10h30/wordpress-export-to-markdown
1
0
Fork 0
mirror of https://github.com/10h30/wordpress-export-to-markdown.git synced 2026-06-05 15:09:59 +09:00
Code Issues Packages Projects Releases Wiki Activity

Split out code for writer and translator

Browse Source
This commit is contained in:
Will Boyd
2019-12-17 13:52:09 -05:00
parent f9e2bc5b0d
commit f4ae769f13
6 changed files with 213 additions and 202 deletions
Download Patch File Download Diff File Expand all files Collapse all files
index.js
+2 -110
View File
@@ -1,11 +1,6 @@
const fs = require('fs');
const luxon = require('luxon');
const path = require('path');
const request = require('request');
const shared = require('./src/shared');
const wizard = require('./src/wizard');
const parser = require('./src/parser');
const writer = require('./src/writer');
// global so various functions can access arguments
let config;
@@ -14,115 +9,12 @@ async function init() {
try {
config = wizard.getConfig();
let posts = await parser.parseFilePromise(config)
writeFiles(posts);
writer.writeFiles(posts, config);
} catch (ex) {
// appease the UnhandledPromiseRejectionWarning
console.error(ex);
}
}
function writeFiles(posts) {
let delay = 0;
posts.forEach(post => {
const postDir = getPostDir(post);
createDir(postDir);
writeMarkdownFile(post, postDir);
if (config.saveimages && post.meta.imageUrls) {
post.meta.imageUrls.forEach(imageUrl => {
const imageDir = path.join(postDir, 'images');
createDir(imageDir);
writeImageFile(imageUrl, imageDir, delay);
delay += 25;
});
}
});
}
function writeMarkdownFile(post, postDir) {
const frontmatter = Object.entries(post.frontmatter)
.reduce((accumulator, pair) => {
return accumulator + pair[0] + ': "' + pair[1] + '"\n'
}, '');
const data = '---\n' + frontmatter + '---\n\n' + post.content + '\n';
const postPath = path.join(postDir, getPostFilename(post));
fs.writeFile(postPath, data, (err) => {
if (err) {
console.log('Unable to write file.')
console.log(err);
} else {
console.log('Wrote ' + postPath + '.');
}
});
}
function writeImageFile(imageUrl, imageDir, delay) {
let imagePath = path.join(imageDir, shared.getFilenameFromUrl(imageUrl));
let stream = fs.createWriteStream(imagePath);
stream.on('finish', () => {
console.log('Saved ' + imagePath + '.');
});
// stagger image requests so we don't piss off hosts
setTimeout(() => {
request
.get(imageUrl)
.on('response', response => {
if (response.statusCode !== 200) {
console.log('Response status code ' + response.statusCode + ' received for ' + imageUrl + '.');
}
})
.on('error', err => {
console.log('Unable to download image.');
console.log(err);
})
.pipe(stream);
}, delay);
}
function createDir(dir) {
try {
fs.accessSync(dir, fs.constants.F_OK);
} catch (ex) {
fs.mkdirSync(dir, { recursive: true });
}
}
function getPostDir(post) {
let dir = config.output;
let dt = luxon.DateTime.fromISO(post.frontmatter.date);
if (config.yearmonthfolders) {
dir = path.join(dir, dt.toFormat('yyyy'), dt.toFormat('LL'));
} else if (config.yearfolders) {
dir = path.join(dir, dt.toFormat('yyyy'));
}
if (config.postfolders) {
let folder = post.meta.slug;
if (config.prefixdate) {
folder = dt.toFormat('yyyy-LL-dd') + '-' + folder;
}
dir = path.join(dir, folder);
}
return dir;
}
function getPostFilename(post) {
if (config.postfolders) {
// the containing folder name will be unique, just use index.md here
return 'index.md';
} else {
let filename = post.meta.slug + '.md';
if (config.prefixdate) {
let dt = luxon.DateTime.fromISO(post.frontmatter.date);
filename = dt.toFormat('yyyy-LL-dd') + '-' + filename;
}
return filename;
}
}
// it's go time!
init();
src/parser.js
+4 -90
View File
@@ -1,9 +1,9 @@
const fs = require('fs');
const luxon = require('luxon');
const turndown = require('turndown');
const xml2js = require('xml2js');
const shared = require('./shared');
const translator = require('./translator');
let config;
@@ -76,7 +76,7 @@ function addContentImages(data, images) {
function collectPosts(data) {
// this is passed into getPostContent() for the markdown conversion
turndownService = initTurndownService();
turndownService = translator.initTurndownService();
return getItemsOfType(data, 'post')
.map(post => ({
@@ -90,65 +90,10 @@ function collectPosts(data) {
title: getPostTitle(post),
date: getPostDate(post)
},
content: getPostContent(post, turndownService)
content: translator.getPostContent(post, turndownService, config)
}));
}
function initTurndownService() {
let turndownService = new turndown({
headingStyle: 'atx',
bulletListMarker: '-',
codeBlockStyle: 'fenced'
});
// preserve embedded tweets
turndownService.addRule('tweet', {
filter: node => node.nodeName === 'BLOCKQUOTE' && node.getAttribute('class') === 'twitter-tweet',
replacement: (content, node) => '\n\n' + node.outerHTML
});
// preserve embedded codepens
turndownService.addRule('codepen', {
filter: node => {
// codepen embed snippets have changed over the years
// but this series of checks should find the commonalities
return (
['P', 'DIV'].includes(node.nodeName) &&
node.attributes['data-slug-hash'] &&
node.getAttribute('class') === 'codepen'
);
},
replacement: (content, node) => '\n\n' + node.outerHTML
});
// preserve embedded scripts (for tweets, codepens, gists, etc.)
