binapp2.py

import gradio as gr
import requests
import json
import os
from typing import List, Dict, Optional

class BinericAPI:
    def __init__(self):
        self.api_key = os.environ.get("Key")
       
        if not self.api_key or self.api_key == "YOUR_API_KEY":
            raise ValueError("API Key not found. Please add your API key in the 'Secrets' tab (Key = 'YOUR_API_KEY')")
        
        self.base_url = "https://api.bineric.com"
        self.headers = {
            'api-key': self.api_key,
            'Content-Type': 'application/json'
        }
    
    def get_balance(self):
        """Fetch balance from Bineric API"""
        try:
            response = requests.get(
                f'{self.base_url}/api/v1/monitoring/balance',
                headers=self.headers
            )
            response.raise_for_status()
            return response.json()
        except requests.exceptions.RequestException as e:
            return {"success": False, "error": f"API request failed: {str(e)}"}
    
    def chat_completion(self, model: str, messages: List[Dict], 
                       stream: bool = False, temperature: float = 0.7,
                       max_tokens: Optional[int] = None,
                       top_p: float = 0.95):
        """Get chat completion from various AI models"""
        # Try multiple possible endpoint patterns
        endpoints_to_try = [
            f'{self.base_url}/api/v1/ai/chat/completions',  # Most likely based on balance endpoint
            f'{self.base_url}/api/ai/chat/completions',
            f'{self.base_url}/api/v1/chat/completions',
            f'{self.base_url}/v1/chat/completions',
            f'{self.base_url}/chat/completions',
            f'{self.base_url}/api/chat/completions',
        ]
        
        # Prepare request payload
        payload = {
            "model": model,
            "messages": messages,
            "options": {
                "temperature": temperature,
                "top_p": top_p
            }
        }
        
        if max_tokens:
            payload["options"]["max_response_length"] = max_tokens
        if stream:
            payload["options"]["stream"] = stream
        
        print(f"\n=== Trying chat completion with model: {model} ===")
        print(f"Messages: {len(messages)}")
        
        last_error = None
        last_response = None
        
        for endpoint in endpoints_to_try:
            try:
                print(f"\nTrying endpoint: {endpoint}")
                print(f"Payload keys: {list(payload.keys())}")
                
                response = requests.post(
                    endpoint,
                    headers=self.headers,
                    json=payload,
                    timeout=30
                )
                
                print(f"Status code: {response.status_code}")
                
                if response.status_code == 200:
                    result = response.json()
                    print(f"Success! Response keys: {list(result.keys())}")
                    return result
                elif response.status_code == 404:
                    print(f"Endpoint not found: {endpoint}")
                    last_response = response
                    continue
                else:
                    # Try to get error message
                    try:
                        error_data = response.json()
                        print(f"Error response: {json.dumps(error_data, indent=2)}")
                    except:
                        print(f"Error text: {response.text[:200]}")
                    last_response = response
                    
            except requests.exceptions.RequestException as e:
                print(f"Request failed: {str(e)}")
                last_error = e
                continue
        
        # If all endpoints failed
        error_msg = "All endpoints failed. "
        if last_response:
            error_msg += f"Last status: {last_response.status_code}. "
            try:
                error_data = last_response.json()
                error_msg += f"Response: {json.dumps(error_data)}"
            except:
                error_msg += f"Response text: {last_response.text[:200]}"
        elif last_error:
            error_msg += f"Last error: {str(last_error)}"
        
        return {"success": False, "error": error_msg}

# Initialize API client
try:
    api_client = BinericAPI()
except ValueError as e:
    api_client = None

# Available models - simplified list for testing
MODELS = [
    "gemini-2.0-flash",
    "gpt-5",
    "gpt-4",
    "claude-haiku-4-5",
    
