Archives/mercury-parser
2
0
Fork 0
mirror of https://github.com/postlight/mercury-parser synced 2024-11-17 03:25:31 +00:00
Code Issues Projects Releases Wiki Activity

fix: duplicate key bug

Browse Source
This commit is contained in:
Adam Pash 2016-09-08 12:15:45 -04:00
parent 93ca688955
commit 6c6451b34b
1 changed files with 14 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

19
src/extractor/root-extractor.js
View File

@ -12,13 +12,14 @@ const RootExtractor = {
const title = extract({ ...opts, type: 'title', extractor }) const title = extract({ ...opts, type: 'title', extractor })
const datePublished = extract({ ...opts, type: 'datePublished', extractor }) const datePublished = extract({ ...opts, type: 'datePublished', extractor })
const author = extract({ ...opts, type: 'author', extractor }) const author = extract({ ...opts, type: 'author', extractor })
const content = extract({ ...opts, type: 'content', extractor, html: true }) const content = extract({ ...opts, type: 'content', extractor, extractHtml: true })
const leadImageUrl = extract({ ...opts, type: 'leadImageUrl', extractor }) const leadImageUrl = extract({ ...opts, type: 'leadImageUrl', extractor })
const dek = extract({ ...opts, type: 'dek', extractor }) const dek = extract({ ...opts, type: 'dek', extractor })
return { return {
title, title,
content, content,
author,
datePublished, datePublished,
leadImageUrl, leadImageUrl,
dek, dek,
@ -27,15 +28,15 @@ const RootExtractor = {
} }
function extract(opts) { function extract(opts) {
const { type, extractor, $, html } = opts const { type, extractor, $, extractHtml } = opts
// If nothing matches the selector, // If nothing matches the selector,
// run the Generic extraction // run the Generic extraction
return select($, extractor[type], html) || return select($, extractor[type], extractHtml) ||
GenericExtractor[type](opts) GenericExtractor[type](opts)
} }
function select($, extractionOpts, html=false) { function select($, extractionOpts, extractHtml=false) {
// Skip if there's not extraction for this type // Skip if there's not extraction for this type
if (!extractionOpts) return if (!extractionOpts) return
@ -44,12 +45,20 @@ function select($, extractionOpts, html=false) {
const matchingSelector = selectors.find((selector) => { const matchingSelector = selectors.find((selector) => {
return $(selector).length === 1 return $(selector).length === 1
}) })
console.log(matchingSelector)
// console.log($(matchingSelector).text())
console.log(extractHtml)
if (!matchingSelector) return if (!matchingSelector) return
// If the selector type requests html as its return type // If the selector type requests html as its return type
// clean the element with provided cleaning selectors // clean the element with provided cleaning selectors
if (html) { if (extractHtml) {
let $content = $(matchingSelector) let $content = $(matchingSelector)
// Wrap in div so transformation can take place on root element
$content.wrap($('<div></div>'))
$content = $content.parent()
$content = cleanBySelectors($content, $, extractionOpts) $content = cleanBySelectors($content, $, extractionOpts)
$content = transformElements($content, $, extractionOpts) $content = transformElements($content, $, extractionOpts)