planner instructions workflow updated

scripts/format_response.py

@@ -291,7 +291,11 @@ def format_response_to_markdown(api_response, agent_name = None, dataframe=None)
                 markdown.append("### Summary\n")
                 for line in summary_lines:
                     if line != "":
-                        markdown.append(f"• {line.strip().replace('•', '').replace('-', '').replace('*', '') if line.strip().startswith('•') or line.strip().startswith('-') or line.strip().startswith('*') else line.strip()}\n")
+                        if line.strip().startswith('•') or line.strip().startswith('-') or line.strip().startswith('*'):
+                            line = line.strip().replace('•', '').replace('-', '').replace('*', '')
+                            markdown.append(f"* {line.strip()}\n")
+                        else:
+                            markdown.append(f"{line.strip()}\n")
 
             if 'refined_complete_code' in content and 'summary' in content:
                 try:

src/agents/agents.py

@@ -4,9 +4,11 @@ import asyncio
 from concurrent.futures import ThreadPoolExecutor
 import os
 from dotenv import load_dotenv
-
+# import logging
+# from src.utils.logger import Logger
 load_dotenv()
 
+# logger = Logger("agents", see_time=True, console_log=False)
 
 
 AGENTS_WITH_DESCRIPTION = {
@@ -920,26 +922,3 @@ class auto_analyst(dspy.Module):
                 yield name, inputs, result
             except Exception as e:
                 yield agent_name, inputs, {"error": str(e)}
-        # Execute code combiner after all agents complete
-        # code_list = [result['code'] for _, result in completed_results if 'code' in result]
-        # # max tokens is number of characters - number of words / 2
-        # char_count = sum(len(code) for code in code_list)
-        # word_count = sum(len(code.split()) for code in code_list)
-        # max_tokens = int((char_count - word_count) / 2)
-        # print(f"Max tokens: {max_tokens}")
-        # try:
-        #     with dspy.context(lm=dspy.LM(model="gemini/gemini-2.5-pro-preview-03-25", api_key = os.environ['GEMINI_API_KEY'], max_tokens=10000)):
-        #         combiner_result = self.code_combiner_agent(agent_code_list=str(code_list), dataset=dict_['dataset'])
-        #         yield 'code_combiner_agent__gemini', str(code_list), dict(combiner_result)
-        # except:
-        #     try: 
-        #         with dspy.context(lm=dspy.LM(model="o3-mini", max_tokens=max_tokens, temperature=1.0, api_key=os.getenv("OPENAI_API_KEY"))):
-        #             combiner_result = self.code_combiner_agent(agent_code_list=str(code_list), dataset=dict_['dataset'])
-        #             yield 'code_combiner_agent__openai', str(code_list), dict(combiner_result)
-        #     except:
-        #         try: 
-        #             with dspy.context(lm=dspy.LM(model="claude-3-7-sonnet-latest", max_tokens=max_tokens, temperature=1.0, api_key=os.getenv("ANTHROPIC_API_KEY"))):
-        #                 combiner_result = self.code_combiner_agent(agent_code_list=str(code_list), dataset=dict_['dataset'])
-        #                 yield 'code_combiner_agent__anthropic', str(code_list), dict(combiner_result)
-        #         except Exception as e:
-        #             yield 'code_combiner_agent__none', str(code_list), {"error": "Error in code combiner: "+str(e)}

src/agents/retrievers/retrievers.py

@@ -55,14 +55,14 @@ def make_data(df, desc):
     dict_['df_name'] = "The data is loaded as df"
     dict_['Description'] = desc
     dict_['dataframe_head_view'] = df.head(2).to_markdown()
-    dict_['all_column_names'] = str(list(df.columns[:20]))
-
-    for c in df.columns:
-        df[c] = correct_num(df,c)
-        try:
-            dict_[c] = {'column_name':c,'type':str(type(df[c].iloc[0])), 'column_information':return_vals(df,c)}
-        except:
-            dict_[c] = {'column_name':c,'type':str(type(df[c].iloc[0])), 'column_information':'NA'}    
+    # dict_['all_column_names'] = str(list(df.columns[:20]))
+
+    # for c in df.columns:
+    #     df[c] = correct_num(df,c)
+    #     try:
+    #         dict_[c] = {'column_name':c,'type':str(type(df[c].iloc[0])), 'column_information':return_vals(df,c)}
+    #     except:
+    #         dict_[c] = {'column_name':c,'type':str(type(df[c].iloc[0])), 'column_information':'NA'}    
     return dict_