Skip to content

Commit

Permalink
Make summarization work for text/plain content
Browse files Browse the repository at this point in the history
document.body is available and text can be extracted from it.

Fixes #521
  • Loading branch information
mrichtarsky authored and josStorer committed Oct 30, 2023
1 parent 9085396 commit f26c566
Showing 1 changed file with 15 additions and 4 deletions.
19 changes: 15 additions & 4 deletions src/utils/get-core-content-text.mjs
Original file line number Diff line number Diff line change
Expand Up @@ -15,6 +15,9 @@ const adapters = {
}

function findLargestElement(e) {
if (!e) {
return null
}
let maxArea = 0
let largestElement = null
const limitedArea = 0.8 * getArea(e)
Expand Down Expand Up @@ -50,14 +53,22 @@ export function getCoreContentText() {
console.log(largestElement)
console.log(secondLargestElement)

if (!largestElement) return
function getTextFrom(e) {
return e.innerText || e.textContent
}

let ret
if (secondLargestElement && getArea(secondLargestElement) > 0.5 * getArea(largestElement)) {
ret = secondLargestElement.innerText || secondLargestElement.textContent
if (!largestElement) {
ret = getTextFrom(document.body)
console.log('use document.body')
} else if (
secondLargestElement &&
getArea(secondLargestElement) > 0.5 * getArea(largestElement)
) {
ret = getTextFrom(secondLargestElement)
console.log('use second')
} else {
ret = largestElement.innerText || largestElement.textContent
ret = getTextFrom(largestElement)
console.log('use first')
}
return ret.trim().replaceAll(' ', '').replaceAll('\n\n', '').replaceAll(',,', '')
Expand Down

0 comments on commit f26c566

Please sign in to comment.