jyapayne/langflow
1
0
Fork 0
Code Issues Pull requests Projects Releases 1 Packages Wiki Activity Actions

Add Upstash Vector Store support (#2004)

Browse source 
* docs(migration): add UPSTASH_VECTOR_REST_URL and UPSTASH_VECTOR_REST_TOKEN to the list of default environment variables
pyproject.toml: add upstash-vector package as a dependency with version 0.4.0
src/backend/base/langflow/components/vectorsearch/UpstashSearch.py: create UpstashSearchComponent for implementing a Vector Store using Upstash
src/backend/base/langflow/components/vectorstores/Upstash.py: create UpstashVectorStoreComponent for implementing a Vector Store using Upstash
src/backend/base/langflow/services/settings/constants.py: add UPSTASH_VECTOR_REST_URL and UPSTASH_VECTOR_REST_TOKEN to the list of variables to get from the environment

* ✨ (UpstashSearch.py): Add support for 'number_of_results' and 'text_key' parameters in UpstashSearchComponent to enhance search functionality
♻️ (Upstash.py): Refactor UpstashVectorStoreComponent to include 'text_key' parameter for consistency and improved functionality

* ♻️ (Upstash.py): refactor UpstashVectorStoreComponent to improve code readability and maintainability by restructuring the instantiation of UpstashVectorStore instances based on conditions and adding support for adding documents directly to the instance.

* feat: Update langchain-core, langchainhub, langsmith, and requests dependencies to latest versions
This commit is contained in:
Gabriel Luiz Freitas Almeida 2024-05-30 12:02:40 -07:00 • committed by GitHub
commit 4c113d5ac0
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
7 changed files with 239 additions and 52 deletions
Download patch file Download diff file Expand all files Collapse all files

79
src/backend/base/langflow/components/vectorsearch/UpstashSearch.py Normal file
View file

@ -0,0 +1,79 @@
from typing import List, Optional
from langchain_core.embeddings import Embeddings
from langflow.components.vectorstores.base.model import LCVectorStoreComponent
from langflow.components.vectorstores.Upstash import UpstashVectorStoreComponent
from langflow.field_typing import Text
from langflow.schema import Record
class UpstashSearchComponent(UpstashVectorStoreComponent, LCVectorStoreComponent):
"""
A custom component for implementing a Vector Store using Upstash.
"""
display_name: str = "Upstash Search"
description: str = "Search an Upstash Vector Store for similar documents."
def build_config(self):
"""
Builds the configuration for the component.
Returns:
- dict: A dictionary containing the configuration options for the component.
"""
return {
"search_type": {
"display_name": "Search Type",
"options": ["Similarity", "MMR"],
},
"input_value": {"display_name": "Input"},
"inputs": {"display_name": "Input", "input_types": ["Document", "Record"]},
"embedding": {
"display_name": "Embedding",
"input_types": ["Embeddings"],
"info": "To use Upstash's embeddings, don't provide an embedding.",
},
"index_url": {
"display_name": "Index URL",
"info": "The URL of the Upstash index.",
},
"index_token": {
"display_name": "Index Token",
"info": "The token for the Upstash index.",
},
"number_of_results": {
"display_name": "Number of Results",
"info": "Number of results to return.",
"advanced": True,
},
"text_key": {
"display_name": "Text Key",
"info": "The key in the record to use as text.",
"advanced": True,
},
}
def build( # type: ignore[override]
self,
input_value: Text,
search_type: str,
text_key: str = "text",
index_url: Optional[str] = None,
index_token: Optional[str] = None,
embedding: Optional[Embeddings] = None,
number_of_results: int = 4,
) -> List[Record]:
vector_store = super().build(
embedding=embedding,
text_key=text_key,
index_url=index_url,
index_token=index_token,
)
if not vector_store:
raise ValueError("Failed to load the Upstash Vector Store.")
return self.search_with_vector_store(
input_value=input_value, search_type=search_type, vector_store=vector_store, k=number_of_results
)

89
src/backend/base/langflow/components/vectorstores/Upstash.py Normal file
View file

@ -0,0 +1,89 @@
from typing import List, Optional, Union
from langchain_community.vectorstores.upstash import UpstashVectorStore
from langchain_core.embeddings import Embeddings
from langchain_core.retrievers import BaseRetriever
from langchain_core.vectorstores import VectorStore
from langflow.custom import CustomComponent
from langflow.schema.schema import Record
class UpstashVectorStoreComponent(CustomComponent):
"""
A custom component for implementing a Vector Store using Upstash.
"""
display_name: str = "Upstash"
description: str = "Create and Utilize an Upstash Vector Store"
def build_config(self):
"""
Builds the configuration for the component.
Returns:
- dict: A dictionary containing the configuration options for the component.
"""
return {
"inputs": {"display_name": "Input", "input_types": ["Document", "Record"]},
"embedding": {
"display_name": "Embedding",
"input_types": ["Embeddings"],
"info": "To use Upstash's embeddings, don't provide an embedding.",
},
"index_url": {
"display_name": "Index URL",
"info": "The URL of the Upstash index.",
},
"index_token": {
"display_name": "Index Token",
"info": "The token for the Upstash index.",
},
"text_key": {
"display_name": "Text Key",
"info": "The key in the record to use as text.",
"advanced": True,
},
}
def build(
self,
inputs: Optional[List[Record]] = None,
text_key: str = "text",
index_url: Optional[str] = None,
index_token: Optional[str] = None,
embedding: Optional[Embeddings] = None,
) -> Union[VectorStore, BaseRetriever]:
documents = []
for _input in inputs or []:
if isinstance(_input, Record):
documents.append(_input.to_lc_document())
else:
documents.append(_input)
use_upstash_embedding = embedding is None
if not documents:
upstash_vs = UpstashVectorStore(
embedding=embedding or use_upstash_embedding,
text_key=text_key,
index_url=index_url,
index_token=index_token,
)
else:
if use_upstash_embedding:
upstash_vs = UpstashVectorStore(
embedding=use_upstash_embedding,
text_key=text_key,
index_url=index_url,
index_token=index_token,
)
upstash_vs.add_documents(documents)
elif embedding:
upstash_vs = UpstashVectorStore.from_documents(
documents=documents, # type: ignore
embedding=embedding,
text_key=text_key,
index_url=index_url,
index_token=index_token,
)
return upstash_vs

