It worked!
I did not get glasto tickets. I ran the bot for Thursday coach tickets. The site kept crashing and pages timed out. Still,I eventually got through. I entered our registrations but I couldn't pick bus tickets due to an error on the page, this seemed to happen to a lot of people on twitter. I fixed the errors and made a few good improvements over the weekend. I ran it on two laptops then on Sunday. I got through on both. It froze up after submitting payment. This seemed to be a common problem on twitter once again. I waited til the 10 mins on ticket lock was over and submitted on the other page. It was too late however. Sold out.
This app launches chrome via puppeteer. It opens a number of browsers set by the user. It will then iterate through each browser and load the set URL in a tab. It will only begin loading the page on the next browser tab when a certain amount of time has passed so it does not surpass the set rate limit (60 a minute on glastonbury site).
After each page has loads it calculates a similarity rating by comparing the text on the loaded page to the text in resources/live.txt . It is using the inner text of all elements within the body of the returned page. The browser will then automatically switch to the tab with the highest similarity rating. This tab will not be reloaded until another tab beats its similarity rating.
You can pause by pressing the enter key on the command line. It should automatically stop when the reg page loads.
--site : URL
--rate-limit : rate limit per minute
--max-tabs : the number of tabs to use. The more the better. A tab will reload after the iteration of loading tabs has looped back around to it. So more tabs means more time for a page to load.
--test : Will use resources/test.txt for comparison. For use with test site.
npm i
Example run command:
node main.js --site="https://glastonbury.seetickets.com" --rate-limit=55 --max-tabs=15
Test site from https://github.com/thomasms/testsites
cd test_site
npm start
Add test flag in run command:
node main.js --site="localhost:3000" --rate-limit=55 --max-tabs=15 --test
When it loads the 20th page it pretends it is in. Runs on localhost:3000
Scalability. How could this scale? It's currently close to maxing out the potential for one device. Not sure how to do this. I think being able to deploy to a few instances, they would need to use a VPN to avoid the IP being identified as being in the cloud providers IP range. Would need some way to get the most similar tab back to the user so details could be entered. Feeding back results and allowing use to select a tab, then tunnel the port back to the user app. Complex. Pass the session and vpn used back? How could this be done simpler? Is it ethical to scale?