do not fix the cli the same links and image links
This commit is contained in:
+17
-6
@@ -567,13 +567,21 @@ function contentDispatch(funcName, args) {
|
|||||||
return document.querySelector(selector) !== null;
|
return document.querySelector(selector) !== null;
|
||||||
}
|
}
|
||||||
function extractLinks() {
|
function extractLinks() {
|
||||||
return Array.from(document.querySelectorAll("a[href]")).map(a => ({
|
const seen = new Set();
|
||||||
|
return Array.from(document.querySelectorAll("a[href]")).reduce((links, a) => {
|
||||||
|
const href = a.href;
|
||||||
|
if (!href || seen.has(href)) return links;
|
||||||
|
seen.add(href);
|
||||||
|
links.push({
|
||||||
text: a.textContent.trim().slice(0, 100),
|
text: a.textContent.trim().slice(0, 100),
|
||||||
href: a.href,
|
href,
|
||||||
}));
|
});
|
||||||
|
return links;
|
||||||
|
}, []);
|
||||||
}
|
}
|
||||||
function extractImages() {
|
function extractImages() {
|
||||||
return Array.from(document.querySelectorAll("img")).map(img => {
|
const seen = new Set();
|
||||||
|
return Array.from(document.querySelectorAll("img")).reduce((images, img) => {
|
||||||
const src =
|
const src =
|
||||||
img.src ||
|
img.src ||
|
||||||
img.getAttribute("data-src") ||
|
img.getAttribute("data-src") ||
|
||||||
@@ -581,10 +589,13 @@ function contentDispatch(funcName, args) {
|
|||||||
img.getAttribute("data-original") ||
|
img.getAttribute("data-original") ||
|
||||||
(img.srcset ? img.srcset.split(",")[0].trim().split(" ")[0] : "") ||
|
(img.srcset ? img.srcset.split(",")[0].trim().split(" ")[0] : "") ||
|
||||||
"";
|
"";
|
||||||
|
if (!src || seen.has(src)) return images;
|
||||||
|
seen.add(src);
|
||||||
const FAKE_ALT = new Set(["true", "false", "null", "undefined", "image", "img"]);
|
const FAKE_ALT = new Set(["true", "false", "null", "undefined", "image", "img"]);
|
||||||
const alt = img.alt && !FAKE_ALT.has(img.alt.trim().toLowerCase()) ? img.alt.trim() : "";
|
const alt = img.alt && !FAKE_ALT.has(img.alt.trim().toLowerCase()) ? img.alt.trim() : "";
|
||||||
return { alt, src };
|
images.push({ alt, src });
|
||||||
}).filter(img => img.src !== "");
|
return images;
|
||||||
|
}, []);
|
||||||
}
|
}
|
||||||
function extractText() {
|
function extractText() {
|
||||||
return document.body.innerText;
|
return document.body.innerText;
|
||||||
|
|||||||
Reference in New Issue
Block a user