clean + clear convo/logging directory locations

README.md

@@ -53,9 +53,10 @@ Where:
 - `pep` is the provider extra parameters
 - `t` is the number of turns
 - `r` is the run per turns
-- `f` is the folder name (defaults to conversations and a subfolder named based on other paramters and datetime)
+- `f` (or `--convo-folder-name`) is the conversation folder name (defaults to `conversations` with a subfolder named based on other parameters and datetime)
+- `lf` (or `--log-folder-name`) is the logging folder name (defaults to `logging` with a subfolder named based on other parameters and datetime)
 - `c` is the maximum concurrent conversations to run (defaults to None, but try this if the provider you're testing times out)
-This will generate conversations and store them in a subfolder of `conversations`
+This will generate conversations and store them in subfolders of `conversations` and `logging`
 
 6. **Judge the conversations**:
    ```bash
@@ -90,7 +91,9 @@ python judge.py -f conversations/my_experiment -j gpt-4o -jep temperature=0.5,ma
 ```
 
 **Note:** Extra parameters are automatically included in the output folder names, making it easy to track experiments:
-- Generation: `conversations/p_gpt_4o_temp0.3__a_claude_3_5_sonnet_temp0.5__t6__r2__{timestamp}/`
+- Generation: 
+  - Conversations: `conversations/p_gpt_4o_temp0.3__a_claude_3_5_sonnet_temp0.5__t6__r2__{timestamp}/`
+  - Logs: `logging/p_gpt_4o_temp0.3__a_claude_3_5_sonnet_temp0.5__t6__r2__{timestamp}/`
 - Evaluation: `evaluations/j_claude_3_5_sonnet_temp0.3_{timestamp}__{conversation_folder}/`
 
 **Multiple judge models**: You can use multiple different judge models and/or multiple instances:
@@ -305,7 +308,8 @@ results = await generate_conversations(
     max_turns=5,
     runs_per_prompt=3,
     persona_names=["Alex M.", "Chloe Kim"],  # Optional: filter specific personas
-    folder_name="custom_experiment"  # Optional: custom output folder
+    convo_folder_name="custom_conversations",  # Optional: custom conversation folder
+    log_folder_name="custom_logging"  # Optional: custom logging folder
 )
 ```
 
@@ -371,15 +375,18 @@ persona_model_config = {
 
 ### Output Organization
 
-Conversations are automatically organized into timestamped folders:
+Conversations and logs are automatically organized into separate timestamped folders:
 
 ```
 conversations/
-├── p_claude_sonnet_4_20250514__a_claude_sonnet_4_20250514_20250120_143022_t5_r3/
-│   ├── abc123_Alex_M_c3s_run1_20250120_143022_123.txt
-│   ├── abc123_Alex_M_c3s_run1_20250120_143022_123.log
-│   ├── def456_Chloe_Kim_c3s_run1_20250120_143022_456.txt
-│   └── def456_Chloe_Kim_c3s_run1_20250120_143022_456.log
+└── p_claude_sonnet_4_20250514__a_claude_sonnet_4_20250514_20250120_143022_t5_r3/
+    ├── abc123_Alex_M_c3s_run1.txt
+    └── def456_Chloe_Kim_c3s_run1.txt
+
+logging/
+└── p_claude_sonnet_4_20250514__a_claude_sonnet_4_20250514_20250120_143022_t5_r3/
+    ├── abc123_Alex_M_c3s_run1.log
+    └── def456_Chloe_Kim_c3s_run1.log
 ```
 
 ### Logging

generate.py

@@ -20,7 +20,8 @@ async def main(
     runs_per_prompt: int = 2,
     persona_names: Optional[List[str]] = None,
     verbose: bool = True,
-    folder_name: Optional[str] = None,
+    convo_folder_name: Optional[str] = None,
+    log_folder_name: Optional[str] = None,
     run_id: Optional[str] = None,
     max_concurrent: Optional[int] = None,
     max_total_words: Optional[int] = None,
@@ -39,7 +40,9 @@ async def main(
         runs_per_prompt: Number of runs per prompt
         persona_names: List of persona names to use. If None, uses all personas.
         verbose: Whether to print status messages
-        folder_name: Custom folder name for saving conversations. If None, uses
+        convo_folder_name: Custom folder name for saving conversations. If None, uses
+            default format.
+        log_folder_name: Custom folder name for saving logs. If None, uses
             default format.
         max_total_words: Optional maximum total words across all responses
         max_concurrent: Maximum number of concurrent conversations. If None, runs all
@@ -70,15 +73,19 @@ async def main(
         print(f"  - Max turns: {max_turns}")
         print(f"  - Runs per prompt: {runs_per_prompt}")
         print(f"  - Persona names: {persona_names}")
-        print(f"  - Folder name: {folder_name}")
+        print(f"  - Convo folder name: {convo_folder_name}")
+        print(f"  - Log folder name: {log_folder_name}")
         print(f"  - Run ID: {run_id}")
         print(f"  - Max concurrent: {max_concurrent}")
         print(f"  - Max total words: {max_total_words}")
         print(f"  - Max personas: {max_personas}")
 
