ConsistentFakeSparseEmbeddings Class — langchain Architecture

Architecture documentation for the ConsistentFakeSparseEmbeddings class in common.py from the langchain codebase.

Class python

Entity Profile

Dependency Diagram

graph TD
  e01269d9_4bda_3502_ba27_5bad4874d109["ConsistentFakeSparseEmbeddings"]
  d2c50637_94ac_9030_8f99_d10858fb4c29["SparseEmbeddings"]
  e01269d9_4bda_3502_ba27_5bad4874d109 -->|extends| d2c50637_94ac_9030_8f99_d10858fb4c29
  b62fe5d9_7b83_0e14_ef50_3645cafa8bc5["common.py"]
  e01269d9_4bda_3502_ba27_5bad4874d109 -->|defined in| b62fe5d9_7b83_0e14_ef50_3645cafa8bc5
  f03b2f6c_27fa_56c7_4cb2_ae73cc7ab8b0["__init__()"]
  e01269d9_4bda_3502_ba27_5bad4874d109 -->|method| f03b2f6c_27fa_56c7_4cb2_ae73cc7ab8b0
  b9e09bee_ce68_fd2b_005f_81e22de97c53["embed_documents()"]
  e01269d9_4bda_3502_ba27_5bad4874d109 -->|method| b9e09bee_ce68_fd2b_005f_81e22de97c53
  5369e966_8af1_33ce_87ff_9b00c71877a5["embed_query()"]
  e01269d9_4bda_3502_ba27_5bad4874d109 -->|method| 5369e966_8af1_33ce_87ff_9b00c71877a5

Relationship Graph

Source Code

libs/partners/qdrant/tests/integration_tests/common.py lines 61–86

class ConsistentFakeSparseEmbeddings(SparseEmbeddings):
    """Fake sparse embeddings which remembers all the texts seen so far
    "to return consistent vectors for the same texts.
    """

    def __init__(self, dimensionality: int = 25) -> None:
        self.known_texts: list[str] = []
        self.dimensionality = dimensionality

    def embed_documents(self, texts: list[str]) -> list[SparseVector]:
        """Return consistent embeddings for each text seen so far."""
        out_vectors = []
        for text in texts:
            if text not in self.known_texts:
                self.known_texts.append(text)
            index = self.known_texts.index(text)
            indices = [i + index for i in range(self.dimensionality)]
            values = [1.0] * (self.dimensionality - 1) + [float(index)]
            out_vectors.append(SparseVector(indices=indices, values=values))
        return out_vectors

    def embed_query(self, text: str) -> SparseVector:
        """Return consistent embeddings for the text, if seen before, or a constant
        one if the text is unknown.
        """
        return self.embed_documents([text])[0]