A sublime text 3 plugin by Aditya Ramesh
Code Faster is a Sublime Text 3 Plugin developed with the intent of fetching problem statements directly into the sublime text editor, without needing to open a browser.
The problem statements are fetched from the Codeforces programming platform (www.codeforces.com) due to their reputation in the field of algorithmic contests. In addition, their API also lent itself well for the purpose of this plug-in.
Assuming you have Sublime Text 3 set up, head over to C:/Users/{YOUR_USERNAME}/Appdata/Roaming/Sublime Text 3/Packages/
and paste the folder 'Code-Faster-A' in this directory.
You'll also need the python modules urllib, json and BeautifulSoup to get this up and running.
Now, open Sublime Text 3 and along the menu bar, to the right of the Help option, you should notice a new option Codeforces for Adi, and if you do, rejoice!
Codeforces has a standard template of contest information -> Contest Number + Letter
This pair uniquely defines every question on the platform.
The codeforces API was used to find the current contest number, by requesting for the information of the problems available on the platform and then the ongoing or scheduled contest's number is determined.
The letter (which is usually 'A', 'B', 'C', 'D' or 'E'} is manually entered by the user due to the input restrictions on sublime text plug-ins.
An alternative is to manually hard code 5 separate plug-ins, each that retrieves questions corresponding to different letters of the same contest number. UPDATE: This has been done now, and the "Codeforces for Adi" button now spawns a dropbox wherein the user can select the difficulty level they want a problem statement to be at.
The advantage of using codeforces as the target platform is that the standard problem URL is - codeforces.com/contest_number/letter
The BeautifulSoup module in python, allows for web scraping. Consequently, after acquiring the correct link by leveraging the codeforces API, BeautifulSoup was used to scrape the problem statement off the webpage.
-
The parsing of the problem statements of the URL was tedious, and while an alternative would be to remove all the text between every
<and>pair, a quick fix was implemented due to lack of time. -
In addition, it's probably advisable to cache the requests being made as the JSON obtained in response is pretty large in size, and in runtime requires ~ 3-4 seconds to work.
-
A creative fix to minimize fetching time would be, to replace calling all the current information of the problems on the platform, and use their API to call just the most recent submissions, as during a live contest, majority of the audience on the platform will actively be submitting solutions of the same contest.
Fix the HTML tag parsing issue.I have gotten a lead! This will solve the issue of LaTeX tags and even images, yes, images in Sublime Text is possible. I am currently working on this, and will update the repo once I've made considerable progress.Handle questions with embedded images - either automate the opening of the images alone in a browser, or directly open the same question in a browser due to lack of image support in- well, text editors.UPDATE: This has now been done -- upon scraping the problem statement, if an image exists, Python's Selenium module is leveraged to retrieve the image URL and is opened in a new Firefox window.- Allow users to submit their solutions to the corresponding links (from within Sublime Text 3)
- Incorporate Machine Learning Algorithms to parse through user's submission history, and perform analysis of his submitted solutions, and pull recommended questions for the user to learn faster.