    "deepseek-v3.1-terminus"
]

def get_balance():
    """Fetch and display balance"""
    if not api_client:
        return "❌ API Key not found. Please add your API key in the 'Secrets' tab (Key = 'YOUR_API_KEY')"
    
    balance_data = api_client.get_balance()
    
    if not balance_data.get("success", True):
        return f"❌ Error: {balance_data.get('error', 'Unknown error')}"
    
    formatted_response = json.dumps(balance_data, indent=2)
    return f"✅ Balance retrieved successfully:\n\n{formatted_response}"

def test_endpoints():
    """Test all possible endpoints"""
    if not api_client:
        return "❌ API Key not found"
    
    test_message = "Hello, please respond with just 'Test successful'."
    messages = [{"role": "user", "content": test_message}]
    
    result = api_client.chat_completion(
        model="gemini-2.0-flash",
        messages=messages,
        temperature=0.7,
        max_tokens=50
    )
    
    return json.dumps(result, indent=2)

def chat_complete(model, user_message, temperature, max_tokens, top_p, history=None):
    """Generate chat completion - simplified version"""
    if not api_client:
        return "", history, "❌ API Key not found. Please add your API key in the 'Secrets' tab"
    
    if not user_message.strip():
        return "", history, "❌ Please enter a message"
    
    # Prepare messages - simple format
    messages = [{"role": "user", "content": user_message}]
    
    # Get completion
    result = api_client.chat_completion(
        model=model,
        messages=messages,
        temperature=temperature,
        max_tokens=max_tokens if max_tokens > 0 else 1000,
        top_p=top_p
    )
    
    # Check for success
    if not result.get("success", True):
        error_msg = result.get("error", "Unknown error")
        
        # Try to provide helpful suggestions
        suggestions = ""
        if "404" in error_msg or "Not Found" in error_msg:
            suggestions = "\n\n🔍 **Possible solutions:**\n"
            suggestions += "1. The chat endpoint might be different\n"
            suggestions += "2. Check the Bineric API documentation\n"
            suggestions += "3. Try a different model\n"
            suggestions += "4. Contact support for the correct endpoint"
        
        return "", history, f"❌ Error: {error_msg}{suggestions}"
    
    # Extract response text
    response_text = result.get("response", "")
    if not response_text:
        # Try other possible response fields
        for field in ["content", "text", "output", "message", "choices"]:
            if field in result:
                if field == "choices" and isinstance(result[field], list) and len(result[field]) > 0:
                    choice = result[field][0]
                    if "message" in choice and "content" in choice["message"]:
                        response_text = choice["message"]["content"]
                        break
                    elif "text" in choice:
                        response_text = choice["text"]
                        break
                else:
                    response_text = str(result[field])
                    break
        
        if not response_text:
            response_text = "⚠️ Received response but could not parse format\n"
            response_text += f"Raw response keys: {list(result.keys())}"
    
    # Update history
    if history is None:
        history = []
    
    history.append({
        "user": user_message,
        "assistant": response_text[:2000]  # Limit length
    })
    
    # Format usage info if available
    usage_info = ""
    if "usage" in result:
        usage = result["usage"]
        usage_info = f"\n\n📊 **Usage:**\n"
        
        # Basic info
        if "prompt_tokens" in usage:
            usage_info += f"- Prompt tokens: {usage['prompt_tokens']}\n"
        if "completion_tokens" in usage:
            usage_info += f"- Completion tokens: {usage['completion_tokens']}\n"
        if "total_tokens" in usage:
            usage_info += f"- Total tokens: {usage['total_tokens']}\n"
        if "total_cost" in usage:
            usage_info += f"- Cost: ${usage['total_cost']:.6f}\n"
    
    return "", history, response_text + usage_info

def clear_chat():
    """Clear chat history"""
    return [], "", ""

# Create Gradio interface
with gr.Blocks(title="Bineric AI Dashboard") as demo:
    gr.Markdown("# 🤖 Bineric AI Dashboard")
    gr.Markdown("Balance checking and AI chat completion")
    
    # Tabs for different functionalities
    with gr.Tabs():
        # Tab 1: Balance Checker
        with gr.Tab("💰 Balance"):
            gr.Markdown("### Check your API balance")
            
            balance_output = gr.Textbox(
                label="Balance Information",
                placeholder="Click 'Get Balance' to see your balance...",
                lines=10
            )
            
            with gr.Row():
                get_balance_btn = gr.Button("🔄 Get Balance", variant="primary")
                clear_balance_btn = gr.Button("🗑️ Clear")
            
            get_balance_btn.click(get_balance, outputs=balance_output)
            clear_balance_btn.click(lambda: "", outputs=balance_output)
        