turndownService.addRule('script', {
filter: 'script',
replacement: (content, node) => {
let before = '\n\n';
let src = node.getAttribute('src');
if (node.previousSibling && node.previousSibling.nodeName !== '#text') {
// keep twitter and codepen <script> tags snug with the element above them
before = '\n';
}
let html = node.outerHTML.replace('async=""', 'async');
return before + html + '\n\n';
}
});
// preserve iframes (common for embedded audio/video)
turndownService.addRule('iframe', {
filter: 'iframe',
replacement: (content, node) => {
let html = node.outerHTML
.replace('allowfullscreen=""', 'allowfullscreen');
return '\n\n' + html + '\n\n';
}
});
return turndownService;
}
function getItemsOfType(data, type) {
return data.rss.channel[0].item.filter(item => item.post_type[0] === type);
}
@@ -176,37 +121,6 @@ function getPostDate(post) {
return luxon.DateTime.fromRFC2822(post.pubDate[0], { zone: 'utc' }).toISODate();
}
function getPostContent(post, turndownService) {
let content = post.encoded[0].trim();
// insert an empty div element between double line breaks
// this nifty trick causes turndown to keep adjacent paragraphs separated
// without mucking up content inside of other elemnts (like <code> blocks)
content = content.replace(/(\r?\n){2}/g, '\n<div></div>\n');
if (config.addcontentimages) {
// writeImageFile() will save all content images to a relative /images
// folder so update references in post content to match
content = content.replace(/(<img[^>]*src=").*?([^\/"]+\.(?:gif|jpg|png))("[^>]*>)/gi, '$1images/$2$3');
}
// this is a hack to make <iframe> nodes non-empty by inserting a "." which
// allows the iframe rule declared in initTurndownService() to take effect
// (using turndown's blankRule() and keep() solution did not work for me)
content = content.replace(/(<\/iframe>)/gi, '.$1');
// use turndown to convert HTML to Markdown
content = turndownService.turndown(content);
// clean up extra spaces in list items
content = content.replace(/(-|\d+\.) +/g, '$1 ');
// clean up the "." from the iframe hack above
content = content.replace(/\.(<\/iframe>)/gi, '$1');
return content;
}
function mergeImagesIntoPosts(images, posts) {
// create lookup table for quicker traversal
let postsLookup = posts.reduce((lookup, post) => {
@@ -229,4 +143,4 @@ function mergeImagesIntoPosts(images, posts) {
});
}
exports.parseFilePromise = parseFilePromise;
exports.parseFilePromise = parseFilePromise;
src/shared.js
+1 -1
View File
@@ -2,4 +2,4 @@ function getFilenameFromUrl(url) {
return url.split('/').slice(-1)[0];
}
exports.getFilenameFromUrl = getFilenameFromUrl;
exports.getFilenameFromUrl = getFilenameFromUrl;
src/translator.js
+90
View File
@@ -0,0 +1,90 @@
const turndown = require('turndown');
function initTurndownService() {
let turndownService = new turndown({
headingStyle: 'atx',
bulletListMarker: '-',
codeBlockStyle: 'fenced'
});
// preserve embedded tweets
turndownService.addRule('tweet', {
filter: node => node.nodeName === 'BLOCKQUOTE' && node.getAttribute('class') === 'twitter-tweet',
replacement: (content, node) => '\n\n' + node.outerHTML
});
// preserve embedded codepens
turndownService.addRule('codepen', {
filter: node => {
// codepen embed snippets have changed over the years
// but this series of checks should find the commonalities
return (
['P', 'DIV'].includes(node.nodeName) &&
node.attributes['data-slug-hash'] &&
node.getAttribute('class') === 'codepen'
);
},
replacement: (content, node) => '\n\n' + node.outerHTML
});
// preserve embedded scripts (for tweets, codepens, gists, etc.)
