You can not select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
139 lines
3.6 KiB
139 lines
3.6 KiB
const puppeteer = require('puppeteer'); |
|
|
|
const { |
|
has_past_events, |
|
has_upcoming_events, |
|
get_upcoming_events_from_page, |
|
get_past_events_from_page, |
|
map_event, |
|
} = require('./logic'); |
|
const { graphql_endpoint } = require('./constants'); |
|
|
|
const open_browser = async ({ headless }) => { |
|
const browser = await puppeteer.launch({ |
|
headless, |
|
args: ['--disable-dev-shm-usage'], |
|
}); |
|
return browser; |
|
}; |
|
|
|
const register_page_scraper = (endpoint, page, past_events = false) => { |
|
let responses = []; |
|
return new Promise((resolve) => { |
|
page.on('response', async (response) => { |
|
if (endpoint === response.request().url()) { |
|
let json = {}; |
|
try { |
|
json = await response.json(); |
|
} catch (error) { |
|
return responses; |
|
} |
|
|
|
const getters = { |
|
upcoming: get_upcoming_events_from_page, |
|
past: get_past_events_from_page, |
|
}; |
|
|
|
const events = getters[past_events ? 'past' : 'upcoming'](json); |
|
if (events !== null) { |
|
responses = [events, ...responses]; |
|
if (!events.page_info.has_next_page) { |
|
resolve(responses); |
|
} |
|
} |
|
} |
|
}); |
|
}); |
|
}; |
|
|
|
const get_body_inner_text = async (page) => |
|
await page.evaluate('document.querySelector("body").innerText;'); |
|
|
|
const get_page_events = async (opt) => { |
|
const browser = await open_browser(opt); |
|
const facebook_page = await browser.newPage(); |
|
|
|
let past_events = []; |
|
let upcoming_events = []; |
|
|
|
let scraping_past_events = false; |
|
let scraping_upcoming_events = false; |
|
|
|
if (opt.get_past_events) { |
|
scraping_past_events = true; |
|
past_events = register_page_scraper(graphql_endpoint, facebook_page, true) |
|
.then((past_events) => { |
|
scraping_past_events = false; |
|
return past_events; |
|
}) |
|
.catch((err) => { |
|
console.error(err); |
|
scraping_past_events = false; |
|
return []; |
|
}); |
|
} else { |
|
past_events = Promise.resolve([]); |
|
} |
|
if (opt.get_upcoming_events) { |
|
scraping_upcoming_events = true; |
|
upcoming_events = register_page_scraper(graphql_endpoint, facebook_page) |
|
.then((upcoming_events) => { |
|
scraping_upcoming_events = false; |
|
return upcoming_events; |
|
}) |
|
.catch((err) => { |
|
console.error(err); |
|
scraping_upcoming_events = false; |
|
return []; |
|
}); |
|
} else { |
|
upcoming_events = Promise.resolve([]); |
|
} |
|
|
|
await facebook_page.goto(opt.page_id); |
|
await facebook_page.waitFor(2000); |
|
const accept_buttons = await facebook_page.$x( |
|
"//button[contains(text(), 'Accept All')]", |
|
); |
|
if (accept_buttons.length > 0) { |
|
accept_buttons[0].click(); |
|
} |
|
|
|
const body_text = (await get_body_inner_text(facebook_page)).toLowerCase(); |
|
const past_resolved = opt.get_past_events && !has_past_events(body_text); |
|
const upcoming_resolved = |
|
opt.get_upcoming_events && !has_upcoming_events(body_text); |
|
|
|
if (past_resolved) { |
|
past_events = Promise.resolve([]); |
|
scraping_past_events = false; |
|
} |
|
|
|
if (upcoming_resolved) { |
|
upcoming_events = Promise.resolve([]); |
|
scraping_upcoming_events = false; |
|
} |
|
|
|
while (scraping_past_events || scraping_upcoming_events) { |
|
await facebook_page.waitFor(1000); |
|
await facebook_page.evaluate(() => window.scrollBy(0, window.innerHeight)); |
|
if (past_resolved && upcoming_resolved) { |
|
break; |
|
} |
|
} |
|
|
|
upcoming_events = await upcoming_events; |
|
past_events = await past_events; |
|
|
|
const responses = [...upcoming_events, ...past_events]; |
|
const nodes = responses.reduce( |
|
(res, current) => [...res, ...current.edges], |
|
[], |
|
); |
|
|
|
return nodes.map(map_event); |
|
}; |
|
|
|
module.exports = { |
|
get_page_events, |
|
};
|
|
|