api key required for openai embeddings

.vscode/launch.json

@@ -2,13 +2,19 @@
   "version": "0.2.0",
   "configurations": [
     {
-      "name": "Know Lang Bot",
+      "name": "Know Lang",
       "type": "debugpy",
       "request": "launch",
       "module": "know_lang_bot",
       "args": [
         "--v"
       ]
+    },
+    {
+      "name": "Know Lang ChatBot ",
+      "type": "debugpy",
+      "request": "launch",
+      "program": "${workspaceFolder}/src/know_lang_bot/chat_bot/gradio_demo.py",
     }
   ]
 }

src/know_lang_bot/chat_bot/chat_graph.py

@@ -154,7 +154,7 @@ class RetrieveContextNode(BaseNode[ChatGraphState, ChatGraphDeps, ChatResult]):
         try:
             question_embedding = generate_embedding(
                 input=ctx.state.polished_question or ctx.state.original_question,
-                model=ctx.deps.config.embedding
+                config=ctx.deps.config.embedding
             )
 
             results = ctx.deps.collection.query(

src/know_lang_bot/config.py

@@ -77,6 +77,18 @@ class EmbeddingConfig(BaseSettings):
         default_factory=dict,
         description="Provider-specific settings"
     )
+    api_key: Optional[str] = Field(
+        default=None,
+        description="API key for the model provider"
+    )
+
+    @field_validator('api_key', mode='after')
+    @classmethod
+    def validate_api_key(cls, v: Optional[str], info: ValidationInfo) -> Optional[str]:
+        """Validate API key is present when required"""
+        if info.data['provider'] in [ModelProvider.OPENAI] and not v:
+            raise ValueError(f"API key required for {info.data['provider']}")
+        return v
 
 class LLMConfig(BaseSettings):
     model_name: str = Field(

src/know_lang_bot/models/embeddings.py

@@ -6,11 +6,6 @@ from typing import Union, List, overload
 # Type definitions
 EmbeddingVector = List[float]
 
-class EmbeddingConfig:
-    def __init__(self, provider: ModelProvider, model_name: str):
-        self.provider = provider
-        self.model_name = model_name
-
 def _process_ollama_batch(inputs: List[str], model_name: str) -> List[EmbeddingVector]:
     """Helper function to process Ollama embeddings in batch."""
     return [
@@ -60,6 +55,7 @@ def generate_embedding(
         if config.provider == ModelProvider.OLLAMA:
             embeddings = _process_ollama_batch(inputs, config.model_name)
         elif config.provider == ModelProvider.OPENAI:
+            openai.api_key = config.api_key
             embeddings = _process_openai_batch(inputs, config.model_name)
         else:
             raise ValueError(f"Unsupported provider: {config.provider}")