     # Generate default folder name if not provided
-    if folder_name is None:
-        folder_name = "conversations"
+    if convo_folder_name is None:
+        convo_folder_name = "conversations"
+
+    if log_folder_name is None:
+        log_folder_name = "logging"
 
     if run_id is None:
         timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
@@ -94,17 +101,20 @@ async def main(
             f"p_{persona_info}__a_{agent_info}__t{max_turns}__"
             f"r{runs_per_prompt}__{timestamp}"
         )
-        folder_name = f"{folder_name}/{run_id}"
+        convo_folder_name = f"{convo_folder_name}/{run_id}"
+        log_folder_name = f"{log_folder_name}/{run_id}"
         # TODO: do we want to give a message if the folder already exists?
-        os.makedirs(folder_name, exist_ok=True)
+        os.makedirs(convo_folder_name, exist_ok=True)
+        os.makedirs(log_folder_name, exist_ok=True)
 
     # Configuration
     runner = ConversationRunner(
         persona_model_config=persona_model_config,
         agent_model_config=agent_model_config,
         max_turns=max_turns,
         runs_per_prompt=runs_per_prompt,
-        folder_name=folder_name,
+        convo_folder_name=convo_folder_name,
+        log_folder_name=log_folder_name,
         run_id=run_id,
         max_concurrent=max_concurrent,
         max_total_words=max_total_words,
@@ -115,7 +125,8 @@ async def main(
     results = await runner.run_conversations(persona_names=persona_names)
 
     if verbose:
-        print(f"✅ Generated {len(results)} conversations → {folder_name}/")
+        print(f"✅ Generated {len(results)} conversations → {convo_folder_name}/")
+        print(f"✅ Logs saved to {log_folder_name}/")
 
     return results
 
@@ -209,6 +220,13 @@ async def main(
         default="conversations",
     )
 
+    parser.add_argument(
+        "--log-folder-name",
+        "-lf",
+        help=("Folder name containing the logs for this run. " "Default is 'logging'."),
+        default="logging",
+    )
+
     parser.add_argument(
         "--max-concurrent",
         "-c",
@@ -272,7 +290,8 @@ async def main(
                 for k, v in agent_model_config.items()
                 if k not in ["model", "model_name", "name", "temperature", "max_tokens"]
             },
-            folder_name=args.folder_name,
+            convo_folder_name=args.folder_name,
+            log_folder_name=args.log_folder_name,
             max_concurrent=args.max_concurrent,
             max_total_words=args.max_total_words,
             max_personas=args.max_personas,

generate_conversations/runner.py

@@ -30,7 +30,8 @@ def __init__(
         run_id: str,
         max_turns: int = 6,
         runs_per_prompt: int = 3,
-        folder_name: str = "conversations",
+        convo_folder_name: str = "conversations",
+        log_folder_name: str = "logging",
         max_concurrent: Optional[int] = None,
         max_total_words: Optional[int] = None,
         max_personas: Optional[int] = None,
@@ -39,7 +40,8 @@ def __init__(
         self.agent_model_config = agent_model_config
         self.max_turns = max_turns
         self.runs_per_prompt = runs_per_prompt
-        self.folder_name = folder_name
+        self.convo_folder_name = convo_folder_name
+        self.log_folder_name = log_folder_name
         self.run_id = run_id
 
         # Limit concurrent conversations to avoid overwhelming the server
@@ -76,10 +78,14 @@ async def run_single_conversation(
         )
         persona_safe = persona_name.replace(" ", "_").replace(".", "")
         filename_base = f"{tag}_{persona_safe}_{model_short}_run{run_number}"
-        os.makedirs(f"{self.folder_name}", exist_ok=True)
+        os.makedirs(f"{self.convo_folder_name}", exist_ok=True)
 
         # Setup logging
-        logger = setup_conversation_logger(filename_base, run_id=self.run_id)
+        logger, log_file_path = setup_conversation_logger(
+            log_filename=filename_base,
+            run_id=self.run_id,
+            log_folder=self.log_folder_name,
+        )
         start_time = time.time()
 
         # Create LLM1 instance with the persona prompt and configuration
@@ -141,16 +147,16 @@ async def run_single_conversation(
         )
 
         # Save conversation file
-        simulator.save_conversation(f"{filename_base}.txt", self.folder_name)
+        simulator.save_conversation(f"{filename_base}.txt", self.convo_folder_name)
 
         result = {
             "id": conversation_id,
             "llm1_model": model_name,
             "llm1_prompt": persona_name,
             "run_number": run_number,
             "turns": len(conversation),
-            "filename": f"{self.folder_name}/{filename_base}.txt",
-            "log_file": f"{self.folder_name}/{filename_base}.log",
+            "filename": f"{self.convo_folder_name}/{filename_base}.txt",
+            "log_file": log_file_path,
             "duration": conversation_time,
             "early_termination": early_termination,
             "conversation": conversation,