feat: Support for Tigris Vector Database for vector search (#5703)

### Changes
- New vector store integration - [Tigris](https://tigrisdata.com)
- Adds [tigrisdb](https://pypi.org/project/tigrisdb/) optional
dependency
- Example notebook demonstrating usage

Fixes #5535 
Closes tigrisdata/tigris-client-python#40

#### Twitter handles
We'd love a shoutout on our
[@TigrisData](https://twitter.com/TigrisData) and
[@adilansari](https://twitter.com/adilansari) twitter handles

#### Who can review?
@dev2049

---------

Co-authored-by: Harrison Chase <hw.chase.17@gmail.com>

docs/modules/indexes/vectorstores/examples/tigris.ipynb

@@ -0,0 +1,199 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "source": [
+    "# Tigris\n",
+    "\n",
+    "> [Tigris](htttps://tigrisdata.com) is an open source Serverless NoSQL Database and Search Platform designed to simplify building high-performance vector search applications.\n",
+    "> Tigris eliminates the infrastructure complexity of managing, operating, and synchronizing multiple tools, allowing you to focus on building great applications instead."
+   ],
+   "metadata": {
+    "collapsed": false
+   }
+  },
+  {
+   "cell_type": "markdown",
+   "source": [
+    "This notebook guides you how to use Tigris as your VectorStore"
+   ],
+   "metadata": {
+    "collapsed": false
+   }
+  },
+  {
+   "cell_type": "markdown",
+   "source": [
+    "**Pre requisites**\n",
+    "1. An OpenAI account. You can sign up for an account [here](https://platform.openai.com/)\n",
+    "2. [Sign up for a free Tigris account](https://console.preview.tigrisdata.cloud). Once you have signed up for the Tigris account, create a new project called `vectordemo`. Next, make a note of the *Uri* for the region you've created your project in, the **clientId** and **clientSecret**. You can get all this information from the **Application Keys** section of the project."
+   ],
+   "metadata": {
+    "collapsed": false
+   }
+  },
+  {
+   "cell_type": "markdown",
+   "source": [
+    "Let's first install our dependencies:"
+   ],
+   "metadata": {
+    "collapsed": false
+   }
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "outputs": [],
+   "source": [
+    "!pip install tigrisdb openapi-schema-pydantic openai tiktoken"
+   ],
+   "metadata": {
+    "collapsed": false
+   }
+  },
+  {
+   "cell_type": "markdown",
+   "source": [
+    "We will load the `OpenAI` api key and `Tigris` credentials in our environment"
+   ],
+   "metadata": {
+    "collapsed": false
+   }
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "outputs": [],
+   "source": [
+    "import os\n",
+    "import getpass\n",
+    "\n",
+    "os.environ['OPENAI_API_KEY'] = getpass.getpass('OpenAI API Key:')\n",
+    "os.environ['TIGRIS_PROJECT'] = getpass.getpass('Tigris Project Name:')\n",
+    "os.environ['TIGRIS_CLIENT_ID'] = getpass.getpass('Tigris Client Id:')\n",
+    "os.environ['TIGRIS_CLIENT_SECRET'] = getpass.getpass('Tigris Client Secret:')"
+   ],
+   "metadata": {
+    "collapsed": false
+   }
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "outputs": [],
+   "source": [
+    "from langchain.embeddings.openai import OpenAIEmbeddings\n",
+    "from langchain.text_splitter import CharacterTextSplitter\n",
+    "from langchain.vectorstores import Tigris\n",
+    "from langchain.document_loaders import TextLoader"
+   ],
+   "metadata": {
+    "collapsed": false
+   }
+  },
+  {
+   "cell_type": "markdown",
+   "source": [
+    "### Initialize Tigris vector store\n",
+    "Let's import our test dataset:"
+   ],
+   "metadata": {
+    "collapsed": false
+   }
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "outputs": [],
+   "source": [
+    "loader = TextLoader('../../../state_of_the_union.txt')\n",
+    "documents = loader.load()\n",
+    "text_splitter = CharacterTextSplitter(chunk_size=1000, chunk_overlap=0)\n",
+    "docs = text_splitter.split_documents(documents)\n",
+    "\n",
+    "embeddings = OpenAIEmbeddings()"
+   ],
+   "metadata": {
+    "collapsed": false
+   }
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "outputs": [],
+   "source": [
+    "vector_store = Tigris.from_documents(docs, embeddings, index_name=\"my_embeddings\")"
+   ],
+   "metadata": {
+    "collapsed": false
+   }
+  },
+  {
+   "cell_type": "markdown",
+   "source": [
+    "### Similarity Search"
+   ],
+   "metadata": {
+    "collapsed": false
+   }
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "outputs": [],
+   "source": [
+    "query = \"What did the president say about Ketanji Brown Jackson\"\n",
+    "found_docs = vector_store.similarity_search(query)\n",
+    "print(found_docs)"
+   ],
+   "metadata": {
+    "collapsed": false
+   }
+  },
+  {
+   "cell_type": "markdown",
+   "source": [
+    "### Similarity Search with score (vector distance)"
+   ],
+   "metadata": {
+    "collapsed": false
+   }
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "outputs": [],
+   "source": [
+    "query = \"What did the president say about Ketanji Brown Jackson\"\n",
+    "result = vector_store.similarity_search_with_score(query)\n",
+    "for (doc, score) in result:\n",
+    "    print(f\"document={doc}, score={score}\")"
+   ],
+   "metadata": {
+    "collapsed": false
+   }
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 2
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython2",
+   "version": "2.7.6"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 0
+}

