vodorod/parcer
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

feat: hierarchical selectors

Browse Source
This commit is contained in:
karishmas6
2024-09-03 05:49:47 +05:30
parent 9843116cbf
commit 0dfb3f6673
1 changed files with 21 additions and 9 deletions
Download Patch File Download Diff File Expand all files Collapse all files

30
server/src/workflow-management/selector.ts
View File

@@ -784,12 +784,14 @@ export const getNonUniqueSelectors = async (page: Page, coordinates: Coordinates
return { generalSelector: '' }; return { generalSelector: '' };
} }
}; };
interface SelectorNode {
selector: string;
children: SelectorNode[];
}
export const getChildSelectors = async (page: Page, parentSelector: string, maxDepth: number = 2): Promise<SelectorNode[]> => {
export const getChildSelectors = async (page: Page, parentSelector: string): Promise<string[]> => {
try { try {
const childSelectors = await page.evaluate((parentSelector: string) => { const childSelectors = await page.evaluate(({ parentSelector, maxDepth }: { parentSelector: string, maxDepth: number }) => {
function getNonUniqueSelector(element: HTMLElement): string { function getNonUniqueSelector(element: HTMLElement): string {
let selector = element.tagName.toLowerCase(); let selector = element.tagName.toLowerCase();
@@ -809,7 +811,6 @@ export const getChildSelectors = async (page: Page, parentSelector: string): Pro
function getSelectorPath(element: HTMLElement | null): string { function getSelectorPath(element: HTMLElement | null): string {
const path: string[] = []; const path: string[] = [];
let depth = 0; let depth = 0;
const maxDepth = 2;
while (element && element !== document.body && depth < maxDepth) { while (element && element !== document.body && depth < maxDepth) {
const selector = getNonUniqueSelector(element); const selector = getNonUniqueSelector(element);
@@ -821,12 +822,23 @@ export const getChildSelectors = async (page: Page, parentSelector: string): Pro
return path.join(' > '); return path.join(' > ');
} }
const parentElement = document.querySelector(parentSelector); function getDescendantSelectors(element: HTMLElement, currentDepth: number = 0): SelectorNode[] {
if (currentDepth >= maxDepth) return [];
const children = Array.from(element.children) as HTMLElement[];
return children.map(child => {
return {
selector: getSelectorPath(child),
children: getDescendantSelectors(child, currentDepth + 1)
};
});
}
const parentElement = document.querySelector(parentSelector) as HTMLElement;
if (!parentElement) return []; if (!parentElement) return [];
const childElements = Array.from(parentElement.children) as HTMLElement[]; // Type assertion to HTMLElement[] return getDescendantSelectors(parentElement);
return childElements.map(child => getSelectorPath(child)); }, { parentSelector, maxDepth });
}, parentSelector);
return childSelectors || []; return childSelectors || [];
} catch (error) { } catch (error) {