turndownService.addRule('script', {
filter: 'script',
replacement: (content, node) => {
let before = '\n\n';
let src = node.getAttribute('src');
if (node.previousSibling && node.previousSibling.nodeName !== '#text') {
// keep twitter and codepen <script> tags snug with the element above them
before = '\n';
}
let html = node.outerHTML.replace('async=""', 'async');
return before + html + '\n\n';
}
});
// preserve iframes (common for embedded audio/video)
turndownService.addRule('iframe', {
filter: 'iframe',
replacement: (content, node) => {
let html = node.outerHTML
.replace('allowfullscreen=""', 'allowfullscreen');
return '\n\n' + html + '\n\n';
}
});
return turndownService;
}
function getPostContent(post, turndownService, config) {
let content = post.encoded[0].trim();
// insert an empty div element between double line breaks
// this nifty trick causes turndown to keep adjacent paragraphs separated
// without mucking up content inside of other elemnts (like <code> blocks)
content = content.replace(/(\r?\n){2}/g, '\n<div></div>\n');
if (config.addcontentimages) {
// writeImageFile() will save all content images to a relative /images
// folder so update references in post content to match
content = content.replace(/(<img[^>]*src=").*?([^\/"]+\.(?:gif|jpg|png))("[^>]*>)/gi, '$1images/$2$3');
}
// this is a hack to make <iframe> nodes non-empty by inserting a "." which
// allows the iframe rule declared in initTurndownService() to take effect
// (using turndown's blankRule() and keep() solution did not work for me)
content = content.replace(/(<\/iframe>)/gi, '.$1');
// use turndown to convert HTML to Markdown
content = turndownService.turndown(content);
// clean up extra spaces in list items
content = content.replace(/(-|\d+\.) +/g, '$1 ');
// clean up the "." from the iframe hack above
content = content.replace(/\.(<\/iframe>)/gi, '$1');
return content;
}
exports.initTurndownService = initTurndownService;
exports.getPostContent = getPostContent;
src/wizard.js
+1 -1
View File
@@ -45,4 +45,4 @@ function checkFileExists(path) {
}
}
exports.getConfig = getConfig;
exports.getConfig = getConfig;
src/writer.js
+115
View File
@@ -0,0 +1,115 @@
const fs = require('fs');
const luxon = require('luxon');
const path = require('path');
const request = require('request');
const shared = require('./shared');
let config;
function writeFiles(posts, configIn) {
config = configIn;
let delay = 0;
posts.forEach(post => {
const postDir = getPostDir(post);
createDir(postDir);
writeMarkdownFile(post, postDir);
if (config.saveimages && post.meta.imageUrls) {
post.meta.imageUrls.forEach(imageUrl => {
const imageDir = path.join(postDir, 'images');
createDir(imageDir);
writeImageFile(imageUrl, imageDir, delay);
delay += 25;
});
}
});
}
function writeMarkdownFile(post, postDir) {
const frontmatter = Object.entries(post.frontmatter)
.reduce((accumulator, pair) => {
return accumulator + pair[0] + ': "' + pair[1] + '"\n'
}, '');
const data = '---\n' + frontmatter + '---\n\n' + post.content + '\n';
const postPath = path.join(postDir, getPostFilename(post));
fs.writeFile(postPath, data, (err) => {
if (err) {
console.log('Unable to write file.')
console.log(err);
} else {
console.log('Wrote ' + postPath + '.');
}
});
}
function writeImageFile(imageUrl, imageDir, delay) {
let imagePath = path.join(imageDir, shared.getFilenameFromUrl(imageUrl));
let stream = fs.createWriteStream(imagePath);
stream.on('finish', () => {
console.log('Saved ' + imagePath + '.');
});
// stagger image requests so we don't piss off hosts
setTimeout(() => {
request
.get(imageUrl)
.on('response', response => {
if (response.statusCode !== 200) {
console.log('Response status code ' + response.statusCode + ' received for ' + imageUrl + '.');
}
})
.on('error', err => {
console.log('Unable to download image.');
console.log(err);
})
.pipe(stream);
}, delay);
}
function createDir(dir) {
try {
fs.accessSync(dir, fs.constants.F_OK);
} catch (ex) {
fs.mkdirSync(dir, { recursive: true });
}
}
function getPostDir(post) {
let dir = config.output;
let dt = luxon.DateTime.fromISO(post.frontmatter.date);
if (config.yearmonthfolders) {
dir = path.join(dir, dt.toFormat('yyyy'), dt.toFormat('LL'));
} else if (config.yearfolders) {
dir = path.join(dir, dt.toFormat('yyyy'));
}
if (config.postfolders) {
let folder = post.meta.slug;
if (config.prefixdate) {
folder = dt.toFormat('yyyy-LL-dd') + '-' + folder;
}
dir = path.join(dir, folder);
}
return dir;
}
function getPostFilename(post) {
if (config.postfolders) {
// the containing folder name will be unique, just use index.md here
return 'index.md';
} else {
let filename = post.meta.slug + '.md';
if (config.prefixdate) {
let dt = luxon.DateTime.fromISO(post.frontmatter.date);
filename = dt.toFormat('yyyy-LL-dd') + '-' + filename;
}
return filename;
}
}
exports.writeFiles = writeFiles;