Externalize OpenAI base URL

.devcontainer/devcontainer.json

@@ -16,6 +16,7 @@
     "USE_FLASK_DEBUG_MODE": "true",
     "FEEDS": "https://feed.infoq.com/development/,https://blog.jetbrains.com/feed/,https://feed.infoq.com/Devops/,https://feed.infoq.com/architecture-design/",
     "ROOT_LOG_LEVEL": "INFO",
-    "STARTER_LOG_LEVEL": "DEBUG"
+    "STARTER_LOG_LEVEL": "DEBUG",
+    "OPEN_AI_BASE_URL": "https://api.openai.com/v1/"
   }
 }

.env.example

@@ -1 +1,2 @@
 export OPEN_AI_KEY=fill_me_in
+export OPEN_AI_BASE_URL=fill_me_in

README.md

@@ -88,7 +88,7 @@ and uses pgvector to store the embeddings in PostgreSQL.
 
 The web application collects the user's query and creates an embedding with the OpenAI Embeddings API.
 It then searches the PostgreSQL for similar embeddings (using pgvector) and provides the corresponding chunk of text as
-context for a query to the [Azure AI Chat Completion API](https://learn.microsoft.com/en-us/azure/ai-services/openai/reference#completions).
+context for a query to the [Azure AI Chat Completion API](https://learn.microsoft.com/en-us/azure/ai-services/openai/reference#chat-completions).
 
 ## Local development
 
@@ -113,7 +113,7 @@ context for a query to the [Azure AI Chat Completion API](https://learn.microsof
     [localhost:5001](http://localhost:5001).
 
     ```shell
-    python collector.py
-    python analyzer.py
+    python collect.py
+    python analyze.py
     python -m starter
     ```

analyze.py

@@ -19,7 +19,7 @@
 chunks_gateway = ChunksGateway(db_template)
 embeddings_gateway = EmbeddingsGateway(db_template)
 ai_client = OpenAIClient(
-    base_url="https://api.openai.com/v1/",
+    base_url=env.open_ai_base_url,
     api_key=env.open_ai_key,
     embeddings_model="text-embedding-3-small",
     chat_model="gpt-4o"

starter/ai/open_ai_client.py

@@ -36,6 +36,7 @@ def fetch_embedding(self, text) -> Result[List[float]]:
             },
         )
         if not response.ok:
+            logger.error(f"Received {response.status_code} response from {self.base_url}: {response.text}")
             return Failure("Failed to fetch embedding")
 
         return Success(response.json()["data"][0]["embedding"])
@@ -55,6 +56,7 @@ def fetch_chat_completion(self, messages: List[ChatMessage]) -> Result[str]:
                 ]},
         )
         if not response.ok:
+            logger.error(f"Received {response.status_code} response from {self.base_url}: {response.text}")
             return Failure("Failed to fetch completion")
 
         return Success(response.json()["choices"][0]["message"]["content"])

starter/app.py

@@ -28,7 +28,7 @@ def create_app(env: Environment = Environment.from_env()) -> Flask:
     chunks_gateway = ChunksGateway(db_template)
     embeddings_gateway = EmbeddingsGateway(db_template)
     ai_client = OpenAIClient(
-        base_url="https://api.openai.com/v1/",
+        base_url=env.open_ai_base_url,
         api_key=env.open_ai_key,
         embeddings_model="text-embedding-3-small",
         chat_model="gpt-4o"

starter/environment.py

@@ -10,6 +10,7 @@ class Environment:
     use_flask_debug_mode: bool
     feeds: List[str]
     open_ai_key: str
+    open_ai_base_url: str
     root_log_level: str
     starter_log_level: str
 
@@ -21,6 +22,7 @@ def from_env(cls) -> 'Environment':
             use_flask_debug_mode=os.environ.get('USE_FLASK_DEBUG_MODE', 'false') == 'true',
             feeds=cls.__require_env('FEEDS').strip().split(','),
             open_ai_key=cls.__require_env('OPEN_AI_KEY'),
+            open_ai_base_url=cls.__require_env('OPEN_AI_BASE_URL'),
             root_log_level=os.environ.get('ROOT_LOG_LEVEL', 'INFO'),
             starter_log_level=os.environ.get('STARTER_LOG_LEVEL', 'INFO'),
         )

tests/ai/test_open_ai_client.py

@@ -5,6 +5,7 @@
 from starter.ai.open_ai_client import OpenAIClient, ChatMessage
 from tests.chat_support import chat_response
 from tests.embeddings_support import embedding_response, embedding_vector
+from tests.logging_support import disable_logging
 
 
 class TestOpenAIClient(unittest.TestCase):
@@ -22,6 +23,7 @@ def test_fetch_embedding(self):
 
         self.assertEqual(embedding_vector(2), self.client.fetch_embedding("some query").value)
 
+    @disable_logging
     @responses.activate
     def test_fetch_embedding_failure(self):
         responses.add(responses.POST, "https://openai.example.com/embeddings", "bad news", status=400)
@@ -39,6 +41,7 @@ def test_fetch_chat_completion(self):
             ]).value,
         )
 
+    @disable_logging
     @responses.activate
     def test_fetch_chat_completion_failure(self):
         responses.add(responses.POST, "https://openai.example.com/chat/completions", "bad news", status=400)