an attempt to use DeterministicFakeEmbedding with pinecone vectorstore fails with PineconeApiTypeError

[MichaelSkralivetsky]

Checked other resources

• I added a very descriptive title to this issue.
• I searched the LangChain documentation with the integrated search.
• I used the GitHub search to find a similar question and didn't find it.
• I am sure that this is a bug in LangChain rather than my code.
• The bug is not resolved by updating to the latest stable version of LangChain (or the specific integration package).

Example Code

import time
from pinecone import Pinecone, ServerlessSpec

pinecone_api_key = "mykey"
pc = Pinecone(api_key=pinecone_api_key)

index_name = "langchain-test-index"

existing_indexes = [index_info["name"] for index_info in pc.list_indexes()]

if index_name not in existing_indexes:
    pc.create_index(
        name=index_name,
        dimension=4096,
        metric="cosine",
        spec=ServerlessSpec(cloud="aws", region="us-east-1"),
    )
    while not pc.describe_index(index_name).status["ready"]:
        time.sleep(1)

from langchain_core.embeddings import DeterministicFakeEmbedding
embeddings = DeterministicFakeEmbedding(size=4096)

from langchain_pinecone import PineconeVectorStore

index = pc.Index(index_name)
vector_store = PineconeVectorStore(index=index, embedding=embeddings)

from uuid import uuid4

from langchain_core.documents import Document

document_1 = Document(
    page_content="I had chocalate chip pancakes and scrambled eggs for breakfast this morning.",
    metadata={"source": "tweet"},
)

documents = [
    document_1
]
uuids = [str(uuid4()) for _ in range(len(documents))]

vector_store.add_documents(documents=documents, ids=uuids)

Error Message and Stack Trace (if applicable)

PineconeApiTypeError                      Traceback (most recent call last)
Cell In[9], line 15
     10 documents = [
     11     document_1
     12 ]
     13 uuids = [str(uuid4()) for _ in range(len(documents))]
---> 15 vector_store.add_documents(documents=documents, ids=uuids)

File ~/.pythonlibs/mlrun-base/lib/python3.9/site-packages/langchain_core/vectorstores/base.py:287, in VectorStore.add_documents(self, documents, **kwargs)
    285     texts = [doc.page_content for doc in documents]
    286     metadatas = [doc.metadata for doc in documents]
--> 287     return self.add_texts(texts, metadatas, **kwargs)
    288 msg = (
    289     f"`add_documents` and `add_texts` has not been implemented "
    290     f"for {self.__class__.__name__} "
    291 )
    292 raise NotImplementedError(msg)

File ~/.pythonlibs/mlrun-base/lib/python3.9/site-packages/langchain_pinecone/vectorstores.py:283, in PineconeVectorStore.add_texts(self, texts, metadatas, ids, namespace, batch_size, embedding_chunk_size, async_req, id_prefix, **kwargs)
    280 vector_tuples = zip(chunk_ids, embeddings, chunk_metadatas)
    281 if async_req:
    282     # Runs the pinecone upsert asynchronously.
--> 283     async_res = [
    284         self._index.upsert(
    285             vectors=batch_vector_tuples,
    286             namespace=namespace,
    287             async_req=async_req,
    288             **kwargs,
    289         )
    290         for batch_vector_tuples in batch_iterate(batch_size, vector_tuples)
    291     ]
    292     [res.get() for res in async_res]
    293 else:

File ~/.pythonlibs/mlrun-base/lib/python3.9/site-packages/langchain_pinecone/vectorstores.py:284, in <listcomp>(.0)
    280 vector_tuples = zip(chunk_ids, embeddings, chunk_metadatas)
    281 if async_req:
    282     # Runs the pinecone upsert asynchronously.
    283     async_res = [
--> 284         self._index.upsert(
    285             vectors=batch_vector_tuples,
    286             namespace=namespace,
    287             async_req=async_req,
    288             **kwargs,
    289         )
    290         for batch_vector_tuples in batch_iterate(batch_size, vector_tuples)
    291     ]
    292     [res.get() for res in async_res]
    293 else:

File ~/.pythonlibs/mlrun-base/lib/python3.9/site-packages/pinecone/utils/error_handling.py:11, in validate_and_convert_errors.<locals>.inner_func(*args, **kwargs)
      8 @wraps(func)
      9 def inner_func(*args, **kwargs):
     10     try:
---> 11         return func(*args, **kwargs)
     12     except MaxRetryError as e:
     13         if isinstance(e.reason, ProtocolError):

