app.py

import os
from fastapi import FastAPI, HTTPException, Query, File, UploadFile, Form
from fastapi.responses import StreamingResponse, JSONResponse
from openai import AsyncOpenAI
import base64
from typing import Optional

app = FastAPI()

# Define available models (unchanged)
AVAILABLE_MODELS = {
    "openai/gpt-4.1": "OpenAI GPT-4.1",
    "openai/gpt-4.1-mini": "OpenAI GPT-4.1-mini",
    "openai/gpt-4.1-nano": "OpenAI GPT-4.1-nano",
    "openai/gpt-4o": "OpenAI GPT-4o",
    "openai/gpt-4o-mini": "OpenAI GPT-4o mini",
    "openai/o4-mini": "OpenAI o4-mini",
    "microsoft/MAI-DS-R1": "MAI-DS-R1",
    "microsoft/Phi-3.5-MoE-instruct": "Phi-3.5-MoE instruct (128k)",
    "microsoft/Phi-3.5-mini-instruct": "Phi-3.5-mini instruct (128k)",
    "microsoft/Phi-3.5-vision-instruct": "Phi-3.5-vision instruct (128k)",
    "microsoft/Phi-3-medium-128k-instruct": "Phi-3-medium instruct (128k)",
    "microsoft/Phi-3-medium-4k-instruct": "Phi-3-medium instruct (4k)",
    "microsoft/Phi-3-mini-128k-instruct": "Phi-3-mini instruct (128k)",
    "microsoft/Phi-3-small-128k-instruct": "Phi-3-small instruct (128k)",
    "microsoft/Phi-3-small-8k-instruct": "Phi-3-small instruct (8k)",
    "microsoft/Phi-4": "Phi-4",
    "microsoft/Phi-4-mini-instruct": "Phi-4-mini-instruct",
    "microsoft/Phi-4-multimodal-instruct": "Phi-4-multimodal-instruct",
    "ai21-labs/AI21-Jamba-1.5-Large": "AI21 Jamba 1.5 Large",
    "ai21-labs/AI21-Jamba-1.5-Mini": "AI21 Jamba 1.5 Mini",
    "mistral-ai/Codestral-2501": "Codestral 25.01",
    "cohere/Cohere-command-r": "Cohere Command R",
    "cohere/Cohere-command-r-08-2024": "Cohere Command R 08-2024",
    "cohere/Cohere-command-r-plus": "Cohere Command R+",
    "cohere/Cohere-command-r-plus-08-2024": "Cohere Command R+ 08-2024",
    "deepseek/DeepSeek-R1": "DeepSeek-R1",
    "deepseek/DeepSeek-V3-0324": "DeepSeek-V3-0324",
    "meta/Llama-3.2-11B-Vision-Instruct": "Llama-3.2-11B-Vision-Instruct",
    "meta/Llama-3.2-90B-Vision-Instruct": "Llama-3.2-90B-Vision-Instruct",
    "meta/Llama-3.3-70B-Instruct": "Llama-3.3-70B-Instruct",
    "meta/Llama-4-Maverick-17B-128E-Instruct-FP8": "Llama 4 Maverick 17B 128E Instruct FP8",
    "meta/Llama-4-Scout-17B-16E-Instruct": "Llama 4 Scout 17B 16E Instruct",
    "meta/Meta-Llama-3.1-405B-Instruct": "Meta-Llama-3.1-405B-Instruct",
    "meta/Meta-Llama-3.1-70B-Instruct": "Meta-Llama-3.1-70B-Instruct",
    "meta/Meta-Llama-3.1-8B-Instruct": "Meta-Llama-3.1-8B-Instruct",
    "meta/Meta-Llama-3-70B-Instruct": "Meta-Llama-3-70B-Instruct",
    "meta/Meta-Llama-3-8B-Instruct": "Meta-Llama-3-8B-Instruct",
    "mistral-ai/Ministral-3B": "Ministral 3B",
    "mistral-ai/Mistral-Large-2411": "Mistral Large 24.11",
    "mistral-ai/Mistral-Nemo": "Mistral Nemo",
    "mistral-ai/Mistral-large-2407": "Mistral Large (2407)",
    "mistral-ai/Mistral-small": "Mistral Small",
    "cohere/cohere-command-a": "Cohere Command A",
    "core42/jais-30b-chat": "JAIS 30b Chat",
    "mistral-ai/mistral-small-2503": "Mistral Small 3.1"
}

