MickSlash
/
F95API
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

Add list check

pull/83/head
MillenniumEarl 2021-03-21 17:10:49 +01:00
parent d0e87e0ead
commit dcc5ed973f
1 changed files with 11 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

16
src/scripts/scrape-data/post-parse.ts
View File

@ -129,6 +129,13 @@ function isNoScriptNode(node: cheerio.Element): boolean {
return node.type === "tag" && node.name === "noscript"; return node.type === "tag" && node.name === "noscript";
} }
/**
* Check if the node is a list element, i.e. `<li>` or `<ul>` tag.
*/
function isListNode(node: cheerio.Element): boolean {
return node.type === "tag" && (node.name === "ul" || node.name === "li");
}
//#endregion Node Type //#endregion Node Type
//#region Parse Cheerio node //#region Parse Cheerio node
@ -353,10 +360,10 @@ function parseCheerioNode($: cheerio.Root, node: cheerio.Element): IPostElement
else if (isSpoilerNode(cheerioNode)) post = parseCheerioSpoilerNode($, cheerioNode); else if (isSpoilerNode(cheerioNode)) post = parseCheerioSpoilerNode($, cheerioNode);
else if (isLinkNode(node)) post = parseCheerioLinkNode(cheerioNode); else if (isLinkNode(node)) post = parseCheerioLinkNode(cheerioNode);
// Check for childrens only if the node is a <b>/<i> element. // Check for childrens only if the node is a <b>/<i> element
// For the link in unnecessary while for the spoilers is // or a list element. For the link in unnecessary while for
// already done in parseCheerioSpoilerNode // the spoilers is already done in parseCheerioSpoilerNode
if (isFormattingNode(node)) { if (isFormattingNode(node) || isListNode(node)) {
// Parse the node's childrens // Parse the node's childrens
const childPosts = cheerioNode const childPosts = cheerioNode
.contents() // @todo Change to children() after cheerio RC6 .contents() // @todo Change to children() after cheerio RC6
@ -380,7 +387,6 @@ function associateNameToElements(elements: IPostElement[]): IPostElement[] {
const pairs: IPostElement[] = []; const pairs: IPostElement[] = [];
const specialCharsRegex = /^[-!$%^&*()_+|~=`{}[\]:";'<>?,./]/; const specialCharsRegex = /^[-!$%^&*()_+|~=`{}[\]:";'<>?,./]/;
const specialRegex = new RegExp(specialCharsRegex); const specialRegex = new RegExp(specialCharsRegex);
const x = pairUp(elements);
for (let i = 0; i < elements.length; i++) { for (let i = 0; i < elements.length; i++) {
// If the text starts with a special char, clean it // If the text starts with a special char, clean it