Add FastAPI backend proxy

app.py

@@ -0,0 +1,212 @@
+import os
+import base64
+import io
+import json
+from typing import Optional
+from fastapi import FastAPI, HTTPException, Request, Response
+from fastapi.staticfiles import StaticFiles
+from fastapi.middleware.cors import CORSMiddleware
+from pydantic import BaseModel
+import httpx
+import asyncio
+
+HF_TOKEN = os.environ.get("HF_TOKEN", "")
+if not HF_TOKEN:
+    print("WARNING: HF_TOKEN not set! AI features will fail.")
+
+app = FastAPI(title="Comic AI Generator", version="2.0")
+
+# Mount static files (the frontend)
+app.mount("/", StaticFiles(directory="static", html=True), name="static")
+
+# ================= MODELS =================
+class TextGenRequest(BaseModel):
+    model: str
+    prompt: str
+    max_new_tokens: int = 512
+    temperature: float = 0.7
+
+class ChatRequest(BaseModel):
+    model: str
+    messages: list
+    max_tokens: int = 1024
+    temperature: float = 0.3
+
+class ImageGenRequest(BaseModel):
+    model: str
+    prompt: str
+    negative_prompt: str = ""
+    width: int = 1024
+    height: int = 1024
+
+class InpaintRequest(BaseModel):
+    model: str
+    prompt: str
+    image_base64: str
+    mask_base64: str
+    negative_prompt: str = ""
+    width: int = 1024
+    height: int = 1024
+
+class TTSRequest(BaseModel):
+    model: str
+    text: str
+
+# ================= PROXY HELPERS =================
+async def hf_api_request(url: str, payload: dict):
+    headers = {
+        "Authorization": f"Bearer {HF_TOKEN}",
+        "Content-Type": "application/json"
+    }
+    for attempt in range(3):
+        try:
+            async with httpx.AsyncClient(timeout=120.0) as client:
+                response = await client.post(url, json=payload, headers=headers, timeout=120)
+                if response.status_code == 200:
+                    return response.json()
+                if response.status_code == 503 and attempt < 2:
+                    print(f"Model loading (503), retrying in 15s...")
+                    await asyncio.sleep(15)
+                    continue
+                error_text = response.text[:500]
+                print(f"HF API Error {response.status_code}: {error_text}")
+                raise HTTPException(status_code=response.status_code, detail=f"HF API Error {response.status_code}: {error_text}")
+        except httpx.RequestError as e:
+            print(f"Request error attempt {attempt+1}: {e}")
+            if attempt == 2:
+                raise HTTPException(status_code=500, detail=f"Network error: {str(e)}")
+            await asyncio.sleep(5)
+
+async def hf_binary_request(url: str, payload: dict):
+    headers = {
+        "Authorization": f"Bearer {HF_TOKEN}",
+        "Content-Type": "application/json"
+    }
+    for attempt in range(3):
+        try:
+            async with httpx.AsyncClient(timeout=120.0) as client:
+                response = await client.post(url, json=payload, headers=headers, timeout=120)
+                if response.status_code == 200:
+                    return Response(content=response.content, media_type="image/png")
+                if response.status_code == 503 and attempt < 2:
+                    await asyncio.sleep(15)
+                    continue
+                raise HTTPException(status_code=response.status_code, detail=f"Image API Error {response.status_code}")
+        except httpx.RequestError as e:
+            if attempt == 2:
+                raise HTTPException(status_code=500, detail=str(e))
+            await asyncio.sleep(5)
+
+# ================= ENDPOINTS =================
+@app.post("/api/text")
+async def generate_text(req: TextGenRequest):
+    url = f"https://api-inference.huggingface.co/models/{req.model}"
+    payload = {
+        "inputs": req.prompt,
+        "parameters": {
+            "max_new_tokens": req.max_new_tokens,
+            "temperature": req.temperature,
+            "return_full_text": False
+        }
+    }
+    result = await hf_api_request(url, payload)
