langroid 0.2.12__py3-none-any.whl → 0.3.1__py3-none-any.whl

langroid/embedding_models/base.py

@@ -24,6 +24,8 @@ class EmbeddingModel(ABC):
     @classmethod
     def create(cls, config: EmbeddingModelsConfig) -> "EmbeddingModel":
         from langroid.embedding_models.models import (
+            FastEmbedEmbeddings,
+            FastEmbedEmbeddingsConfig,
             OpenAIEmbeddings,
             OpenAIEmbeddingsConfig,
             SentenceTransformerEmbeddings,
@@ -40,6 +42,8 @@ class EmbeddingModel(ABC):
             return OpenAIEmbeddings(config)
         elif isinstance(config, SentenceTransformerEmbeddingsConfig):
             return SentenceTransformerEmbeddings(config)
+        elif isinstance(config, FastEmbedEmbeddingsConfig):
+            return FastEmbedEmbeddings(config)
         else:
             raise ValueError(f"Unknown embedding config: {config.__repr_name__}")
 

langroid/embedding_models/models.py

@@ -1,12 +1,14 @@
 import atexit
 import os
-from typing import Callable, List, Optional
+from functools import cached_property
+from typing import Any, Callable, Dict, List, Optional
 
 import tiktoken
 from dotenv import load_dotenv
 from openai import OpenAI
 
 from langroid.embedding_models.base import EmbeddingModel, EmbeddingModelsConfig
+from langroid.exceptions import LangroidImportError
 from langroid.mytypes import Embeddings
 from langroid.parsing.utils import batched
 
@@ -32,6 +34,20 @@ class SentenceTransformerEmbeddingsConfig(EmbeddingModelsConfig):
     devices: Optional[list[str]] = None
 
 
+class FastEmbedEmbeddingsConfig(EmbeddingModelsConfig):
+    """Config for qdrant/fastembed embeddings,
+    see here: https://github.com/qdrant/fastembed
+    """
+
+    model_type: str = "fastembed"
+    model_name: str = "BAAI/bge-small-en-v1.5"
+    batch_size: int = 256
+    cache_dir: Optional[str] = None
+    threads: Optional[int] = None
+    parallel: Optional[int] = None
+    additional_kwargs: Dict[str, Any] = {}
+
+
 class EmbeddingFunctionCallable:
     """
     A callable class designed to generate embeddings for a list of texts using
@@ -189,6 +205,41 @@ class SentenceTransformerEmbeddings(EmbeddingModel):
         return dims  # type: ignore
 
 
+class FastEmbedEmbeddings(EmbeddingModel):
+    def __init__(self, config: FastEmbedEmbeddingsConfig = FastEmbedEmbeddingsConfig()):
+        try:
+            from fastembed import TextEmbedding
+        except ImportError:
+            raise LangroidImportError("fastembed", extra="fastembed")
+
+        super().__init__()
+        self.config = config
+        self._batch_size = config.batch_size
+        self._parallel = config.parallel
+
+        self._model = TextEmbedding(
+            model_name=self.config.model_name,
+            cache_dir=self.config.cache_dir,
+            threads=self.config.threads,
+            **self.config.additional_kwargs,
+        )
+
+    def embedding_fn(self) -> Callable[[List[str]], Embeddings]:
+        def fn(texts: List[str]) -> Embeddings:
+            embeddings = self._model.embed(
+                texts, batch_size=self._batch_size, parallel=self._parallel
+            )
+
+            return [embedding.tolist() for embedding in embeddings]
+
+        return fn
+
+    @cached_property
+    def embedding_dims(self) -> int:
+        embed_func = self.embedding_fn()
+        return len(embed_func(["text"])[0])
+
+
 def embedding_model(embedding_fn_type: str = "openai") -> EmbeddingModel:
     """
     Args:
@@ -198,5 +249,7 @@ def embedding_model(embedding_fn_type: str = "openai") -> EmbeddingModel:
     """
     if embedding_fn_type == "openai":
         return OpenAIEmbeddings  # type: ignore
+    elif embedding_fn_type == "fastembed":
+        return FastEmbedEmbeddings  # type: ignore
     else:  # default sentence transformer
         return SentenceTransformerEmbeddings  # type: ignore

