Update app.py

app.py

@@ -6,7 +6,9 @@ import plotly.graph_objects as go
 from scipy.stats import norm, t
 from scipy.cluster.hierarchy import linkage, dendrogram, fcluster
 import plotly.figure_factory as ff
-import openai
+import requests
+import time
+
 
 def sorting(df):
     df.index = list(map(float, df.index))
@@ -597,6 +599,84 @@ def categorize_responses(df, api_key, prompt=None):
 
     return categorized_df
 
+def categorize_responses(initial_prompt: str, 
+                        dataframe: pd.DataFrame,
+                        id_column: str,
+                        text_column: str,
+                        api_key: str,
+                        max_retries: int = 3,
+                        delay: float = 1.0) -> pd.DataFrame:
+    """
+    Categorizes survey responses using Deepseek API.
+    
+    Args:
+        initial_prompt: Context/problem statement/employer concerns
+        dataframe: DataFrame containing responses
+        id_column: Name of column with unique IDs
+        text_column: Name of column with text responses
+        api_key: Deepseek API key
+        max_retries: Maximum API call retries
+        delay: Delay between retries in seconds
+        
+    Returns:
+        DataFrame with added 'category' column
+    """
+    
+    # Validate dataframe structure
+    if id_column not in dataframe.columns or text_column not in dataframe.columns:
+        raise ValueError("Dataframe must contain specified ID and text columns")
+    
+    # Create API headers
+    headers = {
+        "Authorization": f"Bearer {api_key}",
+        "Content-Type": "application/json"
+    }
+    
+    # Define the API endpoint (verify correct endpoint from Deepseek documentation)
+    api_url = "https://api.deepseek.com/v1/chat/completions"
+    
+    def get_category(answer: str) -> str:
+        """Helper function to get category from API"""
+        messages = [
+            {
+                "role": "system",
+                "content": f"{initial_prompt}\n\nCategorize the following response into one of the appropriate categories."
+            },
+            {
+                "role": "user",
+                "content": answer
+            }
+        ]
+        
+        payload = {
+            "model": "deepseek-chat",  # Verify correct model name
+            "messages": messages,
+            "temperature": 0.2,
+            "max_tokens": 64
+        }
+        
+        for attempt in range(max_retries):
+            try:
+                response = requests.post(api_url, headers=headers, json=payload)
+                response.raise_for_status()
+                
+                # Parse response - adjust according to actual API response structure
+                result = response.json()['choices'][0]['message']['content'].strip()
+                return result
+                
+            except Exception as e:
+                if attempt == max_retries - 1:
+                    print(f"Failed after {max_retries} attempts: {str(e)}")
+                    return "Error: Categorization failed"
+                time.sleep(delay * (attempt + 1))
+        
+        return "Error: Max retries exceeded"
+    
+    # Apply categorization to each response
+    dataframe['category'] = dataframe[text_column].apply(get_category)
+    
+    return dataframe
+
 empty_col1, main_col, empty_col2 = st.columns([1.6, 2.8, 1.6])
 
 with main_col: