langroid 0.45.10__py3-none-any.whl → 0.46.0__py3-none-any.whl

langroid/embedding_models/models.py

@@ -10,6 +10,7 @@ from openai import AzureOpenAI, OpenAI
 
 from langroid.embedding_models.base import EmbeddingModel, EmbeddingModelsConfig
 from langroid.exceptions import LangroidImportError
+from langroid.language_models.openai_gpt import LangDBParams
 from langroid.mytypes import Embeddings
 from langroid.parsing.utils import batched
 
@@ -24,6 +25,7 @@ class OpenAIEmbeddingsConfig(EmbeddingModelsConfig):
     organization: str = ""
     dims: int = 1536
     context_length: int = 8192
+    langdb_params: LangDBParams = LangDBParams()
 
     class Config:
         # enable auto-loading of env vars with OPENAI_ prefix, e.g.
@@ -136,11 +138,13 @@ class EmbeddingFunctionCallable:
         """
         embeds = []
         if isinstance(self.embed_model, (OpenAIEmbeddings, AzureOpenAIEmbeddings)):
-            tokenized_texts = self.embed_model.truncate_texts(input)
+            # Truncate texts to context length while preserving text format
+            truncated_texts = self.embed_model.truncate_texts(input)
 
-            for batch in batched(tokenized_texts, self.batch_size):
+            # Process in batches
+            for batch in batched(truncated_texts, self.batch_size):
                 result = self.embed_model.client.embeddings.create(
-                    input=batch, model=self.embed_model.config.model_name
+                    input=batch, model=self.embed_model.config.model_name  # type: ignore
                 )
                 batch_embeds = [d.embedding for d in result.data]
                 embeds.extend(batch_embeds)
@@ -183,30 +187,66 @@ class OpenAIEmbeddings(EmbeddingModel):
         super().__init__()
         self.config = config
         load_dotenv()
-        self.config.api_key = os.getenv("OPENAI_API_KEY", "")
+
+        # Check if using LangDB
+        self.is_langdb = self.config.model_name.startswith("langdb/")
+
+        if self.is_langdb:
+            self.config.model_name = self.config.model_name.replace("langdb/", "")
+            self.config.api_base = self.config.langdb_params.base_url
+            project_id = self.config.langdb_params.project_id
+            if project_id:
+                self.config.api_base += "/" + project_id + "/v1"
+            self.config.api_key = self.config.langdb_params.api_key
+
+        if not self.config.api_key:
+            self.config.api_key = os.getenv("OPENAI_API_KEY", "")
+
         self.config.organization = os.getenv("OPENAI_ORGANIZATION", "")
+
         if self.config.api_key == "":
-            raise ValueError(
-                """OPENAI_API_KEY env variable must be set to use 
-                OpenAIEmbeddings. Please set the OPENAI_API_KEY value 
-                in your .env file.
-                """
-            )
-        self.client = OpenAI(base_url=self.config.api_base, api_key=self.config.api_key)
+            if self.is_langdb:
+                raise ValueError(
+                    """
+                    LANGDB_API_KEY must be set in .env or your environment 
+                    to use OpenAIEmbeddings via LangDB.
+                    """
+                )
+            else:
+                raise ValueError(
+                    """
+                    OPENAI_API_KEY must be set in .env or your environment 
+                    to use OpenAIEmbeddings.
+                    """
+                )
+
+        self.client = OpenAI(
+            base_url=self.config.api_base,
+            api_key=self.config.api_key,
+            organization=self.config.organization,
+        )
+        model_for_tokenizer = self.config.model_name
+        if model_for_tokenizer.startswith("openai/"):
+            self.config.model_name = model_for_tokenizer.replace("openai/", "")
         self.tokenizer = tiktoken.encoding_for_model(self.config.model_name)
 
-    def truncate_texts(self, texts: List[str]) -> List[List[int]]:
+    def truncate_texts(self, texts: List[str]) -> List[str] | List[List[int]]:
         """
         Truncate texts to the embedding model's context length.
         TODO: Maybe we should show warning, and consider doing T5 summarization?
         """
-        return [
+        truncated_tokens = [
             self.tokenizer.encode(text, disallowed_special=())[
                 : self.config.context_length
             ]
             for text in texts
         ]
 
