@crawlee/utils - npm Package Compare versions

internals/cheerio.js

		@@ -40,13 +40,4 @@ "use strict";
		return '';
		// TODO: Add support for "html" being a Cheerio element, otherwise the only way
		// to use it is e.g. htmlToText($('p').html())) which is inefficient
		// Also, it seems this doesn't work well in CheerioScraper, e.g. htmlToText($)
		// produces really text with a lot of HTML elements in it. Let's just deprecate this sort of usage,
		// and make the parameter "htmlOrCheerioElement"
		const $ = typeof htmlOrCheerioElement === 'function' ? htmlOrCheerioElement : cheerio_1.default.load(htmlOrCheerioElement, { decodeEntities: true });
		let text = '';
		// TODO: the type for elems is very annoying to work with.
		// The correct type is Node[] from cheerio but it needs a lot more casting in each branch, or alternatively,
		// use the is* methods from domhandler (isText, isTag, isComment, etc.)
		// @ts-expect-error
		const process = (elems) => {
		@@ -65,3 +56,3 @@ const len = elems ? elems.length : 0;
		if (compr.startsWith(' ') && /(^\|\s)$/.test(text))
		compr = compr.substr(1);
		compr = compr.substring(1);
		text += compr;
		@@ -68,0 +59,0 @@ }

internals/memory-info.js

		@@ -44,13 +44,8 @@ "use strict";
		mainProcessBytes = process.memoryUsage().rss;
		// TODO this is quite ugly, let's introduce some local variables
		// https://stackoverflow.com/a/55914335/129415
		childProcessesBytes = +(0, node_child_process_1.execSync)('cat /proc/meminfo')
		.toString()
		.split(/[\n: ]/)
		.filter((val) => val.trim())[19]
		* 1000 // meminfo reports in kb, not bytes
		// the total used memory is reported by meminfo
		// subtract memory used by the main node process
		// in order to infer memory used by any child processes
		- mainProcessBytes;
		const memInfo = (0, node_child_process_1.execSync)('cat /proc/meminfo').toString();
		const values = memInfo.split(/[\n: ]/).filter((val) => val.trim());
		// /proc/meminfo reports in kb, not bytes, the total used memory is reported by meminfo
		// subtract memory used by the main node process in order to infer memory used by any child processes
		childProcessesBytes = +values[19] * 1000 - mainProcessBytes;
		}
		@@ -57,0 +52,0 @@ else {

package.json

		{
		"name": "@crawlee/utils",
		"version": "3.0.0-beta.60",
		"version": "3.0.0-beta.61",
		"description": "A set of shared utilities that can be used by crawlers",
		@@ -52,3 +52,3 @@ "engines": {
		"@apify/ps-tree": "^1.1.4",
		"@crawlee/types": "^3.0.0-beta.60",
		"@crawlee/types": "^3.0.0-beta.61",
		"cheerio": "1.0.0-rc.11",
		@@ -55,0 +55,0 @@ "content-type": "^1.0.4",

internals/cheerio.d.ts.map