langchain/vectorstores/__init__.py

@@ -20,6 +20,7 @@ from langchain.vectorstores.redis import Redis
 from langchain.vectorstores.sklearn import SKLearnVectorStore
 from langchain.vectorstores.supabase import SupabaseVectorStore
 from langchain.vectorstores.tair import Tair
+from langchain.vectorstores.tigris import Tigris
 from langchain.vectorstores.typesense import Typesense
 from langchain.vectorstores.vectara import Vectara
 from langchain.vectorstores.weaviate import Weaviate
@@ -54,4 +55,5 @@ __all__ = [
     "Typesense",
     "Clickhouse",
     "ClickhouseSettings",
+    "Tigris",
 ]

langchain/vectorstores/tigris.py

@@ -0,0 +1,142 @@
+from __future__ import annotations
+
+import itertools
+from typing import TYPE_CHECKING, Any, Iterable, List, Optional, Tuple
+
+from langchain.embeddings.base import Embeddings
+from langchain.schema import Document
+from langchain.vectorstores import VectorStore
+
+if TYPE_CHECKING:
+    from tigrisdb import TigrisClient
+    from tigrisdb import VectorStore as TigrisVectorStore
+    from tigrisdb.types.filters import Filter as TigrisFilter
+    from tigrisdb.types.vector import Document as TigrisDocument
+
+
+class Tigris(VectorStore):
+    def __init__(self, client: TigrisClient, embeddings: Embeddings, index_name: str):
+        """Initialize Tigris vector store"""
+        try:
+            import tigrisdb  # noqa: F401
+        except ImportError:
+            raise ValueError(
+                "Could not import tigrisdb python package. "
+                "Please install it with `pip install tigrisdb`"
+            )
+
+        self._embed_fn = embeddings
+        self._vector_store = TigrisVectorStore(client.get_search(), index_name)
+
+    @property
+    def search_index(self) -> TigrisVectorStore:
+        return self._vector_store
+
+    def add_texts(
+        self,
+        texts: Iterable[str],
+        metadatas: Optional[List[dict]] = None,
+        ids: Optional[List[str]] = None,
+        **kwargs: Any,
+    ) -> List[str]:
+        """Run more texts through the embeddings and add to the vectorstore.
+
+        Args:
+            texts: Iterable of strings to add to the vectorstore.
+            metadatas: Optional list of metadatas associated with the texts.
+            ids: Optional list of ids for documents.
+                Ids will be autogenerated if not provided.
+            kwargs: vectorstore specific parameters
+
+        Returns:
+            List of ids from adding the texts into the vectorstore.
+        """
+        docs = self._prep_docs(texts, metadatas, ids)
+        result = self.search_index.add_documents(docs)
+        return [r.id for r in result]
+
+    def similarity_search(
+        self,
+        query: str,
+        k: int = 4,
+        filter: Optional[TigrisFilter] = None,
+        **kwargs: Any,
+    ) -> List[Document]:
+        """Return docs most similar to query."""
+        docs_with_scores = self.similarity_search_with_score(query, k, filter)
+        return [doc for doc, _ in docs_with_scores]
+
+    def similarity_search_with_score(