+        if self.is_langdb:
+            # LangDB embedding endpt only works with strings, not tokens
+            return [self.tokenizer.decode(tokens) for tokens in truncated_tokens]
+        return truncated_tokens
+
     def embedding_fn(self) -> Callable[[List[str]], Embeddings]:
         return EmbeddingFunctionCallable(self, self.config.batch_size)
 
@@ -256,7 +296,7 @@ class AzureOpenAIEmbeddings(EmbeddingModel):
         )
         self.tokenizer = tiktoken.encoding_for_model(self.config.model_name)
 
-    def truncate_texts(self, texts: List[str]) -> List[List[int]]:
+    def truncate_texts(self, texts: List[str]) -> List[str] | List[List[int]]:
         """
         Truncate texts to the embedding model's context length.
         TODO: Maybe we should show warning, and consider doing T5 summarization?

langroid/language_models/openai_gpt.py

@@ -66,7 +66,7 @@ from langroid.language_models.utils import (
     retry_with_exponential_backoff,
 )
 from langroid.parsing.parse_json import parse_imperfect_json
-from langroid.pydantic_v1 import BaseModel
+from langroid.pydantic_v1 import BaseModel, BaseSettings
 from langroid.utils.configuration import settings
 from langroid.utils.constants import Colors
 from langroid.utils.system import friendly_error
@@ -82,9 +82,13 @@ DEEPSEEK_BASE_URL = "https://api.deepseek.com/v1"
 OPENROUTER_BASE_URL = "https://openrouter.ai/api/v1"
 GEMINI_BASE_URL = "https://generativelanguage.googleapis.com/v1beta/openai"
 GLHF_BASE_URL = "https://glhf.chat/api/openai/v1"
+LANGDB_BASE_URL = "https://api.us-east-1.langdb.ai"
 OLLAMA_API_KEY = "ollama"
 DUMMY_API_KEY = "xxx"
 
+VLLM_API_KEY = os.environ.get("VLLM_API_KEY", DUMMY_API_KEY)
+LLAMACPP_API_KEY = os.environ.get("LLAMA_API_KEY", DUMMY_API_KEY)
+
 
 openai_chat_model_pref_list = [
     OpenAIChatModel.GPT4o,
@@ -177,6 +181,24 @@ def noop() -> None:
     return None
 
 
+class LangDBParams(BaseSettings):
+    """
+    Parameters specific to LangDB integration.
+    """
+
+    api_key: str = DUMMY_API_KEY
+    project_id: str = ""
+    label: Optional[str] = None
+    run_id: Optional[str] = None
+    thread_id: Optional[str] = None
+    base_url: str = LANGDB_BASE_URL
+
+    class Config:
+        # allow setting of fields via env vars,
+        # e.g. LANGDB_PROJECT_ID=1234
+        env_prefix = "LANGDB_"
+
+
 class OpenAICallParams(BaseModel):
     """
     Various params that can be sent to an OpenAI API chat-completion call.
@@ -253,6 +275,8 @@ class OpenAIGPTConfig(LLMConfig):
     # e.g. "mistral-instruct-v0.2 (a fuzzy search is done to find the closest match)
     formatter: str | None = None
     hf_formatter: HFFormatter | None = None
+    langdb_params: LangDBParams = LangDBParams()
+    headers: Dict[str, str] = {}
 
     def __init__(self, **kwargs) -> None:  # type: ignore
         local_model = "api_base" in kwargs and kwargs["api_base"] is not None
@@ -496,6 +520,7 @@ class OpenAIGPT(LanguageModel):
         self.is_deepseek = self.is_deepseek_model()
         self.is_glhf = self.config.chat_model.startswith("glhf/")
         self.is_openrouter = self.config.chat_model.startswith("openrouter/")
+        self.is_langdb = self.config.chat_model.startswith("langdb/")
 
