wird beim ersten Start automatisch erstellt.
+###############################################
+
+recording:
+  # Basis-Verzeichnis für Aufnahmen
+  download_path: /app/data/recordings
+
+  # Wiederholungsversuche bei Verbindungsabbruch (0 = unbegrenzt)
+  max_retries: 5
+
+  # Wartezeit zwischen Wiederholungsversuchen (Sekunden)
+  retry_delay: 5
+
+server:
+  host: "0.0.0.0"
+  port: 8484
+
+# Optionale Basic-Auth
+auth:
+  enabled: false
+  username: admin
+  password: stream-recorder
+
+# NTFY Push-Benachrichtigungen (optional)
+ntfy:
+  url: ""
+  token: ""
+  events: "error"
+
+# Plik-Upload (optional)
+plik:
+  url: ""
+  api_key: ""
+  ttl: "30d"
+
+logging:
+  level: INFO

docker-compose.yml

@@ -0,0 +1,33 @@
+services:
+  stream-recorder:
+    build: .
+    image: stream-recorder:latest
+    container_name: stream-recorder
+    restart: unless-stopped
+    ports:
+      - "${APP_PORT_BIND:-8484}:8484"
+    volumes:
+      - ./data:/app/data
+      # Optional: Recordings auf separatem Pfad (dann download_path in Einstellungen anpassen)
+      # - /mnt/storage/recordings:/recordings
+    networks:
+      stream-recorder_net:
+        ipv4_address: ${APP_IPV4_ADDRESS:-172.16.84.10}
+    environment:
+      TZ: ${TZ:-Europe/Berlin}
+    healthcheck:
+      test: ["CMD", "curl", "-f", "http://localhost:8484/api/health"]
+      interval: 30s
+      timeout: 5s
+      retries: 3
+      start_period: 10s
+
+networks:
+  stream-recorder_net:
+    name: ${DOCKER_NETWORK_NAME:-stream-recorder.dockernetwork.local}
+    driver: bridge
+    ipam:
+      config:
+        - subnet: ${DOCKER_NETWORK_SUBNET:-172.16.84.0/24}
+          gateway: ${DOCKER_NETWORK_GATEWAY:-172.16.84.1}
+          ip_range: ${DOCKER_NETWORK_IP_RANGE:-172.16.84.128/25}

docs/api.md

@@ -0,0 +1,65 @@
+# REST-API
+
+Alle Funktionen des Web-UI sind auch per REST-API verfügbar. Basis-URL: `http://<host>:8484`
+
+## Endpunkte
+
+### System
+
+| Methode | Pfad | Beschreibung |
+|---------|------|-------------|
+| `GET` | `/api/health` | Health Check (kein Auth nötig) |
+| `GET` | `/api/config` | Konfiguration lesen |
+| `PUT` | `/api/config` | Konfiguration speichern |
+
+### Jobs
+
+| Methode | Pfad | Beschreibung |
+|---------|------|-------------|
+| `GET` | `/api/jobs` | Alle Jobs auflisten |
+| `POST` | `/api/jobs` | Neuen Job erstellen |
+| `GET` | `/api/jobs/:id` | Job-Details abrufen |
+| `PUT` | `/api/jobs/:id` | Job aktualisieren |
+| `DELETE` | `/api/jobs/:id` | Job löschen |
+| `POST` | `/api/jobs/:id/record` | Job-Aufnahme starten |
+
+### Schnellaufnahme
+
+| Methode | Pfad | Beschreibung |
+|---------|------|-------------|
+| `POST` | `/api/quick-record` | Schnellaufnahme starten |
+
+```json
+{
+    "url": "rtmp://example.com/live/stream",
+    "name": "Meine Aufnahme",
+    "stream_type": "auto",
+    "max_duration": 3600
+}
+```
+
+### Aufnahmen
+
+| Methode | Pfad | Beschreibung |
+|---------|------|-------------|
+| `GET` | `/api/recordings` | Alle Aufnahmen (Parameter: `active_only`, `limit`) |
+| `GET` | `/api/recordings/:id` | Aufnahme-Status |
+| `POST` | `/api/recordings/:id/stop` | Aufnahme stoppen |
+| `POST` | `/api/recordings/:id/extend?minutes=30` | Aufnahme verlängern |
+| `POST` | `/api/recordings/:id/unlimited` | Zeitlimit entfernen |
+
+### Logs
+
+| Methode | Pfad | Beschreibung |
+|---------|------|-------------|
+| `GET` | `/api/logs` | Protokoll (Parameter: `limit`, `level`) |
+
+## Authentifizierung
+
+Bei aktiviertem Auth wird Basic Authentication verwendet:
+
+```bash
+curl -u admin:passwort http://localhost:8484/api/jobs
+```
+
+Der Health-Endpunkt (`/api/health`) ist immer ohne Auth erreichbar.

