一、服务端搭建
1.本地拉取容器镜像
|---------------------------------|
| docker pull chromadb``/chroma |
2.将本地容器镜像上传到远程服务器（如果远程服务器可以直接拉取镜像可以直接拉取，无需这一步）
|------------------------------------------------------------------------------------------------------------------|
| docker save -o chromadb.``tar chromadb``/chroma scp /Users/本地目录/chromadb``.``tar root@远程机器ip:``/远程机器路径 |
3.远程服务器加载并运行容器镜像
|--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
| docker load -i chromadb.``tar docker run --``rm --name chromadb -p 8001:8000 -``v .``/chromadb``:``/chroma/chroma --``env``-``file ./.chroma_env chromadb``/chroma |
注释：
-p 8001:8000 将服务器8001端口映射到容器8000端口-v ./chromadb:/chroma/chroma 将服务器存储路径./chromadb映射到容器路径/chroma/chroma--env-file ./.chroma_env 容器服务chromadb运行时的相关配置.chroma_env配置内容
|--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
| # 容器中数据是否持久化配置 IS_PERSISTENT=TRUE # 容器中向量数据存储路径 chroma-data=``/chroma/chroma/ # 服务端认证token CHROMA_SERVER_AUTHN_CREDENTIALS=``test``-token # 服务端认证方式 CHROMA_SERVER_AUTHN_PROVIDER=chromadb.auth.token_authn.TokenAuthenticationServerProvider |
二、客户端使用
需要安装依赖：pip install chromadb-client
|----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
| import chromadb import chromadb.utils.embedding_functions as embedding_functions import logging import sys from chromadb.config ``import Settings from dotenv ``import load_dotenv from datetime ``import datetime # 使用日志记录查看查询和事件 # log_format = '%(asctime)s - %(levelname)s - %(message)s' # logging.basicConfig(stream=sys.stdout, level=logging.INFO,format=log_format) # 获取向量数据库客户端 chroma_client ``= chromadb.HttpClient(host``=``'服务器ip'``, ``port``=``8001``, ``settings``=``Settings( ``chroma_client_auth_provider``=``"chromadb.auth.token_authn.TokenAuthClientProvider"``, ``chroma_client_auth_credentials``=``"test-token"``) ``) # 定义向量化函数（使用微软的azure） openai_ef ``= embedding_functions.OpenAIEmbeddingFunction( ``api_key``=``""``, ``api_base``=""``, ``api_type``=``"azure"``, ``api_version``=``""``, ``model_name``=``"embedding-3-small" ``) # 创建数据集 # collection = chroma_client.create_collection(name="my_collection", # embedding_function=openai_ef, # metadata={ # "description": "my first Chroma collection", # "created": str(datetime.now()) # } ) # 删除数据集 # chroma_client.delete_collection(name="my_collection") # 获取数据集 collection ``= chroma_client.get_collection(name``=``"my_collection"``,embedding_function``=``openai_ef) # 向数据集中增加数据 # collection.add( # documents=[ # "This is a document about pineapple", # "This is a document about hawai" # ], # metadatas=[{"chapter": "3", "verse": "16"}, {"chapter": "3", "verse": "5"}], # ids=["id1", "id2"] # ) # 文本查询数据 # results = collection.query( # query_texts=["This is a query document about hawai"], # Chroma will embed this for you # n_results=1 # how many results to return # ) # 文本查询并过滤 results ``= collection.query( ``query_texts``=``[``"This is a query document about hawai"``], ``n_results``=``2``, ``# metadatas 元数据过滤 ``# where={"verse": "16"}, ``# 文本内容过滤 ``where_document``=``{ ``"$contains"``:``"pineapple"``} ) print``(results) |