langroid 0.45.6__tar.gz → 0.45.7__tar.gz

{langroid-0.45.6 → langroid-0.45.7}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: langroid
-Version: 0.45.6
+Version: 0.45.7
 Summary: Harness LLMs with Multi-Agent Programming
 Author-email: Prasad Chalasani <pchalasani@gmail.com>
 License: MIT

{langroid-0.45.6 → langroid-0.45.7}/langroid/embedding_models/models.py

@@ -30,6 +30,7 @@ class OpenAIEmbeddingsConfig(EmbeddingModelsConfig):
         # api_base is set from OPENAI_API_BASE env var, in .env or system env
         env_prefix = "OPENAI_"
 
+
 class AzureOpenAIEmbeddingsConfig(EmbeddingModelsConfig):
     model_type: str = "azure-openai"
     model_name: str = "text-embedding-3-small"

{langroid-0.45.6 → langroid-0.45.7}/langroid/language_models/base.py

@@ -42,6 +42,8 @@ FunctionCallTypes = Literal["none", "auto"]
 ToolChoiceTypes = Literal["none", "auto", "required"]
 ToolTypes = Literal["function"]
 
+DEFAULT_CONTEXT_LENGTH = 16_000
+
 
 class StreamEventType(Enum):
     TEXT = 1
@@ -66,9 +68,9 @@ class LLMConfig(BaseSettings):
     chat_model: str = ""
     completion_model: str = ""
     temperature: float = 0.0
-    chat_context_length: int = 8000
+    chat_context_length: int | None = None
     async_stream_quiet: bool = True  # suppress streaming output in async mode?
-    completion_context_length: int = 8000
+    completion_context_length: int | None = None
     # if input length + max_output_tokens > context length of model,
     # we will try shortening requested output
     min_output_tokens: int = 64
@@ -625,10 +627,10 @@ class LanguageModel(ABC):
         return get_model_info(orig_model, model)
 
     def chat_context_length(self) -> int:
-        return self.config.chat_context_length
+        return self.config.chat_context_length or DEFAULT_CONTEXT_LENGTH
 
     def completion_context_length(self) -> int:
-        return self.config.completion_context_length
+        return self.config.completion_context_length or DEFAULT_CONTEXT_LENGTH
 
     def chat_cost(self) -> Tuple[float, float]:
         return self.config.chat_cost_per_1k_tokens

{langroid-0.45.6 → langroid-0.45.7}/langroid/language_models/model_info.py

@@ -60,13 +60,13 @@ class DeepSeekModel(ModelName):
 class GeminiModel(ModelName):
     """Enum for Gemini models"""
 
-    GEMINI_1_5_FLASH = "gemini/gemini-1.5-flash"
-    GEMINI_1_5_FLASH_8B = "gemini/gemini-1.5-flash-8b"
-    GEMINI_1_5_PRO = "gemini/gemini-1.5-pro"
-    GEMINI_2_PRO = "gemini/gemini-2.0-pro-exp-02-05"
-    GEMINI_2_FLASH = "gemini/gemini-2.0-flash"
-    GEMINI_2_FLASH_LITE = "gemini/gemini-2.0-flash-lite-preview"
-    GEMINI_2_FLASH_THINKING = "gemini/gemini-2.0-flash-thinking-exp"
+    GEMINI_1_5_FLASH = "gemini-1.5-flash"
+    GEMINI_1_5_FLASH_8B = "gemini-1.5-flash-8b"
+    GEMINI_1_5_PRO = "gemini-1.5-pro"
+    GEMINI_2_PRO = "gemini-2.0-pro-exp-02-05"
+    GEMINI_2_FLASH = "gemini-2.0-flash"
+    GEMINI_2_FLASH_LITE = "gemini-2.0-flash-lite-preview"
+    GEMINI_2_FLASH_THINKING = "gemini-2.0-flash-thinking-exp"
 
 
 class OpenAI_API_ParamInfo(BaseModel):

{langroid-0.45.6 → langroid-0.45.7}/langroid/language_models/openai_gpt.py

@@ -660,17 +660,24 @@ class OpenAIGPT(LanguageModel):
 
     def chat_context_length(self) -> int:
         """
-        Context-length for chat-completion models/endpoints
-        Get it from the dict, otherwise fail-over to general method
+        Context-length for chat-completion models/endpoints.
+        Get it from the config if explicitly given,
+         otherwise use model_info based on model name, and fall back to
+         generic model_info if there's no match.
         """
-        return self.info().context_length
+        return self.config.chat_context_length or self.info().context_length
 
     def completion_context_length(self) -> int:
         """
-        Context-length for completion models/endpoints
-        Get it from the dict, otherwise fail-over to general method
+        Context-length for completion models/endpoints.
+        Get it from the config if explicitly given,
+         otherwise use model_info based on model name, and fall back to
+         generic model_info if there's no match.
         """
-        return self.completion_info().context_length
+        return (
+            self.config.completion_context_length
+            or self.completion_info().context_length
+        )
 
     def chat_cost(self) -> Tuple[float, float]:
         """

{langroid-0.45.6 → langroid-0.45.7}/langroid/parsing/web_search.py

@@ -51,8 +51,27 @@ class WebSearchResult:
 
     def get_full_content(self) -> str:
         try:
-            response: Response = requests.get(self.link)
-            soup: BeautifulSoup = BeautifulSoup(response.text, "lxml")
+            # First check headers only to get content length and type
+            head_response: Response = requests.head(self.link, timeout=5)
+            if head_response.status_code != 200:
+                return f"Error: HTTP {head_response.status_code} for {self.link}"
+
+            # Skip large files
+            content_length = int(head_response.headers.get("content-length", 0))
+            if content_length > 5_000_000:  # 5MB limit
+                return (
+                    f"Error: Content too large ({content_length} bytes) for {self.link}"
+                )
+
+            response: Response = requests.get(self.link, timeout=10)
+
+            import warnings
+
+            from bs4 import XMLParsedAsHTMLWarning
+
+            warnings.filterwarnings("ignore", category=XMLParsedAsHTMLWarning)
+
+            soup: BeautifulSoup = BeautifulSoup(response.text, "html.parser")
             text = " ".join(soup.stripped_strings)
             return text[: self.max_content_length]
         except Exception as e:

{langroid-0.45.6 → langroid-0.45.7}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "langroid"
-version = "0.45.6"
+version = "0.45.7"
 authors = [
     {name = "Prasad Chalasani", email = "pchalasani@gmail.com"},
 ]