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

feat: rm threshold logic

Browse Source
This commit is contained in:
Rohit
2025-03-20 01:40:51 +05:30
parent bad7255dc1
commit a5f8f9d64e
1 changed files with 0 additions and 102 deletions
Download Patch File Download Diff File Expand all files Collapse all files

102
server/src/workflow-management/selector.ts
View File

@@ -474,40 +474,6 @@ export const getElementInformation = async (
element = tableParent; element = tableParent;
} }
} }
if (element.tagName !== 'TABLE') {
while (element.parentElement) {
if (element.tagName.toLowerCase() === 'body' ||
element.tagName.toLowerCase() === 'html') {
break;
}
const parentRect = element.parentElement.getBoundingClientRect();
const childRect = element.getBoundingClientRect();
const fullyContained =
parentRect.left <= childRect.left &&
parentRect.right >= childRect.right &&
parentRect.top <= childRect.top &&
parentRect.bottom >= childRect.bottom;
const significantOverlap =
(childRect.width * childRect.height) /
(parentRect.width * parentRect.height) > 0.1;
if (fullyContained && significantOverlap) {
const nextParent = element.parentElement;
if (nextParent.tagName.toLowerCase() !== 'body' &&
nextParent.tagName.toLowerCase() !== 'html') {
element = nextParent;
} else {
break;
}
} else {
break;
}
}
}
const ownerDocument = element.ownerDocument; const ownerDocument = element.ownerDocument;
const frameElement = ownerDocument?.defaultView?.frameElement; const frameElement = ownerDocument?.defaultView?.frameElement;
@@ -1027,40 +993,6 @@ export const getRect = async (page: Page, coordinates: Coordinates, listSelector
} }
} }
if (element.tagName !== 'TABLE') {
while (element.parentElement) {
if (element.tagName.toLowerCase() === 'body' ||
element.tagName.toLowerCase() === 'html') {
break;
}
const parentRect = element.parentElement.getBoundingClientRect();
const childRect = element.getBoundingClientRect();
const fullyContained =
parentRect.left <= childRect.left &&
parentRect.right >= childRect.right &&
parentRect.top <= childRect.top &&
parentRect.bottom >= childRect.bottom;
const significantOverlap =
(childRect.width * childRect.height) /
(parentRect.width * parentRect.height) > 0.1;
if (fullyContained && significantOverlap) {
const nextParent = element.parentElement;
if (nextParent.tagName.toLowerCase() !== 'body' &&
nextParent.tagName.toLowerCase() !== 'html') {
element = nextParent;
} else {
break;
}
} else {
break;
}
}
}
const rectangle = element?.getBoundingClientRect(); const rectangle = element?.getBoundingClientRect();
if (rectangle) { if (rectangle) {
const createRectObject = (rect: DOMRect) => ({ const createRectObject = (rect: DOMRect) => ({
@@ -2377,40 +2309,6 @@ export const getNonUniqueSelectors = async (page: Page, coordinates: Coordinates
} }
} }
if (element.tagName !== 'TABLE') {
while (element.parentElement) {
if (element.tagName.toLowerCase() === 'body' ||
element.tagName.toLowerCase() === 'html') {
break;
}
const parentRect = element.parentElement.getBoundingClientRect();
const childRect = element.getBoundingClientRect();
const fullyContained =
parentRect.left <= childRect.left &&
parentRect.right >= childRect.right &&
parentRect.top <= childRect.top &&
parentRect.bottom >= childRect.bottom;
const significantOverlap =
(childRect.width * childRect.height) /
(parentRect.width * parentRect.height) > 0.1;
if (fullyContained && significantOverlap) {
const nextParent = element.parentElement;
if (nextParent.tagName.toLowerCase() !== 'body' &&
nextParent.tagName.toLowerCase() !== 'html') {
element = nextParent;
} else {
break;
}
} else {
break;
}
}
}
const generalSelector = getSelectorPath(element); const generalSelector = getSelectorPath(element);
return { generalSelector }; return { generalSelector };
}, coordinates); }, coordinates);