Wang Zhiyong

> 按照说明文档，启动了docker，端口号7090，但浏览器无法访问是什么原因？ docker pull tugraph/tugraph-db-centos7:3.3.0 docker run -d -p 7070:7070 --name tugraph_demo tugraph/tugraph-db-centos7:3.3.0 lgraph_server web 访问 http://x.x.x.x:7070 默认用户名为 `admin`，密码为 `73@TuGraph`

> 看了一下docker hub上。只有两个。没有详细的说明！ docker pull tugraph/tugraph-db-centos7:3.3.0 docker run -d -p 7070:7070 --name tugraph_demo tugraph/tugraph-db-centos7:3.3.0 lgraph_server web 访问 http://x.x.x.x:7070 默认用户名为 `admin`，密码为 `73@TuGraph`

> Yes, the build scripts need to be improved. Using the latest version of brpc is ok.

> Fulltext index is not supported in cypher statements and currently can only be used in plugins.

Thanks for your tips, these chinese comments will be removed.

> I have a future request for getting changed data capture from kafka streams when having to query real-time data being from other databases in TuGraph. > > I consider...

> Install java 1.8 in your os environment or use `cmake -DENABLE_FULLTEXT_INDEX=OFF ` to remove the dependency on java. BTW, it is recommended to use `centos` or `ubuntu` to compile,...

> The vector similarity searching is popular recently. > > And vector index can be implemented by lucene. > > https://arxiv.org/abs/2308.14963 > > https://neo4j.com/docs/cypher-manual/current/indexes-for-vector-search/#indexes-vector-similarity-cosine Thanks for your idea, we will...

使用neo4j driver 连接tugraph，直接发送cypher 语句 "match (n) return n" 就可以了。结果是流式返回的，不管多少数据都可以读出来，不会引发内存oom。

建议使用neo4j的java客户端： https://tugraph-db.readthedocs.io/zh-cn/latest/5.developer-manual/4.client-tools/6.bolt-client.html 之前自研的java client后续就不更新了，但对neo4j客户端的兼容性会持续更新。