langchain/vectorstores/pinecone.py

from __future__ import annotations

import logging
import uuid
import warnings
from typing import TYPE_CHECKING, Any, Callable, Iterable, List, Optional, Tuple, Union

import numpy as np
from langchain_core.embeddings import Embeddings
from langchain_core.utils.iter import batch_iterate
from langchain_core.vectorstores import VectorStore

from langchain.docstore.document import Document
from langchain.vectorstores.utils import DistanceStrategy, maximal_marginal_relevance

if TYPE_CHECKING:
    from pinecone import Index

logger = logging.getLogger(__name__)


class Pinecone(VectorStore):
    """`Pinecone` vector store.

    To use, you should have the ``pinecone-client`` python package installed.

    Example:
        .. code-block:: python

            from langchain.vectorstores import Pinecone
            from langchain.embeddings.openai import OpenAIEmbeddings
            import pinecone

            # The environment should be the one specified next to the API key
            # in your Pinecone console
            pinecone.init(api_key="***", environment="...")
            index = pinecone.Index("langchain-demo")
            embeddings = OpenAIEmbeddings()
            vectorstore = Pinecone(index, embeddings.embed_query, "text")
    """

    def __init__(
        self,
        index: Any,
        embedding: Union[Embeddings, Callable],
        text_key: str,
        namespace: Optional[str] = None,
        distance_strategy: Optional[DistanceStrategy] = DistanceStrategy.COSINE,
    ):
        """Initialize with Pinecone client."""
        try:
            import pinecone
        except ImportError:
            raise ImportError(
                "Could not import pinecone python package. "
                "Please install it with `pip install pinecone-client`."
            )
        if not isinstance(embedding, Embeddings):
            warnings.warn(
                "Passing in `embedding` as a Callable is deprecated. Please pass in an"
                " Embeddings object instead."
            )
        if not isinstance(index, pinecone.index.Index):
            raise ValueError(
                f"client should be an instance of pinecone.index.Index, "
                f"got {type(index)}"
            )
        self._index = index
        self._embedding = embedding
        self._text_key = text_key
        self._namespace = namespace
        self.distance_strategy = distance_strategy

    @property
    def embeddings(self) -> Optional[Embeddings]:
        """Access the query embedding object if available."""
        if isinstance(self._embedding, Embeddings):
            return self._embedding
        return None

    def _embed_documents(self, texts: Iterable[str]) -> List[List[float]]:
        """Embed search docs."""
        if isinstance(self._embedding, Embeddings):
            return self._embedding.embed_documents(list(texts))
        return [self._embedding(t) for t in texts]

    def _embed_query(self, text: str) -> List[float]:
        """Embed query text."""
        if isinstance(self._embedding, Embeddings):
            return self._embedding.embed_query(text)
        return self._embedding(text)

    def add_texts(
        self,
        texts: Iterable[str],
        metadatas: Optional[List[dict]] = None,
        ids: Optional[List[str]] = None,
        namespace: Optional[str] = None,
        batch_size: int = 32,
        embedding_chunk_size: int = 1000,
        **kwargs: Any,
    ) -> List[str]:
        """Run more texts through the embeddings and add to the vectorstore.

        Upsert optimization is done by chunking the embeddings and upserting them.
        This is done to avoid memory issues and optimize using HTTP based embeddings.
        For OpenAI embeddings, use pool_threads>4 when constructing the pinecone.Index,
        embedding_chunk_size>1000 and batch_size~64 for best performance.
        Args:
            texts: Iterable of strings to add to the vectorstore.
            metadatas: Optional list of metadatas associated with the texts.
            ids: Optional list of ids to associate with the texts.
            namespace: Optional pinecone namespace to add the texts to.
            batch_size: Batch size to use when adding the texts to the vectorstore.
            embedding_chunk_size: Chunk size to use when embedding the texts.

        Returns:
            List of ids from adding the texts into the vectorstore.

        """
        if namespace is None:
            namespace = self._namespace

        texts = list(texts)
        ids = ids or [str(uuid.uuid4()) for _ in texts]
        metadatas = metadatas or [{} for _ in texts]
        for metadata, text in zip(metadatas, texts):
            metadata[self._text_key] = text

        # For loops to avoid memory issues and optimize when using HTTP based embeddings
        # The first loop runs the embeddings, it benefits when using OpenAI embeddings
        # The second loops runs the pinecone upsert asynchronously.
        for i in range(0, len(texts), embedding_chunk_size):
            chunk_texts = texts[i : i + embedding_chunk_size]
            chunk_ids = ids[i : i + embedding_chunk_size]
            chunk_metadatas = metadatas[i : i + embedding_chunk_size]
            embeddings = self._embed_documents(chunk_texts)
            async_res = [
                self._index.upsert(
                    vectors=batch,
                    namespace=namespace,
                    async_req=True,
                    **kwargs,
                )
                for batch in batch_iterate(
                    batch_size, zip(chunk_ids, embeddings, chunk_metadatas)
                )
            ]
            [res.get() for res in async_res]

        return ids

    def similarity_search_with_score(
        self,
        query: str,
        k: int = 4,
        filter: Optional[dict] = None,
        namespace: Optional[str] = None,
    ) -> List[Tuple[Document, float]]:
        """Return pinecone documents most similar to query, along with scores.