langroid/language_models/azure_openai.py

@@ -133,12 +133,15 @@ class AzureGPT(OpenAIGPT):
         """
         Handles the setting of the GPT-4 model in the configuration.
         This function checks the `model_version` in the configuration.
-        If the version is not set, it raises a ValueError indicating that the model
-        version needs to be specified in the ``.env`` file.
-        It sets `OpenAIChatModel.GPT4_TURBO` if the version is
-        '1106-Preview', otherwise, it defaults to setting `OpenAIChatModel.GPT4`.
+        If the version is not set, it raises a ValueError indicating
+        that the model version needs to be specified in the ``.env``
+        file.  It sets `OpenAIChatMode.GPT4o` if the version is
+        '2024-05-13', `OpenAIChatModel.GPT4_TURBO` if the version is
+        '1106-Preview', otherwise, it defaults to setting
+        `OpenAIChatModel.GPT4`.
         """
         VERSION_1106_PREVIEW = "1106-Preview"
+        VERSION_GPT4o = "2024-05-13"
 
         if self.config.model_version == "":
             raise ValueError(
@@ -146,7 +149,9 @@ class AzureGPT(OpenAIGPT):
                 "Please set it to the chat model version used in your deployment."
             )
 
-        if self.config.model_version == VERSION_1106_PREVIEW:
+        if self.config.model_version == VERSION_GPT4o:
+            self.config.chat_model = OpenAIChatModel.GPT4o
+        elif self.config.model_version == VERSION_1106_PREVIEW:
             self.config.chat_model = OpenAIChatModel.GPT4_TURBO
         else:
             self.config.chat_model = OpenAIChatModel.GPT4

langroid/language_models/base.py

@@ -234,6 +234,7 @@ class LLMResponse(BaseModel):
             # in this case we ignore message, since all information is in function_call
             msg = ""
             args = self.function_call.arguments
+            recipient = ""
             if isinstance(args, dict):
                 recipient = args.get("recipient", "")
             return recipient, msg

langroid/language_models/openai_gpt.py

@@ -1,4 +1,3 @@
-import ast
 import hashlib
 import json
 import logging
@@ -49,6 +48,7 @@ from langroid.language_models.utils import (
     async_retry_with_exponential_backoff,
     retry_with_exponential_backoff,
 )
+from langroid.parsing.parse_json import parse_imperfect_json
 from langroid.pydantic_v1 import BaseModel
 from langroid.utils.configuration import settings
 from langroid.utils.constants import Colors
@@ -797,11 +797,24 @@ class OpenAIGPT(LanguageModel):
         args = {}
         if has_function and function_args != "":
             try:
-                args = ast.literal_eval(function_args.strip())
-            except (SyntaxError, ValueError):
+                stripped_fn_args = function_args.strip()
+                dict_or_list = parse_imperfect_json(stripped_fn_args)
+                if not isinstance(dict_or_list, dict):
+                    raise ValueError(
+                        f"""
+                        Invalid function args: {stripped_fn_args} 
+                        parsed as {dict_or_list},
+                        which is not a valid dict.
+                        """
+                    )
+                args = dict_or_list
+            except (SyntaxError, ValueError) as e:
                 logging.warning(
-                    f"Parsing OpenAI function args failed: {function_args};"
-                    " treating args as normal message"
+                    f"""
+                    Parsing OpenAI function args failed: {function_args};
+                    treating args as normal message. Error detail:
+                    {e}
+                    """
                 )
                 has_function = False
                 completion = completion + function_args

langroid/parsing/parse_json.py

@@ -1,5 +1,6 @@
+import ast
 import json
-from typing import Any, Iterator, List
+from typing import Any, Dict, Iterator, List, Union
 
 import yaml
 from pyparsing import nestedExpr, originalTextFor
@@ -73,6 +74,31 @@ def add_quotes(s: str) -> str:
         return s
 
 
+def parse_imperfect_json(json_string: str) -> Union[Dict[str, Any], List[Any]]:
+    if not json_string.strip():
+        raise ValueError("Empty string is not valid JSON")
+
+    # First, try parsing with ast.literal_eval
+    try:
+        result = ast.literal_eval(json_string)
+        if isinstance(result, (dict, list)):
+            return result
+    except (ValueError, SyntaxError):
+        pass
+
+    # If ast.literal_eval fails or returns non-dict/list, try json.loads
+    try:
+        str = add_quotes(json_string)
+        result = json.loads(str)
+        if isinstance(result, (dict, list)):
+            return result
+    except json.JSONDecodeError:
+        pass
+
+    # If all methods fail, raise ValueError
+    raise ValueError(f"Unable to parse as JSON: {json_string}")
+
+
 def repair_newlines(s: str) -> str:
     """
     Attempt to load as json, and if it fails, try with newlines replaced by space.

