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

feat: pass object instead of str for key-val pair

Browse Source
This commit is contained in:
karishmas6
2024-08-03 20:31:55 +05:30
parent 25eb58945e
commit cf06bbe703
1 changed files with 57 additions and 47 deletions
Download Patch File Download Diff File Expand all files Collapse all files

30
maxun-core/src/browserSide/scraper.js
View File

@@ -126,13 +126,18 @@ function scrapableHeuristics(maxCountPerPage = 50, minArea = 20000, scrolls = 3,
return out;
}
// wrap inside an IIFE to avoid polluting the global scope: https://github.com/microsoft/playwright/issues/31864
(function (window) {
/**
* Returns a "scrape" result from the current page.
* @returns {Array<Object>} *Curated* array of scraped information (with sparse rows removed)
*/
// Wrap the entire function in an IIFE (Immediately Invoked Function Expression)
// and attach it to the window object
(function(window) {
/**
* Returns a "scrape" result from the current page.
* @returns {Array<Object>} *Curated* array of scraped information (with sparse rows removed)
*/
window.scrape = function (selector = null) {
window.scrape = function(selector = null) {
/**
* **crudeRecords** contains uncurated rundowns of "scrapable" elements
* @type {Array<Object>}
@@ -180,10 +185,10 @@ function scrapableHeuristics(maxCountPerPage = 50, minArea = 20000, scrolls = 3,
/**
* Given an object with named lists of elements,
* groups the elements by their distance in the DOM tree.
* @param {Object.<string, object[]>} lists The named lists of HTML elements.
* @param {Object.<string, {selector: string, tag: string}>} lists The named lists of HTML elements.
* @returns {Array.<Object.<string, string>>}
*/
window.scrapeSchema = function (lists) {
window.scrapeSchema = function (lists) {
function omap(object, f, kf = (x) => x) {
return Object.fromEntries(
Object.entries(object)
@@ -199,8 +204,8 @@ function scrapableHeuristics(maxCountPerPage = 50, minArea = 20000, scrolls = 3,
}
function getSeedKey(listObj) {
const maxLength = Math.max(...Object.values(omap(listObj, (x) => x.length)));
return Object.keys(ofilter(listObj, (_, v) => v.length === maxLength))[0];
const maxLength = Math.max(...Object.values(omap(listObj, (x) => document.querySelectorAll(x.selector).length)));
return Object.keys(ofilter(listObj, (_, v) => document.querySelectorAll(v.selector).length === maxLength))[0];
}
function getMBEs(elements) {
@@ -219,12 +224,17 @@ function scrapableHeuristics(maxCountPerPage = 50, minArea = 20000, scrolls = 3,
}
const seedName = getSeedKey(lists);
const MBEs = getMBEs(lists[seedName]);
const seedElements = Array.from(document.querySelectorAll(lists[seedName].selector));
const MBEs = getMBEs(seedElements);
return MBEs.map((mbe) => omap(
lists,
(listOfElements) => listOfElements.find((elem) => mbe.contains(elem))?.innerText,
({ selector }) => {
const elem = Array.from(document.querySelectorAll(selector)).find((elem) => mbe.contains(elem));
return elem ? elem.innerText : undefined;
},
(key) => lists[key].selector // Use the selector as the key in the output
));
}
}
})(window);