docs/aufnahmen.md

@@ -0,0 +1,57 @@
+# Aufnahmen
+
+## Aktive Aufnahmen
+
+Das Dashboard zeigt alle laufenden Aufnahmen mit:
+- Laufzeit
+- Dateigröße
+- Verbleibende Zeit (bei Zeitlimit)
+- Segment-Anzahl (bei Segment-Splitting)
+
+## Aufnahme verlängern
+
+Laufende Aufnahmen können verlängert werden:
+- **+15 / +30 / +60 / +120 Minuten** — Schnellauswahl
+- **Eigene Dauer** — Beliebige Minutenzahl eingeben
+- **Unbegrenzt** — Zeitlimit komplett entfernen
+
+## Aufnahme stoppen
+
+Über den **Stoppen**-Button im Dashboard. Die Aufnahme wird sauber beendet (SIGINT → ffmpeg finalisiert die Datei).
+
+## Reconnect
+
+Bei Verbindungsabbrüchen versucht der Recorder automatisch, die Verbindung wiederherzustellen. Konfigurierbar über:
+- `recording.max_retries` — Maximale Versuche (Standard: 5)
+- `recording.retry_delay` — Wartezeit zwischen Versuchen (Standard: 5 Sekunden)
+
+Bei jedem Reconnect wird eine neue Datei angelegt, um Datenverlust zu vermeiden.
+
+## Segment-Splitting
+
+Lange Aufnahmen können automatisch in Teile zerlegt werden. Die Segment-Dauer wird pro Job in Minuten konfiguriert.
+
+Beispiel: Bei 60 Minuten Segment-Dauer wird eine 3-stündige Aufnahme in 3 Dateien aufgeteilt:
+```
+Sendung_20260915_200000_seg000.mp4
+Sendung_20260915_200000_seg001.mp4
+Sendung_20260915_200000_seg002.mp4
+```
+
+## Ausgabeformate
+
+Das Format wird automatisch anhand des Stream-Typs gewählt:
+
+| Stream-Typ | Standard-Format |
+|------------|----------------|
+| RTMP, HLS, HTTP | mp4 |
+| MP3 | mp3 |
+| AAC | aac |
+| OGG | ogg |
+| FLAC | flac |
+
+Kann pro Job manuell überschrieben werden (z.B. `mkv`, `ts`).
+
+## Speicherort
+
+Aufnahmen werden unter `data/recordings/<job-name>/` gespeichert (im Container: `/app/data/recordings/`). Der Pfad ist über `recording.download_path` konfigurierbar.

