fix compatibility issue with newer models

based on https://platform.openai.com/docs/guides/gpt/chat-completions-vs-completions

Author: pilvar222

src/query.py

@@ -22,16 +22,25 @@ def run(self) -> None:
 
         log_debug(f'Sending query: {self.query_string}')
 
-        response = openai.Completion.create(
-            model=self.model,
-            prompt=self.query_string,
-            max_tokens=self.max_token_count,
-        )
-        # Get the response text.
-        result: str = response.choices[0].text
+        if self.model in ["gpt-3.5-turbo","gpt-4","gpt-4-32k"]:
+            response = openai.ChatCompletion.create(
+                model=self.model,
+                messages=[{"role":"user","content":self.query_string}],
+                max_tokens=self.max_token_count,
+            )
+            # Get the response text.
+            result: str = response.choices[0].message.content
+        else:
+            response = openai.Completion.create(
+                model=self.model,
+                prompt=self.query_string,
+                max_tokens=self.max_token_count,
+            )
+            # Get the response text.
+            result: str = response.choices[0].text
         # If there is a callback, do something with it.
         if self.callback:
             self.callback(result)
         # Otherwise, assume we just want to log it.
         else:
-            log_info(result)
+            log_info(result)