        Args:
            query: Text to look up documents similar to.
            k: Number of Documents to return. Defaults to 4.
            filter: Dictionary of argument(s) to filter on metadata
            namespace: Namespace to search in. Default will search in '' namespace.

        Returns:
            List of Documents most similar to the query and score for each
        """
        return self.similarity_search_by_vector_with_score(
            self._embed_query(query), k=k, filter=filter, namespace=namespace
        )

    def similarity_search_by_vector_with_score(
        self,
        embedding: List[float],
        *,
        k: int = 4,
        filter: Optional[dict] = None,
        namespace: Optional[str] = None,
    ) -> List[Tuple[Document, float]]:
        """Return pinecone documents most similar to embedding, along with scores."""

        if namespace is None:
            namespace = self._namespace
        docs = []
        results = self._index.query(
            [embedding],
            top_k=k,
            include_metadata=True,
            namespace=namespace,
            filter=filter,
        )
        for res in results["matches"]:
            metadata = res["metadata"]
            if self._text_key in metadata:
                text = metadata.pop(self._text_key)
                score = res["score"]
                docs.append((Document(page_content=text, metadata=metadata), score))
            else:
                logger.warning(
                    f"Found document with no `{self._text_key}` key. Skipping."
                )
        return docs

    def similarity_search(
        self,
        query: str,
        k: int = 4,
        filter: Optional[dict] = None,
        namespace: Optional[str] = None,
        **kwargs: Any,
    ) -> List[Document]:
        """Return pinecone documents most similar to query.

        Args:
            query: Text to look up documents similar to.
            k: Number of Documents to return. Defaults to 4.
            filter: Dictionary of argument(s) to filter on metadata
            namespace: Namespace to search in. Default will search in '' namespace.

        Returns:
            List of Documents most similar to the query and score for each
        """
        docs_and_scores = self.similarity_search_with_score(
            query, k=k, filter=filter, namespace=namespace, **kwargs
        )
        return [doc for doc, _ in docs_and_scores]

    def _select_relevance_score_fn(self) -> Callable[[float], float]:
        """
        The 'correct' relevance function
        may differ depending on a few things, including:
        - the distance / similarity metric used by the VectorStore
        - the scale of your embeddings (OpenAI's are unit normed. Many others are not!)
        - embedding dimensionality
        - etc.
        """

        if self.distance_strategy == DistanceStrategy.COSINE:
            return self._cosine_relevance_score_fn
        elif self.distance_strategy == DistanceStrategy.MAX_INNER_PRODUCT:
            return self._max_inner_product_relevance_score_fn
        elif self.distance_strategy == DistanceStrategy.EUCLIDEAN_DISTANCE:
            return self._euclidean_relevance_score_fn
        else:
            raise ValueError(
                "Unknown distance strategy, must be cosine, max_inner_product "
                "(dot product), or euclidean"
            )

    @staticmethod
    def _cosine_relevance_score_fn(score: float) -> float:
        """Pinecone returns cosine similarity scores between [-1,1]"""
        return (score + 1) / 2

    def max_marginal_relevance_search_by_vector(
        self,
        embedding: List[float],
        k: int = 4,
        fetch_k: int = 20,
        lambda_mult: float = 0.5,
        filter: Optional[dict] = None,
        namespace: Optional[str] = None,
        **kwargs: Any,
    ) -> List[Document]:
        """Return docs selected using the maximal marginal relevance.

        Maximal marginal relevance optimizes for similarity to query AND diversity
        among selected documents.

        Args:
            embedding: Embedding to look up documents similar to.
            k: Number of Documents to return. Defaults to 4.
            fetch_k: Number of Documents to fetch to pass to MMR algorithm.
            lambda_mult: Number between 0 and 1 that determines the degree
                        of diversity among the results with 0 corresponding
                        to maximum diversity and 1 to minimum diversity.
                        Defaults to 0.5.
        Returns:
            List of Documents selected by maximal marginal relevance.
        """
        if namespace is None:
            namespace = self._namespace
        results = self._index.query(
            [embedding],
            top_k=fetch_k,
            include_values=True,
            include_metadata=True,
            namespace=namespace,
            filter=filter,
        )
        mmr_selected = maximal_marginal_relevance(
            np.array([embedding], dtype=np.float32),
            [item["values"] for item in results["matches"]],
            k=k,
            lambda_mult=lambda_mult,
        )
        selected = [results["matches"][i]["metadata"] for i in mmr_selected]
        return [
            Document(page_content=metadata.pop((self._text_key)), metadata=metadata)
            for metadata in selected
        ]

    def max_marginal_relevance_search(
        self,
        query: str,
        k: int = 4,
        fetch_k: int = 20,
        lambda_mult: float = 0.5,
        filter: Optional[dict] = None,
        namespace: Optional[str] = None,
        **kwargs: Any,
    ) -> List[Document]:
        """Return docs selected using the maximal marginal relevance.