docs/jobs.md

@@ -0,0 +1,58 @@
+# Jobs
+
+Jobs sind vorkonfigurierte Aufnahme-Vorlagen. Jeder Job definiert einen Stream, optional einen Zeitplan und Plugin-Einstellungen.
+
+## Job erstellen
+
+Im Web-UI unter **Jobs → + Neuer Job**.
+
+### Pflichtfelder
+
+| Feld | Beschreibung |
+|------|-------------|
+| **Name** | Eindeutiger Name (wird als Ordnername verwendet) |
+| **Stream-URL** | URL des Streams (RTMP, HLS, HTTP, etc.) |
+
+### Optionale Felder
+
+| Feld | Beschreibung |
+|------|-------------|
+| **Typ** | Automatisch erkannt aus URL. Manuell wählbar: RTMP, HLS, MP3, AAC, OGG, HTTP |
+| **Ausgabeformat** | Dateiendung. Leer = automatisch (mp4 für Video, mp3 für Audio) |
+| **Max. Dauer** | Zeitlimit in Minuten. Entfällt bei Zeitplan mit Endzeit |
+| **Segment-Dauer** | Aufnahme in Teile zerlegen (Minuten pro Segment) |
+| **Extra ffmpeg Argumente** | Zusätzliche ffmpeg-Parameter (z.B. `-b:a 192k`) |
+
+## Zeitplan
+
+Aktiviert automatische Aufnahmen zu festgelegten Zeiten.
+
+| Feld | Beschreibung |
+|------|-------------|
+| **Einmalig** | Nur einmal am angegebenen Datum ausführen |
+| **Datum** | Datum für einmalige Aufnahme |
+| **Tage** | Wochentage: `*` (täglich), `Mo-Fr`, `Sa,So`, `Mo,Mi,Fr` |
+| **Startzeit** | Aufnahme-Beginn (HH:MM) |
+| **Endzeit** | Aufnahme-Ende (HH:MM) — Dauer wird automatisch berechnet |
+
+Wenn Start- und Endzeit gesetzt sind, wird die Aufnahmedauer automatisch berechnet. Das Feld "Max. Dauer" wird dann nicht benötigt.
+
+### Tage-Syntax
+
+| Eingabe | Bedeutung |
+|---------|-----------|
+| `*` | Jeden Tag |
+| `Mo-Fr` | Montag bis Freitag |
+| `Sa,So` | Samstag und Sonntag |
+| `Mo,Mi,Fr` | Montag, Mittwoch, Freitag |
+| `Mo-Mi,Fr` | Montag bis Mittwoch und Freitag |
+
+Unterstützt: `Mo/Mon/Montag`, `Di/Tue/Dienstag`, `Mi/Wed/Mittwoch`, `Do/Thu/Donnerstag`, `Fr/Fri/Freitag`, `Sa/Sat/Samstag`, `So/Sun/Sonntag`
+
+## Manuelle Aufnahme
+
+Jobs können auch manuell über den **Aufnehmen**-Button gestartet werden, unabhängig vom Zeitplan.
+
+## Schnellaufnahme
+
+Für einmalige Aufnahmen ohne Job: **Dashboard → + Schnellaufnahme**. URL eingeben, optional Name und Dauer angeben, sofort starten.

docs/konfiguration.md

@@ -0,0 +1,65 @@
+# Konfiguration
+
+## config.yml
+
+Die Konfiguration liegt in `data/config.yml` und kann über das Web-UI unter **Einstellungen** bearbeitet werden.
+
+Beim ersten Start wird die Datei automatisch mit Standardwerten erstellt. Eine Vorlage mit Kommentaren findet sich in `config/config.yml.dist`.
+
+## Abschnitte
+
+### Aufnahme
+
+| Parameter | Standard | Beschreibung |
+|-----------|----------|-------------|
+| `recording.download_path` | `/app/data/recordings` | Speicherort für Aufnahmen |
+| `recording.max_retries` | `5` | Max. Reconnect-Versuche bei Verbindungsabbruch |
+| `recording.retry_delay` | `5` | Wartezeit (Sekunden) zwischen Reconnect-Versuchen |
+
+### Authentifizierung
+
+| Parameter | Standard | Beschreibung |
+|-----------|----------|-------------|
+| `auth.enabled` | `false` | Basic Auth aktivieren |
+| `auth.username` | `admin` | Benutzername |
+| `auth.password` | `stream-recorder` | Passwort |
+
+### NTFY
+
+| Parameter | Standard | Beschreibung |
+|-----------|----------|-------------|
+| `ntfy.url` | _(leer)_ | NTFY Topic-URL (z.B. `https://ntfy.sh/mein-topic`) |
+| `ntfy.token` | _(leer)_ | Bearer-Token für authentifizierte Topics |
+| `ntfy.events` | `error` | Kommaseparierte Events: `start`, `stop`, `error`, `upload` |
+
+### Plik
+
+| Parameter | Standard | Beschreibung |
+|-----------|----------|-------------|
+| `plik.url` | _(leer)_ | Plik-Server URL (z.B. `https://plik.example.com`) |
+| `plik.api_key` | _(leer)_ | API Key für authentifizierte Uploads |
+| `plik.ttl` | `30d` | Aufbewahrungsdauer der Uploads (`30d`, `12h`, etc.) |
+
+## Beispiel
+
+```yaml
+recording:
+  download_path: /app/data/recordings
+  max_retries: 5
+  retry_delay: 5
+
+auth:
+  enabled: true
+  username: admin
+  password: mein-passwort
+
+ntfy:
+  url: https://ntfy.sh/stream-recorder
+  token: ""
+  events: start,stop,error,upload
+
+plik:
+  url: https://plik.example.com
+  api_key: ""
+  ttl: 30d
+```