langroid/vector_store/momento.py

@@ -1,6 +1,7 @@
 """
 Momento Vector Index.
 https://docs.momentohq.com/vector-index/develop/api-reference
+DEPRECATED: API is unstable.
 """
 
 from __future__ import annotations

langroid/vector_store/qdrantdb.py

@@ -63,6 +63,7 @@ def is_valid_uuid(uuid_to_test: str) -> bool:
 
 class QdrantDBConfig(VectorStoreConfig):
     cloud: bool = True
+    docker: bool = False
     collection_name: str | None = "temp"
     storage_path: str = ".qdrant/data"
     embedding: EmbeddingModelsConfig = OpenAIEmbeddingsConfig()
@@ -102,7 +103,19 @@ class QdrantDB(VectorStore):
         load_dotenv()
         key = os.getenv("QDRANT_API_KEY")
         url = os.getenv("QDRANT_API_URL")
-        if config.cloud and None in [key, url]:
+        if config.docker:
+            if url is None:
+                logger.warning(
+                    f"""The QDRANT_API_URL env variable must be set to use
+                    QdrantDB in local docker mode. Please set this
+                    value in your .env file.
+                    Switching to local storage at {config.storage_path}
+                    """
+                )
+                config.cloud = False
+            else:
+                config.cloud = True
+        elif config.cloud and None in [key, url]:
             logger.warning(
                 f"""QDRANT_API_KEY, QDRANT_API_URL env variable must be set to use 
                 QdrantDB in cloud mode. Please set these values 
@@ -111,6 +124,7 @@ class QdrantDB(VectorStore):
                 """
             )
             config.cloud = False
+
         if config.cloud:
             self.client = QdrantClient(
                 url=url,
@@ -385,7 +399,7 @@ class QdrantDB(VectorStore):
         # Note the records may NOT be in the order of the ids,
         # so we re-order them here.
         id2payload = {record.id: record.payload for record in records}
-        ordered_payloads = [id2payload[id] for id in _ids]
+        ordered_payloads = [id2payload[id] for id in _ids if id in id2payload]
         docs = [Document(**payload) for payload in ordered_payloads]  # type: ignore
         return docs
 

{langroid-0.2.12.dist-info → langroid-0.3.1.dist-info}/METADATA

@@ -1,11 +1,11 @@
 Metadata-Version: 2.1
 Name: langroid
-Version: 0.2.12
+Version: 0.3.1
 Summary: Harness LLMs with Multi-Agent Programming
 License: MIT
 Author: Prasad Chalasani
 Author-email: pchalasani@gmail.com
-Requires-Python: >=3.10,<4.0
+Requires-Python: >=3.10,<3.13
 Classifier: License :: OSI Approved :: MIT License
 Classifier: Programming Language :: Python :: 3
 Classifier: Programming Language :: Python :: 3.10
@@ -17,6 +17,7 @@ Provides-Extra: chromadb
 Provides-Extra: db
 Provides-Extra: doc-chat
 Provides-Extra: docx
+Provides-Extra: fastembed
 Provides-Extra: hf-embeddings
 Provides-Extra: hf-transformers
 Provides-Extra: lancedb
@@ -44,6 +45,7 @@ Requires-Dist: docstring-parser (>=0.15,<0.16)
 Requires-Dist: duckduckgo-search (>=6.0.0,<7.0.0)
 Requires-Dist: faker (>=18.9.0,<19.0.0)
 Requires-Dist: fakeredis (>=2.12.1,<3.0.0)
+Requires-Dist: fastembed (>=0.3.1,<0.4.0) ; extra == "all" or extra == "fastembed"
 Requires-Dist: fire (>=0.5.0,<0.6.0)
 Requires-Dist: google-api-python-client (>=2.95.0,<3.0.0)
 Requires-Dist: google-generativeai (>=0.5.2,<0.6.0)
@@ -57,7 +59,7 @@ Requires-Dist: litellm (>=1.30.1,<2.0.0) ; extra == "all" or extra == "litellm"
 Requires-Dist: lxml (>=4.9.3,<5.0.0)
 Requires-Dist: meilisearch-python-sdk (>=2.2.3,<3.0.0) ; extra == "meilisearch"
 Requires-Dist: metaphor-python (>=0.1.23,<0.2.0) ; extra == "all" or extra == "metaphor"
-Requires-Dist: momento (>=1.10.2,<2.0.0) ; extra == "momento"
+Requires-Dist: momento (>=1.10.2,<1.21) ; extra == "momento"
 Requires-Dist: neo4j (>=5.14.1,<6.0.0) ; extra == "all" or extra == "neo4j"
 Requires-Dist: nest-asyncio (>=1.6.0,<2.0.0)
 Requires-Dist: nltk (>=3.8.1,<4.0.0)
@@ -233,6 +235,8 @@ teacher_task.run()
 <details>
 <summary> <b>Click to expand</b></summary>
 
