html_layout.js

let a = []

function _add_empty(node, raw){
    // 如果多个标签合并成一个，最低保证合并的时候左右存在空格
    function inner(node){
        let innertext = node.innerText
        if (!innertext){
            return 
        }
        let content = innertext.trim()
        pos = raw.indexOf(content)
        if (raw.indexOf(' '+content) == -1 && raw.indexOf(content+' ') == -1){
            raw = raw.replace(content, ' ' + content + ' ')
        }
        var nodes = node.childNodes;
        for (var i = 0; i <nodes.length; i++){
            if(!nodes[i]){
                continue;
            }
    
            if(nodes[i].childNodes.length > 0){
                inner(nodes[i]);
            }
        }
    }
    inner(node)
    return raw
}

function check(node){
    // 返回true表示不再进行分割
    rect = node.getBoundingClientRect()
    if (node.hidden || (rect.width < 10 && rect.height < 10)){
        return true
    }
    let innertext = node.innerText
    if (!innertext){
        return true
    }
    let content = innertext.trim()
    if (node.childNodes.length == 0){
        if (content.length > 0 && content.split(' ').length > 3){
            a.push(content)
        }
        return true
    }

    if (content.includes('\n')){
        return false
    }else{
        if (content.length >0 && content.split(' ').length > 3){
            a.push(_add_empty(node, content))
        }
        return true
    }
}


function loop(node){
    if (check(node)){
        return
    }
    var nodes = node.childNodes;
    for (var i = 0; i <nodes.length; i++){
        if (nodes[i] && typeof nodes[i].getBoundingClientRect === 'function'){
            loop(nodes[i])
        }
    }
}

let s = 'body'
loop(document.querySelector(s))


console.log(a.join('\n'))