        Maximal marginal relevance optimizes for similarity to query AND diversity
        among selected documents.

        Args:
            query: Text to look up documents similar to.
            k: Number of Documents to return. Defaults to 4.
            fetch_k: Number of Documents to fetch to pass to MMR algorithm.
            lambda_mult: Number between 0 and 1 that determines the degree
                        of diversity among the results with 0 corresponding
                        to maximum diversity and 1 to minimum diversity.
                        Defaults to 0.5.
        Returns:
            List of Documents selected by maximal marginal relevance.
        """
        embedding = self._embed_query(query)
        return self.max_marginal_relevance_search_by_vector(
            embedding, k, fetch_k, lambda_mult, filter, namespace
        )

    @classmethod
    def get_pinecone_index(
        cls,
        index_name: Optional[str],
        pool_threads: int = 4,
    ) -> Index:
        """Return a Pinecone Index instance.

        Args:
            index_name: Name of the index to use.
            pool_threads: Number of threads to use for index upsert.
        Returns:
            Pinecone Index instance."""

        try:
            import pinecone
        except ImportError:
            raise ValueError(
                "Could not import pinecone python package. "
                "Please install it with `pip install pinecone-client`."
            )

        indexes = pinecone.list_indexes()  # checks if provided index exists

        if index_name in indexes:
            index = pinecone.Index(index_name, pool_threads=pool_threads)
        elif len(indexes) == 0:
            raise ValueError(
                "No active indexes found in your Pinecone project, "
                "are you sure you're using the right Pinecone API key and Environment? "
                "Please double check your Pinecone dashboard."
            )
        else:
            raise ValueError(
                f"Index '{index_name}' not found in your Pinecone project. "
                f"Did you mean one of the following indexes: {', '.join(indexes)}"
            )
        return index

    @classmethod
    def from_texts(
        cls,
        texts: List[str],
        embedding: Embeddings,
        metadatas: Optional[List[dict]] = None,
        ids: Optional[List[str]] = None,
        batch_size: int = 32,
        text_key: str = "text",
        namespace: Optional[str] = None,
        index_name: Optional[str] = None,
        upsert_kwargs: Optional[dict] = None,
        pool_threads: int = 4,
        embeddings_chunk_size: int = 1000,
        **kwargs: Any,
    ) -> Pinecone:
        """Construct Pinecone wrapper from raw documents.

        This is a user friendly interface that:
            1. Embeds documents.
            2. Adds the documents to a provided Pinecone index

        This is intended to be a quick way to get started.

        The `pool_threads` affects the speed of the upsert operations.
        Example:
            .. code-block:: python

                from langchain.vectorstores import Pinecone
                from langchain.embeddings import OpenAIEmbeddings
                import pinecone

                # The environment should be the one specified next to the API key
                # in your Pinecone console
                pinecone.init(api_key="***", environment="...")
                embeddings = OpenAIEmbeddings()
                pinecone = Pinecone.from_texts(
                    texts,
                    embeddings,
                    index_name="langchain-demo"
                )
        """
        pinecone_index = cls.get_pinecone_index(index_name, pool_threads)
        pinecone = cls(pinecone_index, embedding, text_key, namespace, **kwargs)

        pinecone.add_texts(
            texts,
            metadatas=metadatas,
            ids=ids,
            namespace=namespace,
            batch_size=batch_size,
            embedding_chunk_size=embeddings_chunk_size,
            **(upsert_kwargs or {}),
        )
        return pinecone

    @classmethod
    def from_existing_index(
        cls,
        index_name: str,
        embedding: Embeddings,
        text_key: str = "text",
        namespace: Optional[str] = None,
        pool_threads: int = 4,
    ) -> Pinecone:
        """Load pinecone vectorstore from index name."""
        pinecone_index = cls.get_pinecone_index(index_name, pool_threads)
        return cls(pinecone_index, embedding, text_key, namespace)

    def delete(
        self,
        ids: Optional[List[str]] = None,
        delete_all: Optional[bool] = None,
        namespace: Optional[str] = None,
        filter: Optional[dict] = None,
        **kwargs: Any,
    ) -> None:
        """Delete by vector IDs or filter.
        Args:
            ids: List of ids to delete.
            filter: Dictionary of conditions to filter vectors to delete.
        """

        if namespace is None:
            namespace = self._namespace

        if delete_all:
            self._index.delete(delete_all=True, namespace=namespace, **kwargs)
        elif ids is not None:
            chunk_size = 1000
            for i in range(0, len(ids), chunk_size):
                chunk = ids[i : i + chunk_size]
                self._index.delete(ids=chunk, namespace=namespace, **kwargs)
        elif filter is not None:
            self._index.delete(filter=filter, namespace=namespace, **kwargs)
        else:
            raise ValueError("Either ids, delete_all, or filter must be provided.")

        return None