Update helper/utils.py

helper/utils.py

@@ -52,11 +52,6 @@ def current_year():
 #     return [text_list, sources_list]
 
 
-from typing import List, Tuple
-
-import PyPDF2
-
-
 def read_and_textify(
     files: List[str], chunk_size: int = 2  # Default chunk size set to 50
 ) -> Tuple[List[str], List[str]]:
@@ -134,6 +129,43 @@ def list_to_nums(sentences: List[str]) -> List[List[float]]:
     return embeddings
 
 
+def call_gpt4(prompt: str) -> str:
+    """
+    Sends a prompt to the GPT-4 model and retrieves a response.
+    
+    This function interacts with the OpenAI API, specifically using
+    the gpt-3.5-turbo model to generate a conversational response.
+    It simulates a conversation by providing system messages,
+    past interactions, and the latest user query.
+    
+    Args:
+    - prompt (str): The message from the user for which the GPT model will generate a response.
+    
+    Returns:
+    - str: The content of the message generated by the GPT model in response to the prompt.
+    
+    Note: This function assumes that 'client' is an instance of OpenAI's client object
+          that has been properly authenticated and initialized elsewhere in your code.
+    """
+    
+    # Interact with the OpenAI API to get a response for the prompt provided
+    response = client.chat.completions.create(
+        model="gpt-3.5-turbo",  # Specifies the AI model to use for the response
+        messages=[  # Constructs the context and the prompt for the AI
+            {"role": "system", "content": "You are a helpful assistant."},
+            {"role": "user", "content": "Who won the world series in 2020?"},
+            {
+                "role": "assistant",
+                "content": "The Los Angeles Dodgers won the World Series in 2020."
+            },
+            {"role": "user", "content": "Where was it played?"}
+        ]
+    )
+    
+    # Return the AI's response to the user's most recent prompt
+    return response.choices[0].message.content
+
+
 def quantize_to_kbit(arr: Union[np.ndarray, Any], k: int = 16) -> np.ndarray:
     """Converts an array to a k-bit representation by normalizing and scaling its values.
 
@@ -256,4 +288,4 @@ def query_search(
     # Sort the DataFrame based on the 'qim' score in descending order
     refs = refs.sort_values(by="qim", ascending=False)
 
-    return refs
+    return refs