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

feat: less brittle non unique selectors

Browse Source
This commit is contained in:
karishmas6
2024-08-07 17:55:36 +05:30
parent e113799029
commit a7e4200677
1 changed files with 13 additions and 8 deletions
Download Patch File Download Diff File Expand all files Collapse all files

21
server/src/workflow-management/selector.ts
View File

@@ -730,37 +730,41 @@ export const getSelectors = async (page: Page, coordinates: Coordinates) => {
* @returns {Promise<Selectors|null|undefined>} * @returns {Promise<Selectors|null|undefined>}
*/ */
export const getNonUniqueSelectors = async (page: Page, coordinates: Coordinates) => { export const getNonUniqueSelectors = async (page: Page, coordinates: Coordinates) => {
try { try {
const selectors = await page.evaluate(({ x, y }) => { const selectors = await page.evaluate(({ x, y }) => {
function getSelector(element: Element): string { function getSelector(element: any) {
let selector = element.tagName.toLowerCase(); let selector = element.tagName.toLowerCase();
// Capture a single, relevant class if present
if (element.className) { if (element.className) {
const classes = element.className.split(/\s+/).filter(Boolean); const classes = element.className.split(/\s+/).filter(Boolean);
if (classes.length > 0) { if (classes.length > 0) {
selector += '.' + classes.join('.'); // Use only the first class to avoid over-specificity
selector += '.' + classes[0];
} }
} }
return selector; return selector;
} }
function getSelectorPath(element: Element): string { function getSelectorPath(element: any) {
const path = []; const path = [];
while (element && element !== document.body) { while (element && element !== document.body) {
path.unshift(getSelector(element)); const selector = getSelector(element);
element = element.parentElement!; path.unshift(selector);
element = element.parentElement;
} }
return path.join(' > '); return path.join(' > ');
} }
const element = document.elementFromPoint(x, y) as Element; const element = document.elementFromPoint(x, y);
if (!element) return null; if (!element) return null;
const generalSelector = getSelectorPath(element); const generalSelector = getSelectorPath(element);
return { return {
generalSelector, generalSelector,
}; };
}, coordinates); }, coordinates);
return selectors || {}; return selectors || {};
@@ -770,6 +774,7 @@ export const getNonUniqueSelectors = async (page: Page, coordinates: Coordinates
} }
}; };
/** /**
* Returns the first pair from the given workflow that contains the given selector * Returns the first pair from the given workflow that contains the given selector
* inside the where condition, and it is the only selector there. * inside the where condition, and it is the only selector there.