langroid 0.45.5__py3-none-any.whl → 0.45.7__py3-none-any.whl

langroid/embedding_models/models.py

@@ -25,6 +25,11 @@ class OpenAIEmbeddingsConfig(EmbeddingModelsConfig):
     dims: int = 1536
     context_length: int = 8192
 
+    class Config:
+        # enable auto-loading of env vars with OPENAI_ prefix, e.g.
+        # 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"

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/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/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/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-0.45.5.dist-info → langroid-0.45.7.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: langroid
-Version: 0.45.5
+Version: 0.45.7
 Summary: Harness LLMs with Multi-Agent Programming
 Author-email: Prasad Chalasani <pchalasani@gmail.com>
 License: MIT

{langroid-0.45.5.dist-info → langroid-0.45.7.dist-info}/RECORD

@@ -59,7 +59,7 @@ langroid/cachedb/momento_cachedb.py,sha256=YEOJ62hEcV6iIeMr5aGgRYgWQqFYaej9gEDEc
 langroid/cachedb/redis_cachedb.py,sha256=7kgnbf4b5CKsCrlL97mHWKvdvlLt8zgn7lc528jEpiE,5141
 langroid/embedding_models/__init__.py,sha256=KyYxR3jDFUCfYjSuCL86qjAmrq6mXXjOT4lFNOKVj6Y,955
 langroid/embedding_models/base.py,sha256=Ml7oA6PzQm0wZmIYn3fhF7dvZCi-amviWUwOeBegH3A,2562
-langroid/embedding_models/models.py,sha256=93rQjvcltGLDYbgT1TIW8PZiAbNPAvPAqxCLlLS9GAo,18955
+langroid/embedding_models/models.py,sha256=kYGGG-FkmGxNr0msXM3ANZ2eU-C85iTcoWNo00E9F_4,19151
 langroid/embedding_models/remote_embeds.py,sha256=6_kjXByVbqhY9cGwl9R83ZcYC2km-nGieNNAo1McHaY,5151
 langroid/embedding_models/protoc/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 langroid/embedding_models/protoc/embeddings.proto,sha256=_O-SgFpTaylQeOTgSpxhEJ7CUw7PeCQQJLaPqpPYKJg,321
@@ -68,11 +68,11 @@ langroid/embedding_models/protoc/embeddings_pb2.pyi,sha256=UkNy7BrNsmQm0vLb3NtGX
 langroid/embedding_models/protoc/embeddings_pb2_grpc.py,sha256=9dYQqkW3JPyBpSEjeGXTNpSqAkC-6FPtBHyteVob2Y8,2452
 langroid/language_models/__init__.py,sha256=3aD2qC1lz8v12HX4B-dilv27gNxYdGdeu1QvDlkqqHs,1095
 langroid/language_models/azure_openai.py,sha256=SW0Fp_y6HpERr9l6TtF6CYsKgKwjUf_hSL_2mhTV4wI,5034
-langroid/language_models/base.py,sha256=is4l3x858tdPHbrJU2jxJXe2j9PCGb9kk_c5nyfShxs,26150
+langroid/language_models/base.py,sha256=mDYmFCBCLdq8_Uvws4MiewwEgcOCP8Qb0e5yUXr3zpQ,26249
 langroid/language_models/config.py,sha256=9Q8wk5a7RQr8LGMT_0WkpjY8S4ywK06SalVRjXlfCiI,378
 langroid/language_models/mock_lm.py,sha256=5BgHKDVRWFbUwDT_PFgTZXz9-k8wJSA2e3PZmyDgQ1k,4022
