import {request} from "obsidian";
import {RssFeed} from "../settings/settings";
import {Md5} from "ts-md5";
/**
* parser for .rss files, build from scratch
* because I could not find a parser that
* - works on mobile
* - is up-to-date
* - works for multiple different interpretations of the rss spec
*/
export interface RssFeedContent {
subtitle: string,
title: string,
name: string,
link: string,
image: string,
folder: string,
description: string,
language: string,
hash: string,
items: RssFeedItem[]
}
export interface RssFeedItem {
title: string,
description: string,
content: string,
category: string,
link: string,
creator: string,
language: string,
enclosure: string,
enclosureType: string,
image: string,
pubDate: string,
folder: string,
feed: string,
favorite: boolean,
read: boolean,
created: boolean,
tags: string[],
hash: string,
id: string,
highlights: string[],
}
/**
* return the node with the specified name
* : to get namespaced element
* . to get nested element
* @param element
* @param name
*/
function getElementByName(element: Element | Document, name: string): ChildNode {
let value: ChildNode;
if (typeof element.getElementsByTagName !== 'function' && typeof element.getElementsByTagNameNS !== 'function') {
//the required methods do not exist on element, aborting
return;
}
if (name.includes(":")) {
const [namespace, tag] = name.split(":");
const namespaceUri = element.lookupNamespaceURI(namespace);
const byNamespace = element.getElementsByTagNameNS(namespaceUri, tag);
if (byNamespace.length > 0) {
value = byNamespace[0].childNodes[0];
} else {
//there is no element in that namespace, probably because no namespace has been defined
const tmp = element.getElementsByTagName(name);
if (tmp.length > 0) {
if (tmp[0].childNodes.length === 0) {
value = tmp[0];
} else {
const node = tmp[0].childNodes[0];
if (node !== undefined) {
value = node;
}
}
}
}
} else if (name.includes(".")) {
const [prefix, tag] = name.split(".");
if (element.getElementsByTagName(prefix).length > 0) {
const nodes = Array.from(element.getElementsByTagName(prefix)[0].childNodes);
nodes.forEach((node) => {
if (node.nodeName == tag) {
value = node;
}
});
}
} else if (element.getElementsByTagName(name).length > 0) {
if (element.getElementsByTagName(name)[0].childNodes.length == 0) {
value = element.getElementsByTagName(name)[0];
} else {
const node = element.getElementsByTagName(name)[0].childNodes[0];
if (node !== undefined)
value = node;
}
}
//if(name === "content") console.log(value);
return value;
}
/**
* # to get attribute
* Always returns the last found value for names
* @param element
* @param names possible names
*/
function getContent(element: Element | Document, names: string[]): string {
let value: string;
for (const name of names) {
if (name.includes("#")) {
const [elementName, attr] = name.split("#");
const data = getElementByName(element, elementName);
if (data) {
if (data.nodeName === elementName) {
//@ts-ignore
const tmp = data.getAttribute(attr);
if (tmp.length > 0) {
value = tmp;
}
}
}
} else {
const data = getElementByName(element, name);
if (data) {
//@ts-ignore
if(data.wholeText && data.wholeText.length > 0) {
//@ts-ignore
value = data.wholeText;
}
//@ts-ignore
if (!value && data.nodeValue && data.nodeValue.length > 0) {
value = data.nodeValue;
}
//@ts-ignore
if (!value && data.innerHTML && data.innerHTML.length > 0) {
//@ts-ignore
value = data.innerHTML;
}
}
}
}
if (value === undefined) {
return "";
}
return value;
}
function buildItem(element: Element): RssFeedItem {
return {
title: getContent(element, ["title"]),
description: getContent(element, ["content", "content:encoded", "itunes:summary", "description", "summary", "media:description"]),
content: getContent(element, ["itunes:summary", "description", "summary", "media:description", "content", "content:encoded", "ns0:encoded"]),
category: getContent(element, ["category"]),
link: getContent(element, ["link", "link#href"]),
creator: getContent(element, ["creator", "dc:creator", "author", "author.name"]),
pubDate: getContent(element, ["pubDate", "published", "updated", "dc:date"]),
enclosure: getContent(element, ["enclosure#url", "yt:videoId"]),
enclosureType: getContent(element, ["enclosure#type"]),
image: getContent(element, ["enclosure#url", "media:content#url", "itunes:image#href", "media:thumbnail#url"]),
id: getContent(element, ["id"]),
language: null,
folder: null,
feed: null,
read: null,
favorite: null,
created: null,
tags: [],
hash: null,
highlights: []
}
}
function getAllItems(doc: Document): Element[] {
const items: Element[] = [];
if (doc.getElementsByTagName("item")) {
for (const elementsByTagNameKey in doc.getElementsByTagName("item")) {
const entry = doc.getElementsByTagName("item")[elementsByTagNameKey];
items.push(entry);
}
}
if (doc.getElementsByTagName("entry")) {
for (const elementsByTagNameKey in doc.getElementsByTagName("entry")) {
const entry = doc.getElementsByTagName("entry")[elementsByTagNameKey];
items.push(entry);
}
}
return items;
}
async function requestFeed(feed: RssFeed) : Promise<string> {
return await request({url: feed.url});
}
export async function getFeedItems(feed: RssFeed): Promise<RssFeedContent> {
let data;
try {
const rawData = await requestFeed(feed);
data = new window.DOMParser().parseFromString(rawData, "text/xml");
console.log(data);
} catch (e) {
console.error(e);
return Promise.resolve(undefined);
}
const items: RssFeedItem[] = [];
const rawItems = getAllItems(data);
const language = getContent(data, ["language"]).substr(0, 2);
rawItems.forEach((rawItem) => {
const item = buildItem(rawItem);
if (item.title !== undefined && item.title.length !== 0) {
item.folder = feed.folder;
item.feed = feed.name;
item.read = false;
item.favorite = false;
item.created = false;
item.language = language;
item.hash = <string>new Md5().appendStr(item.title).appendStr(item.folder).appendStr(item.link).end();
if (!item.image && feed.url.includes("youtube.com/feeds")) {
item.image = "https://i3.ytimg.com/vi/" + item.id.split(":")[2] + "/hqdefault.jpg";
}
items.push(item);
}
})
const image = getContent(data, ["image", "image.url", "icon"]);
const content: RssFeedContent = {
title: getContent(data, ["title"]),
subtitle: getContent(data, ["subtitle"]),
link: getContent(data, ["link"]),
//we don't want any leading or trailing slashes in image urls(i.e. reddit does that)
image: image ? image.replace(/^\/|\/$/g, '') : null,
description: getContent(data, ["description"]),
items: items,
folder: feed.folder,
name: feed.name,
language: language,
hash: "",
};
return Promise.resolve(content);
}