docs/plugins.md

@@ -0,0 +1,85 @@
+# Plugins
+
+Plugins erweitern den Recorder um zusätzliche Funktionen. Alle Plugins sind pro Job einzeln aktivierbar.
+
+## NTFY-Benachrichtigungen
+
+Push-Nachrichten per [ntfy.sh](https://ntfy.sh) bei verschiedenen Events.
+
+### Einrichtung
+
+1. **Einstellungen → NTFY URL**: Topic-URL eintragen (z.B. `https://ntfy.sh/mein-topic`)
+2. **Einstellungen → Events**: Kommaseparierte Liste der gewünschten Events
+3. **Pro Job**: Checkbox „Benachrichtigungen (NTFY)" aktivieren
+
+### Events
+
+| Event | Tag | Beschreibung |
+|-------|-----|-------------|
+| `start` | 🔴 | Aufnahme gestartet |
+| `stop` | ✅ | Aufnahme beendet |
+| `error` | 🚨 | Fehler oder Verbindungsabbruch |
+| `upload` | 📤 | Plik-Upload abgeschlossen (mit Link) |
+
+### Authentifizierung
+
+Für private Topics kann ein Bearer-Token in den Einstellungen hinterlegt werden.
+
+---
+
+## Plik-Upload
+
+Automatischer Upload von Aufnahmen auf einen [Plik](https://github.com/root-gg/plik)-Server nach Aufnahme-Ende.
+
+### Einrichtung
+
+1. **Einstellungen → Plik URL**: Server-URL eintragen
+2. **Einstellungen → TTL**: Aufbewahrungsdauer (Standard: `30d`)
+3. **Pro Job**: Checkbox „Plik-Upload" aktivieren
+
+### Optionen pro Job
+
+| Option | Beschreibung |
+|--------|-------------|
+| **Plik-Upload** | Upload nach Aufnahme-Ende aktivieren |
+| **Nach Upload löschen** | Lokale Datei nach erfolgreichem Upload entfernen |
+
+### Links
+
+Nach dem Upload werden zwei URLs generiert:
+- **Browser-URL**: `https://plik.example.com/#/?id=UPLOAD_ID` — Übersichtsseite mit In-Browser-Wiedergabe
+- **Download-URL**: Direkter Datei-Download
+
+Bei aktiviertem NTFY wird die Browser-URL in der Push-Nachricht mitgeschickt.
+
+### Segment-Upload
+
+Bei Segment-Splitting werden alle Segmente einzeln hochgeladen. Die NTFY-Nachricht enthält die Anzahl der hochgeladenen Segmente.
+
+---
+
+## Euer-Radio Metadata-Monitor
+
+Erkennt automatisch, ob eine bestimmte Show live sendet, indem der Stream-Titel via `ffprobe` abgefragt wird. Wenn das konfigurierte Pattern nicht mehr im Titel erscheint, wird die Aufnahme nach einer Karenzzeit gestoppt.
+
+### Einrichtung
+
+Pro Job im Web-UI:
+
+| Option | Standard | Beschreibung |
+|--------|----------|-------------|
+| **Show-Pattern** | _(leer)_ | Text, der im Stream-Titel vorkommen muss (z.B. Sendungsname) |
+| **Karenzzeit** | 5 Min. | Wie lange gewartet wird, nachdem das Pattern verschwunden ist |
+| **Poll-Intervall** | 30 Sek. | Wie oft der Stream-Titel abgefragt wird |
+
+### Funktionsweise
+
+1. Der Monitor fragt regelmäßig den Stream-Titel per `ffprobe` ab
+2. Solange das Pattern im Titel vorkommt, läuft die Aufnahme weiter
+3. Wenn das Pattern verschwindet, startet die Karenzzeit
+4. Taucht das Pattern innerhalb der Karenzzeit wieder auf, wird der Timer zurückgesetzt
+5. Läuft die Karenzzeit ab, wird die Aufnahme gestoppt
+
+### Anwendungsfall
+
+Ideal für Streams mit wechselnden Shows (z.B. Internet-Radio), bei denen die Aufnahmedauer nicht vorher feststeht. Der Stream-Titel zeigt an, welche Show gerade läuft, und der Monitor erkennt automatisch das Ende.