Embedded Files
Develop a Python program to detect when an open reservation is available, and immediately book the reservation for the user
Develop a Python program to detect when an open reservation is available, and immediately book the reservation for the user
Goals and Requirements
Goals and Requirements
Be able to ski and snowboard in the 2020-2021 season
Use for any date at a variety of partner resorts
Automatically detect when a reservation for a given date opens due to a cancellation
Automatically book the reservation and notify that it is booked
Background
Background
As mountains were beginning to open in late 2020, most COVID-19 restrictions were still in place and reduced the capacity for many of the ski resorts. Poorly planned reservation systems made it extremely difficult to get any reservations at local resorts or partner resorts across the country on any day.
Reserving any date essentially required constantly refreshing the resort page, clicking through menus to enter information and check the specific date, and if one became available then continue through more menus to reserve.
This consistently took anywhere from 20 minutes to an hour or so to successfully find an opening. Once found, sometimes the reservation steps took long enough for the spot to no longer be available and the reservation would fail.
Design Process
Design Process
Main HTML requests to be emulated
Check availability of desired dates at desired resort
Submit a reservation request
Check if reservation is confirmed in user account
Each user has specific info to be included in each request
User ID number (found with Inspect after login)
phone number
email
To be included in each HTML requests
User-specific information
Main URL for each of the three requests
Date range
SessionID to extracted from the cookies
"Resort Code" for the specific partner resort
Desired date
__requestverificationtoken
Simple method of handling the authentication process for proof-of-concept prototype
Copied from a specific HMTL request in DevTools after logging in
Front End User Process
Front End User Process
Initial one-time setup
Enter phone number
Enter email
Enter user ID number
Use terminal to get ChromeKey for importing and deciphering cookies
Each session
Enter desired resort name
Select from the resorts with Resort Codes defined in the script
Enter "Dates_To_Book" as a list of strings in "mm-dd" format
Login to account in Chrome browser
Copy "__requestverificationtoken" from the specific header
Run Script and wait for confirmation in the terminal
Back End
Back End
Setup
Import, decipher, and parse cookies from current session given the main URL
Insert "__requestverificationtoken" from user input
Extract SessionID to include in encoded URL
Reformat each user-entered date in the list to the full length
Main Loop
For each date in the list of desired dates
Send availability check request
Interpret json.response
Error handling
If the day is unavailable, it will be included in 'NoInventoryDays'
If date is available
Submit reservation HTML request
Submit Confirmation HTML request
If confirmation request fails, retry loop
Remove date from the desired list
Loop every few seconds until the desired date list is empty
Error Handling
Check json response for specific text involving errors that commonly occurred
The reservation system was occasionally unavailable
Ignore and continue loop, usually re-opens very shortly
Javascript error or "page down" error
Indicates that the current session expired, requiring an update to the __requestverificationtoken and/or Session_ID
Summary
Summary
Overall the program was very successful. I was able to confirm reservations on any day with minimal effort, and was able to book at partner resorts throughout the country. Occasionally it would still take 20-30 minutes to confirm a spot on busy weekends, but no manual work was required past the initial setup or the occasional restart if the session expired. The intent was never to distribute any further than those around me that I would take trips with, so I did not develop a GUI or a more user-friendly method since the proof-of-concept was more than enough for the end goal.
Page updated
Google Sites
Report abuse