-langroid/language_models/model_info.py,sha256=_EidEMIgAMx0RuELAf5Ans0yiE1QllybZALw5o-1HJg,12265
-langroid/language_models/openai_gpt.py,sha256=lOQcExZO5Tja35Xi4F2HcG8pE-2LEnGrHwLTXLOOagk,77367
+langroid/language_models/model_info.py,sha256=tfBBxL0iUf2mVN6CjcvqflzFUVg2oZqOJZexZ8jHTYA,12216
+langroid/language_models/openai_gpt.py,sha256=FMi4rQsdJETof5eSQrxIaBdu-5cOHFg8fFy7Hx7twOQ,77691
 langroid/language_models/utils.py,sha256=L4_CbihDMTGcsg0TOG1Yd5JFEto46--h7CX_14m89sQ,5016
 langroid/language_models/prompt_formatter/__init__.py,sha256=2-5cdE24XoFDhifOLl8yiscohil1ogbP1ECkYdBlBsk,372
 langroid/language_models/prompt_formatter/base.py,sha256=eDS1sgRNZVnoajwV_ZIha6cba5Dt8xjgzdRbPITwx3Q,1221
@@ -94,7 +94,7 @@ langroid/parsing/table_loader.py,sha256=qNM4obT_0Y4tjrxNBCNUYjKQ9oETCZ7FbolKBTcz
 langroid/parsing/url_loader.py,sha256=obi_kj6ehBkdh5mXNtYCXpm3KCuExoy2D1ODVlFbXbQ,4895
 langroid/parsing/urls.py,sha256=Tjzr64YsCusiYkY0LEGB5-rSuX8T2P_4DVoOFKAeKuI,8081
 langroid/parsing/utils.py,sha256=WwqzOhbQRlorbVvddDIZKv9b1KqZCBDm955lgIHDXRw,12828
-langroid/parsing/web_search.py,sha256=wWSmV0METFTGPhHJIs-M4tog2Aur_75Pxr4a49cKDkU,7042
+langroid/parsing/web_search.py,sha256=rGkeUGsv_nopRIRuB-OqbieGASuHUaEFWsXmVdBXi8g,7800
 langroid/prompts/__init__.py,sha256=RW11vK6jiLPuaUh4GpeFvstti73gkm8_rDMtrbo2YsU,142
 langroid/prompts/dialog.py,sha256=SpfiSyofSgy2pwD1YboHR_yHO3LEEMbv6j2sm874jKo,331
 langroid/prompts/prompts_config.py,sha256=p_lp9nbMuQwhhMwAZsOxveRw9C0ZFZvql7pdIfgVZYo,143
@@ -127,7 +127,7 @@ langroid/vector_store/pineconedb.py,sha256=otxXZNaBKb9f_H75HTaU3lMHiaR2NUp5MqwLZ
 langroid/vector_store/postgres.py,sha256=wHPtIi2qM4fhO4pMQr95pz1ZCe7dTb2hxl4VYspGZoA,16104
 langroid/vector_store/qdrantdb.py,sha256=O6dSBoDZ0jzfeVBd7LLvsXu083xs2fxXtPa9gGX3JX4,18443
 langroid/vector_store/weaviatedb.py,sha256=Yn8pg139gOy3zkaPfoTbMXEEBCiLiYa1MU5d_3UA1K4,11847
-langroid-0.45.5.dist-info/METADATA,sha256=aQ1z8Qu1fZty_2w2Xj9fGCptUUNSPUCEFcTp4xVrats,63335
-langroid-0.45.5.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
-langroid-0.45.5.dist-info/licenses/LICENSE,sha256=EgVbvA6VSYgUlvC3RvPKehSg7MFaxWDsFuzLOsPPfJg,1065
-langroid-0.45.5.dist-info/RECORD,,
+langroid-0.45.7.dist-info/METADATA,sha256=iuFtKxQjQwkRMb8-5lLmVKETJMCCxRcRO0REeJyW9dw,63335
+langroid-0.45.7.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
+langroid-0.45.7.dist-info/licenses/LICENSE,sha256=EgVbvA6VSYgUlvC3RvPKehSg7MFaxWDsFuzLOsPPfJg,1065
+langroid-0.45.7.dist-info/RECORD,,