forkhroy / prvt-langchain4j

0 stars 0 forks source link

langchain4j-oracle/src/main/java/dev/langchain4j/data/document/splitter/oracle/OracleDocumentSplitter.java #2

Open forkhroy opened 1 month ago

forkhroy commented 1 month ago
  1. Splitter constructor is following a different pattern than Loader. It takes both conn and pref whereas Loader constructor takes only conn. Maybe we want to keep them consistent meaning both follow the same pattern.

    // splitter public OracleDocumentSplitter(Connection conn, String pref) { this.conn = conn; this.pref = pref; }

    // loader public OracleDocumentLoader(Connection conn) { this.conn = conn; }

  2. Can we see same samples from other providers?

  3. Any plural version for split?

hackerdave commented 3 weeks ago
  1. Tried to follow their convention where split only passes the text

https://github.com/langchain4j/langchain4j/blob/main/langchain4j/src/main/java/dev/langchain4j/data/document/splitter/DocumentByWordSplitter.java

protected abstract String[] split(String text)

  1. Their splitters are here

https://github.com/langchain4j/langchain4j/tree/main/langchain4j/src/main/java/dev/langchain4j/data/document/splitter

  1. Their base class is this. I didn't derive from it as it was pretty complicated.

https://github.com/langchain4j/langchain4j/blob/main/langchain4j/src/main/java/dev/langchain4j/data/document/splitter/HierarchicalDocumentSplitter.java

They provide these 2. We may also want to provide a Document version.

protected abstract String[] split(String text); public List split(Document document)