File ~/.pythonlibs/mlrun-base/lib/python3.9/site-packages/pinecone/data/index.py:175, in Index.upsert(self, vectors, namespace, batch_size, show_progress, **kwargs)
    168     raise ValueError(
    169         "async_req is not supported when batch_size is provided."
    170         "To upsert in parallel, please follow: "
    171         "https://docs.pinecone.io/docs/insert-data#sending-upserts-in-parallel"
    172     )
    174 if batch_size is None:
--> 175     return self._upsert_batch(vectors, namespace, _check_type, **kwargs)
    177 if not isinstance(batch_size, int) or batch_size <= 0:
    178     raise ValueError("batch_size must be a positive integer")

File ~/.pythonlibs/mlrun-base/lib/python3.9/site-packages/pinecone/data/index.py:206, in Index._upsert_batch(self, vectors, namespace, _check_type, **kwargs)
    201 args_dict = self._parse_non_empty_args([("namespace", namespace)])
    202 vec_builder = lambda v: VectorFactory.build(v, check_type=_check_type)
    204 return self._vector_api.upsert(
    205     UpsertRequest(
--> 206         vectors=list(map(vec_builder, vectors)),
    207         **args_dict,
    208         _check_type=_check_type,
    209         **{k: v for k, v in kwargs.items() if k not in _OPENAPI_ENDPOINT_PARAMS},
    210     ),
    211     **{k: v for k, v in kwargs.items() if k in _OPENAPI_ENDPOINT_PARAMS},
    212 )

File ~/.pythonlibs/mlrun-base/lib/python3.9/site-packages/pinecone/data/index.py:202, in Index._upsert_batch.<locals>.<lambda>(v)
    194 def _upsert_batch(
    195     self,
    196     vectors: Union[List[Vector], List[tuple], List[dict]],
   (...)
    199     **kwargs,
    200 ) -> UpsertResponse:
    201     args_dict = self._parse_non_empty_args([("namespace", namespace)])
--> 202     vec_builder = lambda v: VectorFactory.build(v, check_type=_check_type)
    204     return self._vector_api.upsert(
    205         UpsertRequest(
    206             vectors=list(map(vec_builder, vectors)),
   (...)
    211         **{k: v for k, v in kwargs.items() if k in _OPENAPI_ENDPOINT_PARAMS},
    212     )

File ~/.pythonlibs/mlrun-base/lib/python3.9/site-packages/pinecone/data/vector_factory.py:26, in VectorFactory.build(item, check_type)
     24     return item
     25 elif isinstance(item, tuple):
---> 26     return VectorFactory._tuple_to_vector(item, check_type)
     27 elif isinstance(item, Mapping):
     28     return VectorFactory._dict_to_vector(item, check_type)

File ~/.pythonlibs/mlrun-base/lib/python3.9/site-packages/pinecone/data/vector_factory.py:42, in VectorFactory._tuple_to_vector(item, check_type)
     38     raise ValueError(
     39         "Sparse values are not supported in tuples. Please use either dicts or Vector objects as inputs."
     40     )
     41 else:
---> 42     return Vector(
     43         id=id,
     44         values=convert_to_list(values),
     45         metadata=metadata or {},
     46         _check_type=check_type,
     47     )

File ~/.pythonlibs/mlrun-base/lib/python3.9/site-packages/pinecone/core/openapi/shared/model_utils.py:33, in convert_js_args_to_python_args.<locals>.wrapped_init(_self, *args, **kwargs)
     31 if spec_property_naming:
     32     kwargs = change_keys_js_to_python(kwargs, _self if isinstance(_self, type) else _self.__class__)
---> 33 return fn(_self, *args, **kwargs)

File ~/.pythonlibs/mlrun-base/lib/python3.9/site-packages/pinecone/core/openapi/data/model/vector.py:289, in Vector.__init__(self, id, values, *args, **kwargs)
    286 self._visited_composed_classes = _visited_composed_classes + (self.__class__,)
    288 self.id = id
--> 289 self.values = values
    290 for var_name, var_value in kwargs.items():
    291     if (
    292         var_name not in self.attribute_map
    293         and self._configuration is not None
   (...)
    296     ):
    297         # discard variable.

