summaryrefslogtreecommitdiffstatshomepage
path: root/frontend/src/lib/Scraper.ts
diff options
context:
space:
mode:
authorWolfgang Müller2024-03-05 18:08:09 +0100
committerWolfgang Müller2024-03-05 19:25:59 +0100
commitd1d654ebac2d51e3841675faeb56480e440f622f (patch)
tree56ef123c1a15a10dfd90836e4038e27efde950c6 /frontend/src/lib/Scraper.ts
downloadhircine-d1d654ebac2d51e3841675faeb56480e440f622f.tar.gz
Initial commit0.1.0
Diffstat (limited to '')
-rw-r--r--frontend/src/lib/Scraper.ts156
1 files changed, 156 insertions, 0 deletions
diff --git a/frontend/src/lib/Scraper.ts b/frontend/src/lib/Scraper.ts
new file mode 100644
index 0000000..4baf370
--- /dev/null
+++ b/frontend/src/lib/Scraper.ts
@@ -0,0 +1,156 @@
+import {
+ Category,
+ Censorship,
+ Direction,
+ Language,
+ Layout,
+ OnMissing,
+ Rating,
+ type FullComicFragment,
+ type ScrapedComic,
+ type UpsertComicInput,
+ type UpsertOptions
+} from '$gql/graphql';
+import {
+ CategoryLabel,
+ CensorshipLabel,
+ DirectionLabel,
+ LanguageLabel,
+ LayoutLabel,
+ RatingLabel
+} from '$lib/Enums';
+import { getContext, setContext } from 'svelte';
+import { writable, type Writable } from 'svelte/store';
+
+interface ScraperContext {
+ scraper: string;
+ warnings: string[];
+ selector?: ScrapedComicSelector;
+}
+
+export function initScraperContext() {
+ return setContext<Writable<ScraperContext>>('scraper', writable({ scraper: '', warnings: [] }));
+}
+
+export function getScraperContext() {
+ return getContext<Writable<ScraperContext>>('scraper');
+}
+
+export class Selector<T extends string> {
+ keep = true;
+ value: T;
+ display: string | undefined;
+
+ constructor(value: T, display?: string) {
+ this.value = value;
+ this.display = display;
+ }
+
+ toString() {
+ return this.display ?? this.value;
+ }
+
+ static from<T extends string>(
+ scraped: T | undefined | null,
+ have: string | undefined | null,
+ label?: Record<string, string>
+ ) {
+ if (scraped && have !== scraped) {
+ return new Selector(scraped, label ? label[scraped] : undefined);
+ }
+ return undefined;
+ }
+
+ static fromList(scraped: string[], have: { name: string }[]) {
+ const haves = new Set(have.map((i) => i.name));
+
+ return scraped.filter((i) => !haves.has(i)).map((i) => new Selector(i));
+ }
+}
+
+function keepItem<T extends string>(selector?: Selector<T>): T | undefined | null {
+ if (selector?.keep) {
+ return selector.value;
+ }
+ return undefined;
+}
+
+function keepList<T extends string>(
+ selectorList: Selector<T>[],
+ onMissing: OnMissing
+): { names: T[]; options: UpsertOptions } {
+ return {
+ names: selectorList.filter((v) => v.keep).map((v) => v.value),
+ options: { onMissing }
+ };
+}
+
+export class ScrapedComicSelector {
+ title?: Selector<string>;
+ originalTitle?: Selector<string>;
+ url?: Selector<string>;
+ date?: Selector<string>;
+ category?: Selector<Category>;
+ censorship?: Selector<Censorship>;
+ rating?: Selector<Rating>;
+ language?: Selector<Language>;
+ direction?: Selector<Direction>;
+ layout?: Selector<Layout>;
+ artists: Selector<string>[];
+ circles: Selector<string>[];
+ characters: Selector<string>[];
+ worlds: Selector<string>[];
+ tags: Selector<string>[];
+
+ constructor(scraped: ScrapedComic, comic: FullComicFragment) {
+ this.title = Selector.from(scraped.title, comic.title);
+ this.originalTitle = Selector.from(scraped.originalTitle, comic.originalTitle);
+ this.url = Selector.from(scraped.url, comic.url);
+ this.date = Selector.from(scraped.date, comic.date);
+ this.category = Selector.from(scraped.category, comic.category, CategoryLabel);
+ this.censorship = Selector.from(scraped.censorship, comic.censorship, CensorshipLabel);
+ this.rating = Selector.from(scraped.rating, comic.rating, RatingLabel);
+ this.language = Selector.from(scraped.language, comic.language, LanguageLabel);
+ this.direction = Selector.from(scraped.direction, comic.direction, DirectionLabel);
+ this.layout = Selector.from(scraped.layout, comic.layout, LayoutLabel);
+
+ this.artists = Selector.fromList(scraped.artists, comic.artists);
+ this.circles = Selector.fromList(scraped.circles, comic.circles);
+ this.characters = Selector.fromList(scraped.characters, comic.characters);
+ this.tags = Selector.fromList(scraped.tags, comic.tags);
+ this.worlds = Selector.fromList(scraped.worlds, comic.worlds);
+ }
+
+ hasData() {
+ return (
+ Object.values(this).filter((i) => {
+ if (i === undefined) {
+ return false;
+ } else if (Array.isArray(i) && i.length === 0) {
+ return false;
+ }
+ return true;
+ }).length > 0
+ );
+ }
+
+ toInput(onMissing: OnMissing): UpsertComicInput {
+ return {
+ title: keepItem(this.title),
+ originalTitle: keepItem(this.originalTitle),
+ url: keepItem(this.url),
+ date: keepItem(this.date),
+ category: keepItem(this.category),
+ censorship: keepItem(this.censorship),
+ rating: keepItem(this.rating),
+ language: keepItem(this.language),
+ direction: keepItem(this.direction),
+ layout: keepItem(this.layout),
+ artists: keepList(this.artists, onMissing),
+ circles: keepList(this.circles, onMissing),
+ characters: keepList(this.characters, onMissing),
+ worlds: keepList(this.worlds, onMissing),
+ tags: keepList(this.tags, onMissing)
+ };
+ }
+}