You cannot select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

212 lines
5.9 KiB
JavaScript

"use strict";
const cheerio = require("cheerio");
const url = require("url");
const assert = require("assert");
const syncpipe = require("syncpipe");
const htmlEntities = require("html-entities");
const createDatasheet = require("../../shared/create-datasheet");
const assureResponse = require("../../shared/assure-response");
const getUntaggedText = require("../../shared/get-untagged-text");
// ST Microelectronics
function extractID(string) {
// Quick-and-dirty category ID parsing from category pages
let match = /"prmisID":"([^"]+)"/.exec(string);
if (match != null) {
return match[1];
} else {
throw new Error(`ST: prmis ID expected but not found`);
}
}
module.exports = function ({ session }) {
return {
seed: [{
id: "st:home",
tags: [ "st:home" ],
data: {}
}],
tags: {
"st:home": [ "st:findCategories" ],
"st:category": [ "st:scrapeCategory" ],
"st:product": [ "st:scrapeProduct", "st:normalizeProduct" ],
},
tasks: {
"st:findCategories": {
ttl: "15d",
run: async function ({ createItem }) {
let response = await session.get("https://www.st.com/content/st_com/en.html");
assureResponse(response);
let $ = cheerio.load(response.body);
let links = $("#Top_Menu_Products :is(a.st-nav__blockmenu-title, a.st-nav__blockmenu-link)")
.toArray()
.map((element) => $(element).attr("href"))
.map((relativeURL) => url.resolve("https://www.st.com/", relativeURL));
for (let link of links) {
createItem({
id: `st:category:${link}`,
tags: [ "st:category" ],
data: { url: link }
});
}
}
},
"st:scrapeCategory": {
ttl: "1d",
taskInterval: "60s",
version: "2",
run: async function({ data, createItem }) {
let response = await session.get(data.url);
assureResponse(response);
let prmisID = extractID(response.body.toString());
let listingResponse = await session.get(`https://www.st.com/en/documentation/scraper.cxst-ps-grid.html/${encodeURIComponent(prmisID)}.json`, { noDecode: true });
assureResponse(response);
let listingBuffer = listingResponse.body;
if (listingBuffer.length > 0) {
// This is a category that has a product explorer
let listing = JSON.parse(listingBuffer.toString());
let cellNames = listing.columns.map((column) => {
let cellName = (column.identifier != null)
? `${column.identifier}_${column.qualifier_identifier}`
: `nonstandard:${column.name}:${column.qualifier}`
createItem({
id: `st:column:${cellName}`,
tags: [ "st:column" ],
data: column
});
return cellName;
});
for (let row of listing.rows) {
assert(row.productId != null);
let cellData = syncpipe(row.cells, [
(_) => _.map((cell, i) => [ cellNames[i], cell.value ]),
(_) => Object.fromEntries(_)
]);
createItem({
id: `st:product:${row.productId}`,
tags: [ "st:product" ],
data: {
... row,
cells: undefined,
cellData: cellData
}
});
}
} else {
console.warn("Warning: empty response, category does not have product explorer");
}
}
},
"st:scrapeProduct": {
ttl: "15d",
taskInterval: "5s",
run: async function({ data, createItem, updateData, expireDependents }) {
if (data.productFolderUrl == null) {
throw new Error(`No known product page URL`);
}
let productPageURL = url.resolve("https://www.st.com/", data.productFolderUrl);
let response = await session.get(productPageURL);
assureResponse(response);
let $ = cheerio.load(response.body);
let datasheetLink = $("a[data-js='datasheetLink']").attr("href");
let datasheetURL = (datasheetLink != null)
? url.resolve(productPageURL, datasheetLink)
: null;
let resources = $(".st-table--resources")
.find("h3").toArray()
.map((heading) => {
let $heading = $(heading);
let sectionID = $heading.attr("id");
let sectionTitle = $heading.text().trim();
let $table = $heading.next("table");
let items = $table
.find("tbody tr").toArray()
.map((row) => {
let $row = $(row);
let $mainView = $row.find(".visible-on-desktop-only");
let $link = $mainView.find("a.st-link");
return {
url: url.resolve(productPageURL, $link.attr("href")),
documentID: $link.find("span.st-font--bold").text().trim(),
description: $link.find("span:not(.st-font--bold)").text().trim(),
version: getUntaggedText($link),
date: $row.find(".visible-on-desktop-only[data-latest-update]").text().trim()
};
});
return {
sectionID: sectionID,
sectionTitle: sectionTitle,
items: items
};
});
updateData((data) => {
return {
... data,
datasheetLink: datasheetURL,
resources: resources
};
});
expireDependents();
for (let section of resources) {
for (let resource of section.items) {
createItem({
id: `st:resource:${resource.url}`,
tags: (resource.url === datasheetURL)
? [ "st:resource", "st:datasheet" ]
: [ "st_resource" ],
data: { url: resource.url }
});
}
}
}
},
"st:normalizeProduct": {
dependsOn: [ "st:scrapeProduct" ],
version: "8",
parallelTasks: 50,
run: async function (api) {
let { data } = api;
createDatasheet(api, {
priority: 0.8,
source: "st",
manufacturer: "STMicroelectronics",
productID: data.productId,
name: data.cellData["XJE010_VT-007"],
description: htmlEntities.decode(data.cellData["XJE014_VT-007"]),
url: data.datasheetLink
});
}
},
}
};
};