Added functionality to allow the use of GPT on a Microsoft Azure instance

scripts/llm_utils.py

@@ -6,11 +6,20 @@ openai.api_key = cfg.openai_api_key
 
 # Overly simple abstraction until we create something better
 def create_chat_completion(messages, model=None, temperature=None, max_tokens=None)->str:
-    response = openai.ChatCompletion.create(
-        model=model,
-        messages=messages,
-        temperature=temperature,
-        max_tokens=max_tokens
-    )
+    if cfg.use_azure:
+        response = openai.ChatCompletion.create(
+            deployment_id=cfg.openai_deployment_id,
+            model=model,
+            messages=messages,
+            temperature=temperature,
+            max_tokens=max_tokens
+        )
+    else:
+        response = openai.ChatCompletion.create(
+            model=model,
+            messages=messages,
+            temperature=temperature,
+            max_tokens=max_tokens
+        )
 
     return response.choices[0].message["content"]