langroid 0.16.0__tar.gz → 0.16.1__tar.gz

{langroid-0.16.0 → langroid-0.16.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: langroid
-Version: 0.16.0
+Version: 0.16.1
 Summary: Harness LLMs with Multi-Agent Programming
 License: MIT
 Author: Prasad Chalasani
@@ -242,6 +242,11 @@ teacher_task.run()
 <details>
 <summary> <b>Click to expand</b></summary>
 
+- **Sep 2024:**
+  - **[0.16.0](https://github.com/langroid/langroid/releases/tag/0.16.0)**  Support for OpenAI `o1-mini` and `o1-preview` models.
+  - **[0.15.0](https://github.com/langroid/langroid/releases/tag/0.15.0)** Cerebras API support -- run llama-3.1 models hosted on Cerebras Cloud (very fast inference).
+  - **[0.14.0](https://github.com/langroid/langroid/releases/tag/0.14.0)** `DocChatAgent` uses Reciprocal Rank Fusion (RRF) to rank chunks retrieved by different methods.
+  - **[0.12.0](https://github.com/langroid/langroid/releases/tag/0.12.0)** `run_batch_task` new option -- `stop_on_first_result` - allows termination of batch as soon as any task returns a result.  
 - **Aug 2024:**
   - **[0.11.0](https://github.com/langroid/langroid/releases/tag/0.11.0)** Polymorphic `Task.run(), Task.run_async`.
   - **[0.10.0](https://github.com/langroid/langroid/releases/tag/0.10.0)** Allow tool handlers to return arbitrary result type, including other tools.

{langroid-0.16.0 → langroid-0.16.1}/README.md

@@ -134,6 +134,11 @@ teacher_task.run()
 <details>
 <summary> <b>Click to expand</b></summary>
 
+- **Sep 2024:**
+  - **[0.16.0](https://github.com/langroid/langroid/releases/tag/0.16.0)**  Support for OpenAI `o1-mini` and `o1-preview` models.
+  - **[0.15.0](https://github.com/langroid/langroid/releases/tag/0.15.0)** Cerebras API support -- run llama-3.1 models hosted on Cerebras Cloud (very fast inference).
+  - **[0.14.0](https://github.com/langroid/langroid/releases/tag/0.14.0)** `DocChatAgent` uses Reciprocal Rank Fusion (RRF) to rank chunks retrieved by different methods.
+  - **[0.12.0](https://github.com/langroid/langroid/releases/tag/0.12.0)** `run_batch_task` new option -- `stop_on_first_result` - allows termination of batch as soon as any task returns a result.  
 - **Aug 2024:**
   - **[0.11.0](https://github.com/langroid/langroid/releases/tag/0.11.0)** Polymorphic `Task.run(), Task.run_async`.
   - **[0.10.0](https://github.com/langroid/langroid/releases/tag/0.10.0)** Allow tool handlers to return arbitrary result type, including other tools.

{langroid-0.16.0 → langroid-0.16.1}/langroid/language_models/openai_gpt.py

@@ -606,6 +606,17 @@ class OpenAIGPT(LanguageModel):
             case _:
                 return []
 
+    def rename_params(self) -> Dict[str, str]:
+        """
+        Map of param name -> new name for specific models.
+        Currently main troublemaker is o1* series.
+        """
+        match self.config.chat_model:
+            case OpenAIChatModel.O1_MINI | OpenAIChatModel.O1_PREVIEW:
+                return {"max_tokens": "max_completion_tokens"}
+            case _:
+                return {}
+
     def chat_context_length(self) -> int:
         """
         Context-length for chat-completion models/endpoints
@@ -1186,7 +1197,7 @@ class OpenAIGPT(LanguageModel):
             kwargs["prompt"] = prompt
         args = dict(
             **kwargs,
-            max_completion_tokens=max_tokens,  # for output/completion
+            max_tokens=max_tokens,  # for output/completion
             stream=self.get_stream(),
         )
         args = self._openai_api_call_params(args)
@@ -1261,7 +1272,7 @@ class OpenAIGPT(LanguageModel):
             kwargs["prompt"] = prompt
         cached, hashed_key, response = await completions_with_backoff(
             **kwargs,
-            max_completion_tokens=max_tokens,
+            max_tokens=max_tokens,
             stream=False,
         )
         if not isinstance(response, dict):
@@ -1468,7 +1479,7 @@ class OpenAIGPT(LanguageModel):
         args: Dict[str, Any] = dict(
             model=chat_model,
             messages=[m.api_dict() for m in llm_messages],
-            max_completion_tokens=max_tokens,
+            max_tokens=max_tokens,
             stream=self.get_stream(),
         )
         args.update(self._openai_api_call_params(args))
@@ -1498,6 +1509,11 @@ class OpenAIGPT(LanguageModel):
             # some models e.g. o1-mini (as of sep 2024) don't support some params,
             # like temperature and stream, so we need to remove them.
             args.pop(p, None)
+
+        param_rename_map = self.rename_params()
+        for old_param, new_param in param_rename_map.items():
+            if old_param in args:
+                args[new_param] = args.pop(old_param)
         return args
 
     def _process_chat_completion_response(

{langroid-0.16.0 → langroid-0.16.1}/pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "langroid"
-version = "0.16.0"
+version = "0.16.1"
 description = "Harness LLMs with Multi-Agent Programming"
 authors = ["Prasad Chalasani <pchalasani@gmail.com>"]
 readme = "README.md"