mirror of
https://github.com/hwchase17/langchain.git
synced 2025-09-29 07:19:59 +00:00
844955d6e1
Assigning missed defaults in various classes. Most clients were being
assigned during the `model_validator(mode="before")` step, so this
change should amount to a no-op in those cases.
---
This PR was autogenerated using gritql
```shell
grit apply 'class_definition(name=$C, $body, superclasses=$S) where {
$C <: ! "Config", // Does not work in this scope, but works after class_definition
$body <: block($statements),
$statements <: some bubble assignment(left=$x, right=$y, type=$t) as $A where {
or {
$y <: `Field($z)`,
$x <: "model_config"
}
},
// And has either Any or Optional fields without a default
$statements <: some bubble assignment(left=$x, right=$y, type=$t) as $A where {
$t <: or {
r"Optional.*",
r"Any",
r"Union[None, .*]",
r"Union[.*, None, .*]",
r"Union[.*, None]",
},
$y <: ., // Match empty node
$t => `$t = None`,
},
}
' --language python .
```
75 lines
2.3 KiB
Python
75 lines
2.3 KiB
Python
from typing import Any, List
|
|
|
|
from langchain_core.embeddings import Embeddings
|
|
from pydantic import BaseModel, ConfigDict
|
|
|
|
DEFAULT_MODEL_URL = "https://tfhub.dev/google/universal-sentence-encoder-multilingual/3"
|
|
|
|
|
|
class TensorflowHubEmbeddings(BaseModel, Embeddings):
|
|
"""TensorflowHub embedding models.
|
|
|
|
To use, you should have the ``tensorflow_text`` python package installed.
|
|
|
|
Example:
|
|
.. code-block:: python
|
|
|
|
from langchain_community.embeddings import TensorflowHubEmbeddings
|
|
url = "https://tfhub.dev/google/universal-sentence-encoder-multilingual/3"
|
|
tf = TensorflowHubEmbeddings(model_url=url)
|
|
"""
|
|
|
|
embed: Any = None #: :meta private:
|
|
model_url: str = DEFAULT_MODEL_URL
|
|
"""Model name to use."""
|
|
|
|
def __init__(self, **kwargs: Any):
|
|
"""Initialize the tensorflow_hub and tensorflow_text."""
|
|
super().__init__(**kwargs)
|
|
try:
|
|
import tensorflow_hub
|
|
except ImportError:
|
|
raise ImportError(
|
|
"Could not import tensorflow-hub python package. "
|
|
"Please install it with `pip install tensorflow-hub``."
|
|
)
|
|
try:
|
|
import tensorflow_text # noqa
|
|
except ImportError:
|
|
raise ImportError(
|
|
"Could not import tensorflow_text python package. "
|
|
"Please install it with `pip install tensorflow_text``."
|
|
)
|
|
|
|
self.embed = tensorflow_hub.load(self.model_url)
|
|
|
|
model_config = ConfigDict(
|
|
extra="forbid",
|
|
)
|
|
|
|
def embed_documents(self, texts: List[str]) -> List[List[float]]:
|
|
"""Compute doc embeddings using a TensorflowHub embedding model.
|
|
|
|
Args:
|
|
texts: The list of texts to embed.
|
|
|
|
Returns:
|
|
List of embeddings, one for each text.
|
|
"""
|
|
texts = list(map(lambda x: x.replace("\n", " "), texts))
|
|
embeddings = self.embed(texts).numpy()
|
|
return embeddings.tolist()
|
|
|
|
def embed_query(self, text: str) -> List[float]:
|
|
"""Compute query embeddings using a TensorflowHub embedding model.
|
|
|
|
Args:
|
|
text: The text to embed.
|
|
|
|
Returns:
|
|
Embeddings for the text.
|
|
"""
|
|
text = text.replace("\n", " ")
|
|
embedding = self.embed([text]).numpy()[0]
|
|
return embedding.tolist()
|