This ticket addressed #10

What was done

Changing the test runner from jest to vitest for better compatibility with ESM
Generate the grammar in ts instead of js to allow the test to run parse stage separately
Remove the Classify stage tests (since the scraper's url fetch is unreliable)
Add tests for tokenize stage in the form of snapshots
- The tests use locally stored html since the fetch step is unreliable (as mentioned above with the Classify stage test)
Add tests for parse stage
- 10 can be boiled down to the "AND course followed by an OR course" test case
Fix #10 by adding a check whether the reqs is an array or not
- This is mainly because nearley will always wrap the parsed results inside an array when using a post-processor

Future works

Add more tests for XOMs and any other cases that weren't covered here
Review the grammar and the post-processor to make sure the fix is robust enough/there is a better way to do it through the grammar
Address the failing test for unbounded range courses (similar nested array problem)