# Vision-capable models (subset of AVAILABLE_MODELS)
VISION_MODELS = [
    "openai/gpt-4o",
    "openai/gpt-4o-mini",
    "microsoft/Phi-3.5-vision-instruct",
    "meta/Llama-3.2-11B-Vision-Instruct",
    "meta/Llama-3.2-90B-Vision-Instruct",
    "microsoft/Phi-4-multimodal-instruct"
]

async def generate_ai_response(prompt: str, model: str):
    token = os.getenv("GITHUB_TOKEN")
    if not token:
        raise HTTPException(status_code=500, detail="GitHub token not configured")
    
    endpoint = "https://models.github.ai/inference"
    
    if model not in AVAILABLE_MODELS:
        raise HTTPException(status_code=400, detail=f"Model not available. Choose from: {', '.join(AVAILABLE_MODELS.keys())}")

    client = AsyncOpenAI(base_url=endpoint, api_key=token)

    try:
        stream = await client.chat.completions.create(
            messages=[
                {"role": "user", "content": prompt}
            ],
            model=model,
            temperature=1.0,
            top_p=1.0,
            stream=True
        )

        async for chunk in stream:
            if chunk.choices and chunk.choices[0].delta.content:
                yield chunk.choices[0].delta.content

    except Exception as err:
        yield f"Error: {str(err)}"
        raise HTTPException(status_code=500, detail="AI generation failed")

async def process_image_with_vision(image: bytes, question: str, model: str):
    token = os.getenv("GITHUB_TOKEN")
    if not token:
        raise HTTPException(status_code=500, detail="GitHub token not configured")
    
    endpoint = "https://models.github.ai/inference"
    
    if model not in VISION_MODELS:
        raise HTTPException(status_code=400, detail=f"Model not vision-capable. Choose from: {', '.join(VISION_MODELS)}")

    client = AsyncOpenAI(base_url=endpoint, api_key=token)

    # Encode image to base64
    base64_image = base64.b64encode(image).decode("utf-8")

    try:
        # Non-streaming request for vision task
        response = await client.chat.completions.create(
            messages=[
                {
                    "role": "user",
                    "content": [
                        {"type": "text", "text": question},
                        {
                            "type": "image_url",
                            "image_url": {"url": f"data:image/jpeg;base64,{base64_image}"}
                        }
                    ]
                }
            ],
            model=model,
            temperature=1.0,
            top_p=1.0,
            stream=False  # Vision tasks typically don't stream
        )

        return response.choices[0].message.content

    except Exception as err:
        raise HTTPException(status_code=500, detail=f"Vision processing failed: {str(err)}")

@app.post("/generate")
async def generate_response(
    prompt: str = Query(..., description="The prompt for the AI"),
    model: str = Query("openai/gpt-4.1-mini", description="The model to use for generation")
):
    if not prompt:
        raise HTTPException(status_code=400, detail="Prompt cannot be empty")
    
    return StreamingResponse(
        generate_ai_response(prompt, model),
        media_type="text/event-stream"
    )

@app.post("/process-image")
async def process_image(
    image: UploadFile = File(..., description="Image file (PNG, JPEG, GIF)"),
    question: str = Form(..., description="Question about the image"),
    model: str = Form("openai/gpt-4o", description="Vision-capable model")
):
    # Validate image format
    if not image.filename.lower().endswith((".png", ".jpg", ".jpeg", ".gif")):
        raise HTTPException(status_code=400, detail="Unsupported image format. Use PNG, JPEG, or GIF.")

    # Read image content
    image_data = await image.read()

    # Process image with vision model
    response = await process_image_with_vision(image_data, question, model)

    return response

def get_app():
    return app