feat: complete production-quality Model Router AI traffic control platform

backend/app/providers/base.py

@@ -0,0 +1,46 @@
+import abc
+import time
+import asyncio
+from typing import AsyncGenerator, Dict, Any, Optional
+from app.models.schemas import ProviderResponse
+
+
+class ModelProvider(abc.ABC):
+    """
+    Abstract Base Class for all Model Providers (Mock, Ollama, OpenAI, Anthropic, Gemini, etc.)
+    Ensures complete provider decoupling from routing logic.
+    """
+
+    def __init__(self, provider_id: str, name: str, base_url: Optional[str] = None):
+        self.provider_id = provider_id
+        self.name = name
+        self.base_url = base_url
+
+    @abc.abstractmethod
+    async def generate(
+        self,
+        prompt: str,
+        model_id: str,
+        system_prompt: Optional[str] = None,
+        temperature: float = 0.7,
+        max_tokens: Optional[int] = None,
+    ) -> ProviderResponse:
+        """Execute a text generation call against the model provider."""
+        pass
+
+    @abc.abstractmethod
+    async def stream(
+        self,
+        prompt: str,
+        model_id: str,
+        system_prompt: Optional[str] = None,
+        temperature: float = 0.7,
+        max_tokens: Optional[int] = None,
+    ) -> AsyncGenerator[str, None]:
+        """Stream generated response chunks."""
+        pass
+
+    @abc.abstractmethod
+    async def check_health(self) -> Dict[str, Any]:
+        """Check provider connectivity, status, and loaded models."""
+        pass