EmbedInDB
v0.2.1
Embedin est une base de données vectorielle open source et une bibliothèque efficace qui convertit de manière transparente des bases de données populaires comme MySQL, PostgreSQL et MS SQL Server en bases de données vectorielles avec un effort nul.
L'intègre est une solution idéale pour les applications d'IA comme le traitement du langage naturel, la reconnaissance d'image et les systèmes de recommandation, offrant une indexation et une récupération rapides. Son API simple et son langage de requête assurent la facilité d'utilisation et l'intégration transparente.
Python 3,7 ou plus.
pip install embedin from embedin import Embedin
client = Embedin ( collection_name = "test_collection" , texts = [ "This is a test" , "Hello world!" ])
result = client . query ( "These are tests" , top_k = 1 ) # Query the most similar text from the collection
print ( result ) from embedin import Embedin
url = 'sqlite:///test.db'
client = Embedin ( collection_name = "test_collection" , texts = [ "This is a test" , "Hello world!" ], url = url )
result = client . query ( "These are tests" , top_k = 1 ) cd docker
docker-compose up embedin-postgresexemple
import os
from embedin import Embedin
url = os . getenv ( 'EMBEDIN_POSGRES_URL' , "postgresql+psycopg2://embedin:embedin@localhost/embedin_db" )
client = Embedin ( collection_name = "test_collection" , texts = [ "This is a test" , "Hello world!" ], url = url )
result = client . query ( "These are tests" , top_k = 1 ) cd docker
docker-compose up embedin-mysqlexemple
import os
from embedin import Embedin
url = os . getenv ( 'EMBEDIN_MYSQL_URL' , "mysql+pymysql://embedin:embedin@localhost/embedin_db" )
client = Embedin ( collection_name = "test_collection" , texts = [ "This is a test" , "Hello world!" ], url = url )
result = client . query ( "These are tests" , top_k = 1 ) cd docker
docker-compose up embedin-mssqlexemple
import os
from embedin import Embedin
url = os . getenv ( 'EMBEDIN_MSSQL_URL' , "mssql+pymssql://sa:StrongPassword123@localhost/tempdb" )
client = Embedin ( collection_name = "test_collection" , url = url )
client . add_data ( texts = [ "This is a test" ], meta_data = [{ "source" : "abc4" }])
result = client . query ( "These are tests" , top_k = 1 )Veuillez référer le guide des contributeurs