openai_embedding v1.0.13

openai_embedding

package to help inputting and storing long reference text, and extract information when asking prompt

Example usage:

//this will return all vector map in {text:{embedding:num[],tokensNum:num}}

let embedMap=await computeDocEmbeddings("hello, hows it going","come on!","long long long text long long long text long long long text long long long text" + " long long long text long long long text long long long text long long long text"+ " long long long text long long long text long long long text long long long text"+ " long long long text long long long text long long long text long long long text"+ " long long long text long long long text long long long text long long long text"+ " long long long text long long long text long long long text long long long text");

//pass the vector map and question text, then //this will return related context in string[]

await constructPrompt("How long is the text?",embedMap);

//this will return an array of text, separated in to coherent sections

intoParagraphs({raw:"some paragraphs text", maxtoken=400});