+    return result
+
+@app.post("/api/chat")
+async def chat(req: ChatRequest):
+    url = "https://router.huggingface.co/v1/chat/completions"
+    payload = {
+        "model": req.model,
+        "messages": req.messages,
+        "max_tokens": req.max_tokens,
+        "temperature": req.temperature
+    }
+    result = await hf_api_request(url, payload)
+    return result
+
+@app.post("/api/image")
+async def generate_image(req: ImageGenRequest):
+    url = f"https://api-inference.huggingface.co/models/{req.model}"
+    payload = {
+        "inputs": req.prompt,
+        "parameters": {
+            "negative_prompt": req.negative_prompt,
+            "width": req.width,
+            "height": req.height,
+            "num_inference_steps": 30,
+            "guidance_scale": 7.5
+        }
+    }
+    return await hf_binary_request(url, payload)
+
+@app.post("/api/inpaint")
+async def inpaint_image(req: InpaintRequest):
+    url = f"https://api-inference.huggingface.co/models/{req.model}"
+    payload = {
+        "inputs": req.prompt,
+        "parameters": {
+            "image": f"data:image/png;base64,{req.image_base64}",
+            "mask": f"data:image/png;base64,{req.mask_base64}",
+            "negative_prompt": req.negative_prompt,
+            "num_inference_steps": 30,
+            "guidance_scale": 7.5,
+            "width": req.width,
+            "height": req.height
+        }
+    }
+    return await hf_binary_request(url, payload)
+
+@app.post("/api/tts")
+async def text_to_speech(req: TTSRequest):
+    url = f"https://api-inference.huggingface.co/models/{req.model}"
+    payload = {"inputs": req.text}
+    
+    for attempt in range(3):
+        try:
+            async with httpx.AsyncClient(timeout=60.0) as client:
+                response = await client.post(url, json=payload,
+                    headers={"Authorization": f"Bearer {HF_TOKEN}", "Content-Type": "application/json"},
+                    timeout=60
+                )
+                if response.status_code == 200:
+                    return Response(content=response.content, media_type="audio/wav")
+                if response.status_code == 503 and attempt < 2:
+                    await asyncio.sleep(10)
+                    continue
+                raise HTTPException(status_code=response.status_code, detail=f"TTS error: {response.status_code}")
+        except Exception as e:
+            if attempt == 2:
+                raise HTTPException(status_code=500, detail=str(e))
+            await asyncio.sleep(3)
+
+@app.get("/api/health")
+async def health_check():
+    if not HF_TOKEN:
+        return {"status": "error", "message": "HF_TOKEN not configured in backend"}
+    try:
+        async with httpx.AsyncClient(timeout=30.0) as client:
+            response = await client.post(
+                "https://api-inference.huggingface.co/models/Qwen/Qwen3-0.6B",
+                json={"inputs": "Hi", "parameters": {"max_new_tokens": 3}},
+                headers={"Authorization": f"Bearer {HF_TOKEN}"},
+                timeout=30
+            )
+            if response.status_code == 200:
+                return {"status": "ok", "message": "API connected successfully"}
+            elif response.status_code == 503:
+                return {"status": "loading", "message": "Model is warming up, please wait ~1 min"}
+            else:
+                return {"status": "error", "message": f"HF API returned {response.status_code}"}
+    except Exception as e:
+        return {"status": "error", "message": f"Connection failed: {str(e)}"}
+
+@app.get("/api/models")
+async def list_models():
+    return {
+        "text_model": "Qwen/Qwen3-0.6B",
+        "vision_model": "Qwen/Qwen3-VL-2B-Instruct",
+        "image_model": "stabilityai/stable-diffusion-xl-base-1.0",
+        "inpaint_model": "diffusers/stable-diffusion-xl-1.0-inpainting-0.1",
+        "tts_model": "microsoft/speecht5_tts",
+        "token_configured": bool(HF_TOKEN)
+    }