app.py

# app.py
import os
import base64
import json
import gradio as gr
from huggingface_hub import upload_file, InferenceClient
from datetime import datetime
import traceback
from typing import Optional, Dict, Any

from fastmcp import FastMCP

# --- Configuration ---
HF_DATASET_REPO = os.environ.get("HF_DATASET_REPO", "OppaAI/Robot_MCP")
HF_VLM_MODEL = os.environ.get("HF_VLM_MODEL", "Qwen/Qwen2.5-VL-7B-Instruct")

mcp = FastMCP("Robot_MCP")

# -----------------------------------------------------
# Register Robot Tools (MCP)
# -----------------------------------------------------
@mcp.tool()
def speak(text: str, emotion: str = "neutral"):
    """Makes the robot speak a given text with an emotion."""
    return {"status": "success", "action_executed": "speak", "payload": {"text": text, "emotion": emotion}}

@mcp.tool()
def navigate(direction: str, distance_meters: float):
    """Moves the robot a specified distance in a direction (max 5m)."""
    if distance_meters > 5.0:
        return {"status": "error", "message": "Safety limit exceeded"}
    return {"status": "success", "action_executed": "navigate", "payload": {"direction": direction, "distance": distance_meters}}

@mcp.tool()
def scan_hazard(hazard_type: str, severity: str):
    """Logs a potential hazard detected by the robot."""
    timestamp = datetime.now().isoformat()
    return {"status": "warning_logged", "log": f"[{timestamp}] HAZARD: {hazard_type} (Severity: {severity})"}

@mcp.tool()
def analyze_human(clothing_color: str, estimated_action: str):
    """Tracks human activity based on visual input."""
    return {"status": "human_tracked", "details": f"Human wearing {clothing_color} is {estimated_action}"}

# -----------------------------------------------------
# Save and upload image to HF
# -----------------------------------------------------
def save_and_upload_image(image_b64: str, hf_token: str):
    """Decodes a base64 image, saves it locally, and uploads to Hugging Face Hub."""
    try:
        image_bytes = base64.b64decode(image_b64)
        size_bytes = len(image_bytes)

        # Ensure the /tmp directory exists
        os.makedirs("/tmp", exist_ok=True)
        timestamp = datetime.now().strftime("%Y%m%d_%H%M%S_%f")
        local_path = f"/tmp/robot_img_{timestamp}.jpg"

        with open(local_path, "wb") as f:
            f.write(image_bytes)

        filename = f"robot_{timestamp}.jpg"

        # Corrected Hugging Face Hub upload
        from huggingface_hub import HfApi
        api = HfApi()
        api.upload_file(
            path_or_fileobj=local_path,
            path_in_repo=f"tmp/{filename}",
            repo_id=HF_DATASET_REPO,
            repo_type="dataset",
            token=hf_token
        )

        url = f"https://huggingface.co/datasets/{HF_DATASET_REPO}/resolve/main/{filename}"
        return local_path, url, filename, size_bytes

    except Exception as e:
        print(f"[Error] during image upload: {e}")
        traceback.print_exc()
        return None, None, None, 0

# -----------------------------------------------------
# JSON parsing helper
# -----------------------------------------------------
def safe_parse_json_from_text(text: str) -> Optional[Dict[str, Any]]:
    """Safely extract JSON from messy VLM output"""
    if not text:
        return None
    try:
        return json.loads(text)
    except:
        pass
    cleaned = text.strip().strip("`").strip()
    if cleaned.lower().startswith("json"):
        cleaned = cleaned[4:].strip()
    try:
        start = cleaned.find("{")
        end = cleaned.rfind("}")
        if start >= 0 and end > start:
            return json.loads(cleaned[start:end+1])
    except:
        return None
    return None

# -----------------------------------------------------
# Call MCP tool safely using public API
# -----------------------------------------------------
def validate_and_call_tool(tool_name: str, tool_args: dict) -> Dict[str, Any]:
    """Use public API instead of _tools"""
    try:
        # FastMCP v2.11.2 provides call_tool
        if hasattr(mcp, "call_tool"):
            return mcp.call_tool(tool_name, tool_args)
        # fallback: call the registered function directly
        if hasattr(mcp, tool_name):
            tool_fn = getattr(mcp, tool_name)
            return tool_fn(**tool_args)
        return {"error": f"Unknown tool '{tool_name}'"}
    except Exception as e:
        traceback.print_exc()
        return {"error": f"Tool execution error: {str(e)}"}

# -----------------------------------------------------
# Main pipeline: image → VLM → tool
# -----------------------------------------------------
def process_and_describe(payload: Dict[str, Any]) -> Dict[str, Any]:
    if isinstance(payload, str):
        try:
            payload = json.loads(payload)
        except:
            return {"error": "Invalid JSON payload"}

    hf_token = payload.get("hf_token")
    if not hf_token:
        return {"error": "hf_token missing"}

    robot_id = payload.get("robot_id", "unknown")
    image_b64 = payload.get("image_b64")
    if not image_b64:
        return {"error": "image_b64 missing"}

    # Save + Upload
    _, hf_url, _, size_bytes = save_and_upload_image(image_b64, hf_token)
    if not hf_url:
        return {"error": "Image upload failed"}

    # VLM system prompt
    system_prompt = f"""
Respond in STRICT JSON ONLY:
{{
 "description": "short visual description",
 "tool_name": "speak | navigate | scan_hazard | analyze_human",
 "arguments": {{ ... }}
}}
"""
    messages = [
        {"role": "system", "content": system_prompt},
        {"role": "user", "content": [
            {"type": "text", "text": "Analyze the image and choose ONE tool."},
            {"type": "image_url", "image_url": {"url": f"data:image/jpeg;base64,{image_b64}"}}
        ]}
    ]

    client = InferenceClient(token=hf_token)
    try:
        response = client.chat.completions.create(
            model=HF_VLM_MODEL,
            messages=messages,
            max_tokens=300,
            temperature=0.1,
        )
    except Exception as e:
        return {"status": "error", "message": f"Inference API call failed: {e}"}

    vlm_output = response.choices[0].message.content.strip()
    parsed = safe_parse_json_from_text(vlm_output)
    if parsed is None:
        return {"status": "model_no_json", "robot_id": robot_id, "image_url": hf_url, "vlm_raw": vlm_output, "message": "VLM returned invalid JSON"}

    tool_name = parsed.get("tool_name")
    tool_args = parsed.get("arguments") or {}
    tool_result = validate_and_call_tool(tool_name, tool_args)

    return {
        "status": "success",
        "robot_id": robot_id,
        #"image_url": hf_url,
        "file_size_bytes": size_bytes,
        "vlm_description": parsed.get("description"),
        "chosen_tool": tool_name,
        "tool_arguments": tool_args,
        "tool_execution_result": tool_result,
        "vlm_raw": vlm_output
    }

# ------------------------------
# Gradio Interface
# ------------------------------
iface = gr.Interface(
    fn=process_and_describe,
    inputs=gr.JSON(label="Input JSON Payload (must include hf_token & image_b64)"),
    outputs=gr.JSON(label="Output JSON Result"),
    api_name="predict",
    flagging_mode="never"
)

# ------------------------------
# Main Entry
# ------------------------------
if __name__ == "__main__":
    print(f"[Config] HF_DATASET_REPO: {HF_DATASET_REPO}")
    print(f"[Config] HF_VLM_MODEL: {HF_VLM_MODEL}")
    print("[Gradio] Launching interface...")
    iface.launch(server_name="0.0.0.0", server_port=7860)