FOSS
/
urltomarkdown
mirror of https://github.com/macsplit/urltomarkdown
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

rollback code-block enhancement due to issues downstream with ToMarkdown

main
macsplit 2022-05-13 17:47:42 +01:00
parent 4c66e91b73
commit b73a9cef1b
1 changed files with 5 additions and 34 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

39
index.js
View File

@ -8,7 +8,6 @@ const table_to_markdown = require('./html_table_to_markdown.js');
const validURL = require('@7c/validurl'); const validURL = require('@7c/validurl');
const express = require('express'); const express = require('express');
const rateLimit = require('express-rate-limit'); const rateLimit = require('express-rate-limit');
const htmlentities = require('html-entities');
const port = process.env.PORT; const port = process.env.PORT;
@ -72,14 +71,14 @@ app.post('/', function(req, res) {
if (!html) { if (!html) {
res.status(400).send("Please provide a POST parameter called html"); res.status(400).send("Please provide a POST parameter called html");
} else { } else {
//try { try {
let document = new JSDOM(html); let document = new JSDOM(html);
let markdown = process_dom(url, document, res, inline_title, ignore_links); let markdown = process_dom(url, document, res, inline_title, ignore_links);
send_headers(res); send_headers(res);
res.send(markdown); res.send(markdown);
//} catch (error) { } catch (error) {
// res.status(400).send("Could not parse that document"); res.status(400).send("Could not parse that document");
//} }
} }
}); });
@ -100,8 +99,7 @@ function process_dom(url, document, res, inline_title, ignore_links) {
let reader = new Readability(document.window.document); let reader = new Readability(document.window.document);
let readable = reader.parse().content; let readable = reader.parse().content;
let replacements = [] let replacements = []
readable = format_tables(readable, replacements); readable = format_tables(readable, replacements);
readable = format_code_blocks(readable, replacements);
let markdown = service.turndown(readable); let markdown = service.turndown(readable);
for (let i=0;i<replacements.length;i++) { for (let i=0;i<replacements.length;i++) {
markdown = markdown.replace(replacements[i].placeholder, replacements[i].replacement); markdown = markdown.replace(replacements[i].placeholder, replacements[i].replacement);
@ -152,30 +150,3 @@ function format_tables(html, replacements) {
return html; return html;
} }
function format_code_blocks(html, replacements) {
const start = replacements.length;
const code_blocks = html.match(/(<pre[^>]*>(?:.|\n)*?<\/pre>)/gi);
if (code_blocks) {
for (let cb=0;cb<code_blocks.length;cb++) {
let code_block = code_blocks[cb];
let markdown = code_block_to_markdown(code_block);
let placeholder = "urltomarkdowncodeblockplaceholder"+cb+Math.random();
replacements[start+cb] = { placeholder: placeholder, replacement: markdown};
html = html.replace(code_block, "<p>"+placeholder+"</p>");
}
}
return html;
}
function code_block_to_markdown (html) {
const match_pre = /^<pre[^>]*>([\s\S]*)<\/pre>$/ig.exec(html);
let inner_html = match_pre[1];
const match_code = /^\s*<code[^>]*>[\r\n]*([\s\S]*)<\/code>\s*$/ig.exec(inner_html);
if (match_code && match_code[1])
inner_html = match_code[1];
inner_html = inner_html.replaceAll(/<br[^>]*>/ig,"\n");
inner_html = inner_html.replaceAll(/<[^>]+>/ig, "");
inner_html = htmlentities.decode(inner_html);
const markdown = "```\n"+inner_html+"\n```\n";
return markdown;
}