+- **Jul 2024:**
+  - **[0.3.0](https://github.com/langroid/langroid/releases/tag/0.3.0)**: Added [FastEmbed](https://qdrant.github.io/fastembed/qdrant/Usage_With_Qdrant/) embeddings from Qdrant
 - **Jun 2024:**
   - **0.2.0:** Improved lineage tracking, granular sub-task configs, and a new tool, `RewindTool`, 
     that lets an agent "rewind and redo" a past message (and all dependent messages are cleared out 

{langroid-0.2.12.dist-info → langroid-0.3.1.dist-info}/RECORD

@@ -33,6 +33,7 @@ langroid/agent/special/sql/utils/system_message.py,sha256=qKLHkvQWRQodTtPLPxr1GS
 langroid/agent/special/sql/utils/tools.py,sha256=vFYysk6Vi7HJjII8B4RitA3pt_z3gkSglDNdhNVMiFc,1332
 langroid/agent/special/table_chat_agent.py,sha256=d9v2wsblaRx7oMnKhLV7uO_ujvk9gh59pSGvBXyeyNc,9659
 langroid/agent/task.py,sha256=vKM2dmRYSH4i_VA0lf2axUtZcTGU44rVHz6EyxI4kG0,73990
+langroid/agent/team.py,sha256=88VNRSmK35WEl620GfBzuIrBASXYSeBZ8yDKX-nP_Bo,75778
 langroid/agent/tool_message.py,sha256=wIyZnUcZpxkiRPvM9O3MO3b5BBAdLEEan9kqPbvtApc,9743
 langroid/agent/tools/__init__.py,sha256=e-63cfwQNk_ftRKQwgDAJQK16QLbRVWDBILeXIc7wLk,402
 langroid/agent/tools/duckduckgo_search_tool.py,sha256=NhsCaGZkdv28nja7yveAhSK_w6l_Ftym8agbrdzqgfo,1935
@@ -52,9 +53,9 @@ langroid/cachedb/base.py,sha256=ztVjB1DtN6pLCujCWnR6xruHxwVj3XkYniRTYAKKqk0,1354
 langroid/cachedb/momento_cachedb.py,sha256=YEOJ62hEcV6iIeMr5aGgRYgWQqFYaej9gEDEcY0sm7M,3172
 langroid/cachedb/redis_cachedb.py,sha256=7kgnbf4b5CKsCrlL97mHWKvdvlLt8zgn7lc528jEpiE,5141
 langroid/embedding_models/__init__.py,sha256=lsu8qxCjfGujXGueJWU-VI3LMZYGjLSYgqUKDd4F3Qo,715
-langroid/embedding_models/base.py,sha256=MSjaTkFcfoMGY6SHPOqAsbZbKctj8-1N6zgaFYmOFTg,1830
+langroid/embedding_models/base.py,sha256=3dK0nW3XNjK3Vyh2kxhIffzDuUYumVVkCIimB3UPHeU,2009
 langroid/embedding_models/clustering.py,sha256=tZWElUqXl9Etqla0FAa7og96iDKgjqWjucZR_Egtp-A,6684
-langroid/embedding_models/models.py,sha256=-xeN0irBPc1tUgRFHGM1ki4NwOIHr6F3SKuEjD5nTOg,7144
+langroid/embedding_models/models.py,sha256=NQ1Cfw5MmeR69fzthIWjnkSAQuWcxnUecS5z_U2IPNs,8900
 langroid/embedding_models/protoc/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 langroid/embedding_models/protoc/embeddings.proto,sha256=_O-SgFpTaylQeOTgSpxhEJ7CUw7PeCQQJLaPqpPYKJg,321
 langroid/embedding_models/protoc/embeddings_pb2.py,sha256=4Q57PhOunv-uZNJrxYrWBXAI0ZtfnVZXFRhRj5JuRSg,1662
@@ -63,11 +64,11 @@ langroid/embedding_models/protoc/embeddings_pb2_grpc.py,sha256=9dYQqkW3JPyBpSEje
 langroid/embedding_models/remote_embeds.py,sha256=6_kjXByVbqhY9cGwl9R83ZcYC2km-nGieNNAo1McHaY,5151
 langroid/exceptions.py,sha256=w_Cr41nPAmsa6gW5nNFaO9yDcBCWdQqRspL1jYvZf5w,2209
 langroid/language_models/__init__.py,sha256=1sUGobooTqq77XC7LxKsvME0RgSd5GGmeyrPo9SMh4U,940
-langroid/language_models/azure_openai.py,sha256=ncRCbKooqLVOY-PWQUIo9C3yTuKEFbAwyngXT_M4P7k,5989
-langroid/language_models/base.py,sha256=oAK2lXBqksMglqWqE2CjC03X3qPFXWgtjFWpH9hJ3C8,17500
+langroid/language_models/azure_openai.py,sha256=G4le3j4YLHV7IwgB2C37hO3MKijZ1KjynbYlEvpIF7Y,6214
+langroid/language_models/base.py,sha256=nhY-AdSkfqaW4hzeIekxxZs29AWLd7X7GYhRygU9L74,17527