         if self.is_groq:
             # use groq-specific client
@@ -544,18 +569,39 @@ class OpenAIGPT(LanguageModel):
                 self.api_base = DEEPSEEK_BASE_URL
                 if self.api_key == OPENAI_API_KEY:
                     self.api_key = os.getenv("DEEPSEEK_API_KEY", DUMMY_API_KEY)
+            elif self.is_langdb:
+                self.config.chat_model = self.config.chat_model.replace("langdb/", "")
+                self.api_base = self.config.langdb_params.base_url
+                project_id = self.config.langdb_params.project_id
+                if project_id:
+                    self.api_base += "/" + project_id + "/v1"
+                if self.api_key == OPENAI_API_KEY:
+                    self.api_key = self.config.langdb_params.api_key or DUMMY_API_KEY
+
+                if self.config.langdb_params:
+                    params = self.config.langdb_params
+                    if params.project_id:
+                        self.config.headers["x-project-id"] = params.project_id
+                    if params.label:
+                        self.config.headers["x-label"] = params.label
+                    if params.run_id:
+                        self.config.headers["x-run-id"] = params.run_id
+                    if params.thread_id:
+                        self.config.headers["x-thread-id"] = params.thread_id
 
             self.client = OpenAI(
                 api_key=self.api_key,
                 base_url=self.api_base,
                 organization=self.config.organization,
                 timeout=Timeout(self.config.timeout),
+                default_headers=self.config.headers,
             )
             self.async_client = AsyncOpenAI(
                 api_key=self.api_key,
                 organization=self.config.organization,
                 base_url=self.api_base,
                 timeout=Timeout(self.config.timeout),
+                default_headers=self.config.headers,
             )
 
         self.cache: CacheDB | None = None
@@ -1028,6 +1074,7 @@ class OpenAIGPT(LanguageModel):
                 OpenAIResponse object (with choices, usage)
 
         """
+
         completion = ""
         reasoning = ""
         function_args = ""
@@ -1075,7 +1122,9 @@ class OpenAIGPT(LanguageModel):
         )
 
     @staticmethod
-    def tool_deltas_to_tools(tools: List[Dict[str, Any]]) -> Tuple[
+    def tool_deltas_to_tools(
+        tools: List[Dict[str, Any]],
+    ) -> Tuple[
         str,
         List[OpenAIToolCall],
         List[Dict[str, Any]],

{langroid-0.45.10.dist-info → langroid-0.46.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: langroid
-Version: 0.45.10
+Version: 0.46.0
 Summary: Harness LLMs with Multi-Agent Programming
 Author-email: Prasad Chalasani <pchalasani@gmail.com>
 License: MIT

{langroid-0.45.10.dist-info → langroid-0.46.0.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=kYGGG-FkmGxNr0msXM3ANZ2eU-C85iTcoWNo00E9F_4,19151
+langroid/embedding_models/models.py,sha256=iGRrQR7ehDunA_7cPMu3CiHFugYWDkauOsiqHH-bv9s,20725
 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
@@ -72,7 +72,7 @@ langroid/language_models/base.py,sha256=mDYmFCBCLdq8_Uvws4MiewwEgcOCP8Qb0e5yUXr3
 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=tfBBxL0iUf2mVN6CjcvqflzFUVg2oZqOJZexZ8jHTYA,12216
-langroid/language_models/openai_gpt.py,sha256=FMi4rQsdJETof5eSQrxIaBdu-5cOHFg8fFy7Hx7twOQ,77691
+langroid/language_models/openai_gpt.py,sha256=Re4T1my9rhOPI-w4JCluhAZUVUIbW2AZJ3MIJMYjRuk,79633
 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
@@ -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-0.45.10.dist-info/METADATA,sha256=wprB9hS3tbbIaV9FvHE7VTrTZq0Xv6b0TLu3YG21k6A,63390
-langroid-0.45.10.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
-langroid-0.45.10.dist-info/licenses/LICENSE,sha256=EgVbvA6VSYgUlvC3RvPKehSg7MFaxWDsFuzLOsPPfJg,1065
-langroid-0.45.10.dist-info/RECORD,,
+langroid-0.46.0.dist-info/METADATA,sha256=S5uBIAjkQEV4KbQSZ2OH-YMoMfGIJtnulBdaYNouSpw,63389
+langroid-0.46.0.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
+langroid-0.46.0.dist-info/licenses/LICENSE,sha256=EgVbvA6VSYgUlvC3RvPKehSg7MFaxWDsFuzLOsPPfJg,1065
+langroid-0.46.0.dist-info/RECORD,,