        # Tab 2: Chat
        with gr.Tab("💬 Chat Test"):
            gr.Markdown("### Test Chat Completion")
            
            with gr.Row():
                # Left column
                with gr.Column(scale=1):
                    model_selector = gr.Dropdown(
                        label="Select Model",
                        choices=MODELS,
                        value="gemini-2.0-flash",
                        interactive=True
                    )
                    
                    user_input = gr.Textbox(
                        label="Message",
                        placeholder="Type your message here...",
                        lines=3,
                        value="Hello, how are you?"
                    )
                    
                    with gr.Accordion("Parameters", open=False):
                        temperature = gr.Slider(0.0, 2.0, 0.7, step=0.1, label="Temperature")
                        max_tokens = gr.Slider(0, 4000, 500, step=50, label="Max Tokens")
                        top_p = gr.Slider(0.0, 1.0, 0.95, step=0.05, label="Top-p")
                    
                    send_btn = gr.Button("🚀 Send Message", variant="primary")
                
                # Right column
                with gr.Column(scale=2):
                    response_output = gr.Textbox(
                        label="Response",
                        placeholder="Response will appear here...",
                        lines=15
                    )
            
            # Simple send function
            def send_message(model, message, temp, tokens, top_p_val):
                if not api_client:
                    return "❌ API Key not found"
                
                if not message.strip():
                    return "❌ Please enter a message"
                
                messages = [{"role": "user", "content": message}]
                
                result = api_client.chat_completion(
                    model=model,
                    messages=messages,
                    temperature=temp,
                    max_tokens=tokens if tokens > 0 else None,
                    top_p=top_p_val
                )
                
                if not result.get("success", True):
                    return f"❌ Error: {result.get('error', 'Unknown error')}"
                
                # Extract response
                response_text = result.get("response", "")
                if not response_text:
                    return f"⚠️ No response text. Raw result:\n{json.dumps(result, indent=2)}"
                
                # Add usage info
                if "usage" in result:
                    usage = result["usage"]
                    response_text += f"\n\n📊 Usage: {usage.get('total_tokens', '?')} tokens"
                    if "total_cost" in usage:
                        response_text += f", ${usage['total_cost']:.6f} cost"
                
                return response_text
            
            send_btn.click(
                send_message,
                inputs=[model_selector, user_input, temperature, max_tokens, top_p],
                outputs=response_output
            )
            
            user_input.submit(
                send_message,
                inputs=[model_selector, user_input, temperature, max_tokens, top_p],
                outputs=response_output
            )
        
        # Tab 3: Endpoint Debug
        with gr.Tab("🔧 Debug"):
            gr.Markdown("### Endpoint Testing")
            
            with gr.Accordion("Test Endpoints", open=True):
                test_output = gr.Textbox(
                    label="Endpoint Test Results",
                    lines=20,
                    interactive=False
                )
                test_btn = gr.Button("🧪 Run Endpoint Tests", variant="primary")
                test_btn.click(test_endpoints, outputs=test_output)
            
            with gr.Accordion("Current Balance", open=False):
                debug_balance = gr.Textbox(
                    label="Balance",
                    lines=5,
                    interactive=False
                )
                refresh_btn = gr.Button("🔄 Refresh")
                refresh_btn.click(get_balance, outputs=debug_balance)
            
            gr.Markdown("""
            **Endpoint patterns being tested:**
            1. `/api/v1/ai/chat/completions`
            2. `/api/ai/chat/completions`
            3. `/api/v1/chat/completions`
            4. `/v1/chat/completions`
            5. `/chat/completions`
            6. `/api/chat/completions`
            
            **Note:** Check console/logs for detailed output
            """)
    
    # Footer
    gr.Markdown("""
    ---
    **Status:** 
    - ✅ Balance endpoint works: `/api/v1/monitoring/balance`
    - 🔍 Testing chat completion endpoint...
    - Check the Debug tab for endpoint testing results
    """)

# Launch the app
if __name__ == "__main__":
    demo.launch(share=True)