mirror of
https://github.com/hwchase17/langchain
synced 2024-11-08 07:10:35 +00:00
78 lines
2.0 KiB
Python
78 lines
2.0 KiB
Python
import os
|
|
from typing import List, Optional
|
|
|
|
import nomic # type: ignore
|
|
from langchain_core.embeddings import Embeddings
|
|
from nomic import embed # type: ignore
|
|
|
|
|
|
class NomicEmbeddings(Embeddings):
|
|
"""NomicEmbeddings embedding model.
|
|
|
|
Example:
|
|
.. code-block:: python
|
|
|
|
from langchain_nomic import NomicEmbeddings
|
|
|
|
model = NomicEmbeddings()
|
|
"""
|
|
|
|
def __init__(
|
|
self,
|
|
*,
|
|
model: str,
|
|
nomic_api_key: Optional[str] = None,
|
|
dimensionality: Optional[int] = None,
|
|
):
|
|
"""Initialize NomicEmbeddings model.
|
|
|
|
Args:
|
|
model: model name
|
|
nomic_api_key: optionally, set the Nomic API key. Uses the NOMIC_API_KEY
|
|
environment variable by default.
|
|
"""
|
|
_api_key = nomic_api_key or os.environ.get("NOMIC_API_KEY")
|
|
if _api_key:
|
|
nomic.login(_api_key)
|
|
self.model = model
|
|
self.dimensionality = dimensionality
|
|
|
|
def embed(self, texts: List[str], *, task_type: str) -> List[List[float]]:
|
|
"""Embed texts.
|
|
|
|
Args:
|
|
texts: list of texts to embed
|
|
task_type: the task type to use when embedding. One of `search_query`,
|
|
`search_document`, `classification`, `clustering`
|
|
"""
|
|
|
|
output = embed.text(
|
|
texts=texts,
|
|
model=self.model,
|
|
task_type=task_type,
|
|
dimensionality=self.dimensionality,
|
|
)
|
|
return output["embeddings"]
|
|
|
|
def embed_documents(self, texts: List[str]) -> List[List[float]]:
|
|
"""Embed search docs.
|
|
|
|
Args:
|
|
texts: list of texts to embed as documents
|
|
"""
|
|
return self.embed(
|
|
texts=texts,
|
|
task_type="search_document",
|
|
)
|
|
|
|
def embed_query(self, text: str) -> List[float]:
|
|
"""Embed query text.
|
|
|
|
Args:
|
|
text: query text
|
|
"""
|
|
return self.embed(
|
|
texts=[text],
|
|
task_type="search_query",
|
|
)[0]
|