 langroid/language_models/config.py,sha256=9Q8wk5a7RQr8LGMT_0WkpjY8S4ywK06SalVRjXlfCiI,378
 langroid/language_models/mock_lm.py,sha256=qdgj-wtbQBXlibo_0rIRfCt0hGTPRoxy1C4VjN6quI4,2707
-langroid/language_models/openai_gpt.py,sha256=84oXRT6hyY_MiavmRw61Jtw67xCmzBQEt5NKuqsxnQo,51680
+langroid/language_models/openai_gpt.py,sha256=JOgENlOGBTg9r94hjvuqB2OteHRbX2JtMkrApoNu-jc,52257
 langroid/language_models/prompt_formatter/__init__.py,sha256=2-5cdE24XoFDhifOLl8yiscohil1ogbP1ECkYdBlBsk,372
 langroid/language_models/prompt_formatter/base.py,sha256=eDS1sgRNZVnoajwV_ZIha6cba5Dt8xjgzdRbPITwx3Q,1221
 langroid/language_models/prompt_formatter/hf_formatter.py,sha256=TFL6ppmeQWnzr6CKQzRZFYY810zE1mr8DZnhw6i85ok,5217
@@ -82,7 +83,7 @@ langroid/parsing/config.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 langroid/parsing/document_parser.py,sha256=WGnA5ADwMHliGJt6WW9rc4RiFXQcKU33b5zdPiGrtEY,24265
 langroid/parsing/image_text.py,sha256=sbLIQ5nHe2UnYUksBaQsmZGaX-X0qgEpPd7CEzi_z5M,910
 langroid/parsing/para_sentence_split.py,sha256=AJBzZojP3zpB-_IMiiHismhqcvkrVBQ3ZINoQyx_bE4,2000
-langroid/parsing/parse_json.py,sha256=tgB_oatcrgt6L9ZplC-xBBXjLzL1gjSQf1L2_W5kwFA,4230
+langroid/parsing/parse_json.py,sha256=Rcy0iCyMMICk2QstrlJtW7zDPs9lOZ3N_mlaJGWkC40,5033
 langroid/parsing/parser.py,sha256=AgtmlVUvrkSG1l7-YZPX8rlldgXjh_HqXAMqpXkBxUo,11746
 langroid/parsing/repo_loader.py,sha256=3GjvPJS6Vf5L6gV2zOU8s-Tf1oq_fZm-IB_RL_7CTsY,29373
 langroid/parsing/routing.py,sha256=-FcnlqldzL4ZoxuDwXjQPNHgBe9F9-F4R6q7b_z9CvI,1232
@@ -126,11 +127,11 @@ langroid/vector_store/base.py,sha256=tuEPaxJcuU_39sRnUjjNd8D8n8IjP6jrbwQv_ecNpSw
 langroid/vector_store/chromadb.py,sha256=bZ5HjwgKgfJj1PUHsatYsrHv-v0dpOfMR2l0tJ2H0_A,7890
 langroid/vector_store/lancedb.py,sha256=9x7e_5zo7nLhMbhjYby2ZpBJ-vyawcC0_XAuatfHJf8,20517
 langroid/vector_store/meilisearch.py,sha256=6frB7GFWeWmeKzRfLZIvzRjllniZ1cYj3HmhHQICXLs,11663
-langroid/vector_store/momento.py,sha256=QaPzUnTwlswoawGB-paLtUPyLRvckFXLfLDfvbTzjNQ,10505
+langroid/vector_store/momento.py,sha256=qR-zBF1RKVHQZPZQYW_7g-XpTwr46p8HJuYPCkfJbM4,10534
 langroid/vector_store/qdrant_cloud.py,sha256=3im4Mip0QXLkR6wiqVsjV1QvhSElfxdFSuDKddBDQ-4,188
-langroid/vector_store/qdrantdb.py,sha256=wYOuu5c2vIKn9ZgvTXcAiZXMpV8AOXEWFAzI8S8UP-0,16828
-pyproject.toml,sha256=IQC1A8fpg0cKWje_Gt3YxZpRQw9ie1xlJ8TaiS4Vr3w,6960
-langroid-0.2.12.dist-info/LICENSE,sha256=EgVbvA6VSYgUlvC3RvPKehSg7MFaxWDsFuzLOsPPfJg,1065
-langroid-0.2.12.dist-info/METADATA,sha256=66JdF6SmaH5yisjOpT6dgYsjXAG7FbnD8OaTAmVnV7I,54101
-langroid-0.2.12.dist-info/WHEEL,sha256=FMvqSimYX_P7y0a7UY-_Mc83r5zkBZsCYPm7Lr0Bsq4,88
-langroid-0.2.12.dist-info/RECORD,,
+langroid/vector_store/qdrantdb.py,sha256=HkcK6jOf-FEDoOiG94MpsYDJr98T7vZkDyG__1BlnWI,17354
+pyproject.toml,sha256=x0YGXi9ennkubMYlFO-Eeyp6h2YE_aOBbeRJrUtTm34,7063
+langroid-0.3.1.dist-info/LICENSE,sha256=EgVbvA6VSYgUlvC3RvPKehSg7MFaxWDsFuzLOsPPfJg,1065
+langroid-0.3.1.dist-info/METADATA,sha256=9WLpuCfOtRfjB30PZa2jwGmnlotxXRZgHqt6UWiNh4E,54402
+langroid-0.3.1.dist-info/WHEEL,sha256=FMvqSimYX_P7y0a7UY-_Mc83r5zkBZsCYPm7Lr0Bsq4,88
+langroid-0.3.1.dist-info/RECORD,,

pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "langroid"
-version = "0.2.12"
+version = "0.3.1"
 description = "Harness LLMs with Multi-Agent Programming"
 authors = ["Prasad Chalasani <pchalasani@gmail.com>"]
 readme = "README.md"
@@ -10,11 +10,11 @@ include = ["pyproject.toml"]
 
 # =============== MAIN DEPS ==============
 [tool.poetry.dependencies]
-python = "^3.10"
+python = ">=3.10,<3.13"
 
 # =========== OPTIONALS ==============================
 chromadb = {version=">=0.4.21, <=0.4.23", optional=true}
-momento = {version="^1.10.2", optional=true}
+momento = {version=">=1.10.2, < 1.21", optional=true}
 unstructured = {extras = ["docx", "pptx", "pdf"], version = ">=0.10.16,<0.10.18", optional=true}
 sentence-transformers = {version="^2.2.2", optional=true}
 torch = {version="^2.0.0", optional=true}
@@ -39,6 +39,7 @@ pyarrow = {version="15.0.0", optional=true}
 pdfplumber = {version="^0.10.2", optional=true}
 python-docx = {version="^1.1.0", optional=true}
 scrapy = {version="^2.11.0", optional=true}
+fastembed = {version="^0.3.1", optional=true}
 
 # ====CORE================================
 pyyaml = "^6.0.1"
@@ -111,6 +112,7 @@ all = [
     "metaphor-python", "neo4j",
     "litellm",
     "chainlit", "python-socketio",
+    "fastembed"
 ]
 # more granular groupings
 lancedb = ["lancedb", "tantivy", "pyarrow"]
@@ -135,6 +137,7 @@ mkdocs = [
 ]
 meilisearch = ["meilisearch-python-sdk"]
 momento = ["momento"]
+fastembed = ["fastembed"]
 
 
 # ================= DEV DEPS =================