router: Bildgenerierung mit FLUX.2 [klein] 4B Base (GPU-Hotswap)

- POST /v1/images/generations (OpenAI-kompatibel, prompt/size/n/seed/quality)
- quality: standard=30 Steps (Default), high=50 Steps
- Größen: 1024x1024, 1536x1024, 1024x1536, 1920x1088, 1088x1920
- GPU-Hotswap: Qwen stoppen -> FLUX laden -> Bild -> FLUX entladen -> Qwen
  wiederherstellen (exakt vorheriges Profil)
- Zentrales GPU/Modell-Lock (Profilwechsel und Bild teilen sich das Lock)
- Chat-Requests warten während Bild-Job (kein 502), Timeout CHAT_WAIT_TIMEOUT
- Robuste Recovery: try/finally, Worker-Beendigung, VRAM-Check, Qwen-Readiness
- /status: image.phase, image.worker, image.model_loaded, qwen.available,
  qwen.active_chats
- GET /images, GET /images/<datei> (validiert, nur images/-Verzeichnis)
- image_worker.py: FLUX-Worker (eigener Prozess, JSON-Protokoll, bf16 +
  enable_model_cpu_offload)
- deploy: venv (torch/diffusers/transformers/accelerate), Modell-Download,
  Image-Dir, systemd-Unit mit Image-Umgebungsvariablen
- dev: Mock-Worker, fake-systemctl, Benchmarks (GPU-Resident, Offload, Steps,
  Quality-Compare), 32 lokale Tests
- README: Bildgenerierung, Hotswap, Recovery, Benchmarks (RTX 5080),
  Python-Pakete

Benchmarks (RTX 5080, 16 GB, CPU-Offload):
- 512x512 / 10 Steps: ~9.3 s
- 1024x1024 / 30 Steps: ~31.3 s
- 1024x1024 / 50 Steps: ~45.3 s
- 1920x1088 / 50 Steps: ~91 s
- Peak-VRAM: ~8.4-8.9 GB
- Hotswap-Gesamtzeit: ~41-42 s (1024x1024 / 30 Steps)

dev/fake-systemctl.sh

@@ -0,0 +1,49 @@
+#!/bin/bash
+# Fake systemctl für lokale Tests: verwaltet den Mock-llama.cpp-Prozess.
+# Simuliert: systemctl stop|start|status <service>
+#
+# Umgebungsvariablen (vom Router geerbt):
+#   FAKE_SYSTEMD_PIDFILE   PID-Datei des Mocks (Default /tmp/mock_upstream_pid)
+#   FAKE_SYSTEMD_PORT      Mock-Port (Default 18080)
+#   FAKE_SYSTEMD_PROFILE_DIR  Profil-Dir für den Mock
+#   FAKE_SYSTEMD_MOCK      Mock-Skript (Default dev/mock_upstream.py)
+#   FAKE_SYSTEMD_LOG       Log-Datei (Default /tmp/mock_upstream_fake.log)
+
+CMD="${1:-}"
+PIDFILE="${FAKE_SYSTEMD_PIDFILE:-/tmp/mock_upstream_pid}"
+PORT="${FAKE_SYSTEMD_PORT:-18080}"
+PROFILE_DIR="${FAKE_SYSTEMD_PROFILE_DIR:-}"
+MOCK="${FAKE_SYSTEMD_MOCK:-dev/mock_upstream.py}"
+LOG="${FAKE_SYSTEMD_LOG:-/tmp/mock_upstream_fake.log}"
+
+is_running() {
+  [ -f "$PIDFILE" ] && kill -0 "$(cat "$PIDFILE")" 2>/dev/null
+}
+
+case "$CMD" in
+  stop)
+    if is_running; then
+      kill "$(cat "$PIDFILE")" 2>/dev/null || true
+      rm -f "$PIDFILE"
+      for _ in $(seq 1 50); do
+        is_running || break
+        sleep 0.1
+      done
+    fi
+    exit 0
+    ;;
+  start)
+    if ! is_running; then
+      MOCK_PROFILE_DIR="$PROFILE_DIR" MOCK_PORT="$PORT" \
+        python3 "$MOCK" >>"$LOG" 2>&1 &
+      echo $! > "$PIDFILE"
+    fi
+    exit 0
+    ;;
+  status)
+    is_running && exit 0 || exit 3
+    ;;
+  *)
+    exit 0
+    ;;
+esac