+        self,
+        query: str,
+        k: int = 4,
+        filter: Optional[TigrisFilter] = None,
+    ) -> List[Tuple[Document, float]]:
+        """Run similarity search with Chroma with distance.
+
+        Args:
+            query (str): Query text to search for.
+            k (int): Number of results to return. Defaults to 4.
+            filter (Optional[TigrisFilter]): Filter by metadata. Defaults to None.
+
+        Returns:
+            List[Tuple[Document, float]]: List of documents most similar to the query
+                text with distance in float.
+        """
+        vector = self._embed_fn.embed_query(query)
+        result = self.search_index.similarity_search(
+            vector=vector, k=k, filter_by=filter
+        )
+        docs: List[Tuple[Document, float]] = []
+        for r in result:
+            docs.append(
+                (
+                    Document(
+                        page_content=r.doc["text"], metadata=r.doc.get("metadata")
+                    ),
+                    r.score,
+                )
+            )
+        return docs
+
+    @classmethod
+    def from_texts(
+        cls,
+        texts: List[str],
+        embedding: Embeddings,
+        metadatas: Optional[List[dict]] = None,
+        ids: Optional[List[str]] = None,
+        client: Optional[TigrisClient] = None,
+        index_name: Optional[str] = None,
+        **kwargs: Any,
+    ) -> Tigris:
+        """Return VectorStore initialized from texts and embeddings."""
+        if not index_name:
+            raise ValueError("`index_name` is required")
+
+        if not client:
+            client = TigrisClient()
+        store = cls(client, embedding, index_name)
+        store.add_texts(texts=texts, metadatas=metadatas, ids=ids)
+        return store
+
+    def _prep_docs(
+        self,
+        texts: Iterable[str],
+        metadatas: Optional[List[dict]],
+        ids: Optional[List[str]],
+    ) -> List[TigrisDocument]:
+        embeddings: List[List[float]] = self._embed_fn.embed_documents(list(texts))
+        docs: List[TigrisDocument] = []
+        for t, m, e, _id in itertools.zip_longest(
+            texts, metadatas or [], embeddings or [], ids or []
+        ):
+            doc: TigrisDocument = {
+                "text": t,
+                "embeddings": e or [],
+                "metadata": m or {},
+            }
+            if _id:
+                doc["id"] = _id
+            docs.append(doc)
+        return docs

pyproject.toml

@@ -103,6 +103,7 @@ py-trello = {version = "^0.19.0", optional = true}
 momento = {version = "^1.5.0", optional = true}
 bibtexparser = {version = "^1.4.0", optional = true}
 pyspark = {version = "^3.4.0", optional = true}
+tigrisdb = {version = "^1.0.0b6", optional = true}
 
 [tool.poetry.group.docs.dependencies]
 autodoc_pydantic = "^1.8.0"
@@ -278,7 +279,8 @@ all = [
     "azure-ai-formrecognizer",
     "azure-ai-vision",
     "azure-cognitiveservices-speech",
-    "momento"
+    "momento",
+    "tigrisdb"
 ]
 
 # An extra used to be able to add extended testing.