Implement HUJI course import feature and some other frontend changes

[nmdanny]

Import feature

• Closes Parse student's past courses from HUJI website #44

This feature allows a user to run a script within the "מידע אישי" site to scrape all his registered courses, and pass them
to the frontend. The script is executed by running a bookmark as explained on the import instructions page.

The script does not communicate with the back-end, rather, it communicates with the frontend page(locally, no network communication) by the use of cross-origin messaging API. This is why users are require to open HUJI's site via a pop-up.

This is good from several standpoints

• Security-wise, the scrape script doesn't need authentication (which would've required embedding/passing auth credentials to the script, which is running in a potentially malicious origin)
• UX: the results of the scrape are presented to the user in a table, allowing him to be aware of which courses cannot
  be added(missing from our database or summer/annual courses), and manually choose a semester for courses whose
  semester wasn't identified.
• Separation of concerns/less code duplication: The front-end already has components for dealing with authentication
  and communication with the DB ($http and $auth), no need to embed them in the script as well

Other front-end changes

• Use of a store object for holding courses, which should fix some difficulties regarding accessing the courses information from analytics view #96

• The course store also saves to local storage whenever a change is made. In the future we should also save
  to DB

• Some fixes and refactorings of ProgressBox (I'm guessing that feature is still WIP but it
  caused exceptions due to missing type, which would occur be if the student has no track assigned)

Future improvement ideas

• Currently, the front-end makes a separate GET for each course(via the /course/?course_id=?? endpoint)
  for each scraped course to verify that said course exists in the DB(and to obtain its semester if it wasn't inferred
  from the personal information site)

  We should implement an endpoint that does all this at once (gets a list of course IDs and returns a list of courses),

• Use proper, styled HTML elements instead of window.alert within the scrape script

  Done

• Perhaps find a way to scrape the student's track? The grades page only seems to include his faculty, maybe this info
  appears somewhere else within "מידע אישי"

• Implement Scrape course grade histogram from HUJI personal user information site #77, scrape grade statistics too

• Paginate the parsed courses table which can get pretty long, and other UI improvements

• Currently, when validating the existence of the parsed courses, I compare them to the Shnaton entries
  of 2021, because we didn't seem to parse other years yet. This can potentially cause mismatches(e.g, courses
  whose semester changed)

[ScDor]

Very nice work, please see the attached comments

[ScDor]

Are there any changes required here due to the new data year mechanism, or can we merge?

[nmdanny]

Are there any changes required here due to the new data year mechanism, or can we merge?

Made a tiny fix, should be good to go