2
src/backend/base/langflow/services/settings/constants.py
View file

@ -17,6 +17,8 @@ VARIABLES_TO_GET_FROM_ENVIRONMENT = [
"PINECONE_API_KEY",
"SEARCHAPI_API_KEY",
"SERPAPI_API_KEY",
"UPSTASH_VECTOR_REST_URL",
"UPSTASH_VECTOR_REST_TOKEN",
"VECTARA_CUSTOMER_ID",
"VECTARA_CORPUS_ID",
"VECTARA_API_KEY",

24
src/backend/base/poetry.lock generated
View file

@ -1224,13 +1224,13 @@ extended-testing = ["aiosqlite (>=0.19.0,<0.20.0)", "aleph-alpha-client (>=2.15.
[[package]]
name = "langchain-core"
version = "0.2.1"
version = "0.2.2"
description = "Building applications with LLMs through composability"
optional = false
python-versions = "<4.0,>=3.8.1"
files = [
{file = "langchain_core-0.2.1-py3-none-any.whl", hash = "sha256:3521e1e573988c47399fca9739270c5d34f8ecec147253ad829eb9ff288f76d5"},
{file = "langchain_core-0.2.1.tar.gz", hash = "sha256:49383126168d934559a543ce812c485048d9e6ac9b6798fbf3d4a72b6bba5b0c"},
{file = "langchain_core-0.2.2-py3-none-any.whl", hash = "sha256:4b3b55a5f214acbcf8d6d8e322da3a9d6248d6b6f45ac1b86ab0494fd3716128"},
{file = "langchain_core-0.2.2.tar.gz", hash = "sha256:6884a87f7ac8e0d43e4d83c5f9efa95236c7bd535e22a0a51db19156875b4cd6"},
]
[package.dependencies]
@ -1281,13 +1281,13 @@ extended-testing = ["beautifulsoup4 (>=4.12.3,<5.0.0)", "lxml (>=4.9.3,<6.0)"]
[[package]]
name = "langchainhub"
version = "0.1.16"
version = "0.1.17"
description = "The LangChain Hub API client"
optional = false
python-versions = "<4.0,>=3.8.1"
files = [
{file = "langchainhub-0.1.16-py3-none-any.whl", hash = "sha256:a4379a1879cc6b441b8d02cc65e28a54f160fba61c9d1d4b0eddc3a276dff99a"},
{file = "langchainhub-0.1.16.tar.gz", hash = "sha256:9f11e68fddb575e70ef4b28800eedbd9eeb180ba508def04f7153ea5b246b6fc"},
{file = "langchainhub-0.1.17-py3-none-any.whl", hash = "sha256:4c609b3948252c71670f0d98f73413b515cfd2f6701a7b40ce959203e6133e04"},
{file = "langchainhub-0.1.17.tar.gz", hash = "sha256:af7df0cb1cebc7a6e0864e8632ae48ecad39ed96568f699c78657b9d04e50b46"},
]
[package.dependencies]
@ -1296,13 +1296,13 @@ types-requests = ">=2.31.0.2,<3.0.0.0"
[[package]]
name = "langsmith"
version = "0.1.63"
version = "0.1.65"
description = "Client library to connect to the LangSmith LLM Tracing and Evaluation Platform."
optional = false
python-versions = "<4.0,>=3.8.1"
files = [
{file = "langsmith-0.1.63-py3-none-any.whl", hash = "sha256:7810afdf5e3f3b472fc581a29371fb96cd843dde2149e048d1b9610325159d1e"},
{file = "langsmith-0.1.63.tar.gz", hash = "sha256:a609405b52f6f54df442a142cbf19ab38662d54e532f96028b4c546434d4afdf"},
{file = "langsmith-0.1.65-py3-none-any.whl", hash = "sha256:ab4487029240e69cca30da1065f1e9138e5a7ca2bbe8c697f0bd7d5839f71cf7"},
{file = "langsmith-0.1.65.tar.gz", hash = "sha256:d3c2eb2391478bd79989f02652cf66e29a7959d677614b6993a47cef43f7f43b"},
]
[package.dependencies]
@ -2466,13 +2466,13 @@ files = [
[[package]]
name = "requests"
version = "2.32.2"
version = "2.32.3"
description = "Python HTTP for Humans."
optional = false
python-versions = ">=3.8"
files = [
{file = "requests-2.32.2-py3-none-any.whl", hash = "sha256:fc06670dd0ed212426dfeb94fc1b983d917c4f9847c863f313c9dfaaffb7c23c"},
{file = "requests-2.32.2.tar.gz", hash = "sha256:dd951ff5ecf3e3b3aa26b40703ba77495dab41da839ae72ef3c8e5d8e2433289"},
{file = "requests-2.32.3-py3-none-any.whl", hash = "sha256:70761cfe03c773ceb22aa2f671b4757976145175cdfca038c02654d061d6dcc6"},
{file = "requests-2.32.3.tar.gz", hash = "sha256:55365417734eb18255590a9ff9eb97e9e1da868d4ccd6402399eaf68af20a760"},
]
[package.dependencies]