File ~/.pythonlibs/mlrun-base/lib/python3.9/site-packages/pinecone/core/openapi/shared/model_utils.py:156, in OpenApiModel.__setattr__(self, attr, value)
    154 def __setattr__(self, attr, value):
    155     """set the value of an attribute using dot notation: `instance.attr = val`"""
--> 156     self[attr] = value

File ~/.pythonlibs/mlrun-base/lib/python3.9/site-packages/pinecone/core/openapi/shared/model_utils.py:432, in ModelNormal.__setitem__(self, name, value)
    429     self.__dict__[name] = value
    430     return
--> 432 self.set_attribute(name, value)

File ~/.pythonlibs/mlrun-base/lib/python3.9/site-packages/pinecone/core/openapi/shared/model_utils.py:132, in OpenApiModel.set_attribute(self, name, value)
    129     raise PineconeApiTypeError(error_msg, path_to_item=path_to_item, valid_classes=(str,), key_type=True)
    131 if self._check_type:
--> 132     value = validate_and_convert_types(
    133         value,
    134         required_types_mixed,
    135         path_to_item,
    136         self._spec_property_naming,
    137         self._check_type,
    138         configuration=self._configuration,
    139     )
    140 if (name,) in self.allowed_values:
    141     check_allowed_values(self.allowed_values, (name,), value)

File ~/.pythonlibs/mlrun-base/lib/python3.9/site-packages/pinecone/core/openapi/shared/model_utils.py:1489, in validate_and_convert_types(input_value, required_types_mixed, path_to_item, spec_property_naming, _check_type, configuration)
   1487         inner_path = list(path_to_item)
   1488         inner_path.append(index)
-> 1489         input_value[index] = validate_and_convert_types(
   1490             inner_value,
   1491             inner_required_types,
   1492             inner_path,
   1493             spec_property_naming,
   1494             _check_type,
   1495             configuration=configuration,
   1496         )
   1497 elif isinstance(input_value, dict):
   1498     if input_value == {}:
   1499         # allow an empty dict

File ~/.pythonlibs/mlrun-base/lib/python3.9/site-packages/pinecone/core/openapi/shared/model_utils.py:1453, in validate_and_convert_types(input_value, required_types_mixed, path_to_item, spec_property_naming, _check_type, configuration)
   1451         return converted_instance
   1452     else:
-> 1453         raise get_type_error(input_value, path_to_item, valid_classes, key_type=False)
   1455 # input_value's type is in valid_classes
   1456 if len(valid_classes) > 1 and configuration:
   1457     # there are valid classes which are not the current class

PineconeApiTypeError: Invalid type for variable '0'. Required value type is float and passed type was float64 at ['values'][0]

Description

the error is unexpected. the code is identical to the example in https://python.langchain.com/docs/integrations/vectorstores/pinecone/
same code works if using OllamaEmbeddings

System Info

System Information
------------------
> OS:  Linux
> OS Version:  #1 SMP Mon Dec 2 06:32:20 EST 2024
> Python Version:  3.9.18 | packaged by conda-forge | (main, Dec 23 2023, 16:33:10) 
[GCC 12.3.0]

Package Information
-------------------
> langchain_core: 0.3.28
> langchain: 0.3.13
> langchain_community: 0.3.13
> langsmith: 0.1.147
> langchain_chroma: 0.1.4
> langchain_milvus: 0.1.7
> langchain_ollama: 0.2.2
> langchain_pinecone: 0.2.0
> langchain_text_splitters: 0.3.4

Optional packages not installed
-------------------------------
> langserve

Other Dependencies
------------------
> aiohttp: 3.10.11
> async-timeout: 4.0.3
> chromadb: 0.5.23
> dataclasses-json: 0.6.7
> fastapi: 0.115.6
> httpx: 0.27.2
> httpx-sse: 0.4.0
> jsonpatch: 1.33
> langsmith-pyo3: Installed. No version info available.
> numpy: 1.26.4
> ollama: 0.4.5
> orjson: 3.10.12
> packaging: 24.0
> pinecone-client: 5.0.1
> pydantic: 2.10.4
> pydantic-settings: 2.7.0
> pymilvus: 2.5.3
> PyYAML: 6.0.2
> requests: 2.32.3
> requests-toolbelt: 1.0.0
> SQLAlchemy: 1.4.54
> tenacity: 8.5.0
> typing-extensions: 4.12.2