Fix parsing fail error for Parse phase

[rael346]

closes #14 #15

This pull request mainly fixes most of the common errors in the scraper's Parse phase (parsing fail, unexpected end of token, ambiguous grammar) by doing the following

• Refactor the snapshot tests to output into individual files to be more readable (This is mainly to avoid large changes that affect multiple majors like the missing sections error near the end of the semester)
• Tokenize Section Info only if it is after a header/subheader/subsubheader or it is the first element in the section. This is because the Section Info phrases ("Complete one of the following" for example) can also be tokenized as a XOM
• Add a new Token type SUBSUBHEADER (basically a SUBHEADER but with indentation). This header sometimes doesn't have any meaning (see CS and English's 19th century section for an example), but other times it does. This is mainly here so that we have more options to parse later on (in this PR the Parse Phase just ignore the subsubheader token and the comment).
• Tokenize Phase now aggregates adjacent course lists table. This is because some majors decided to separate two tables in a section (see this major for an example), which the Tokenize Phase would have understood as two different sections.
• Rewrite the grammar for the parse phase to be more succinct and contains less recursion (since it is easier to debug and less chance of having ambiguous grammar error)
• Refactor the Parse Phase code to write snapshot tests

With this PR, most of the errors in the Parse Phase are gone. If you run the scraper for 2023, there are only 13 Parse Phase errors, and all of them are specific to one major.

For future works

• Address comments that are not tokenized as XOM #16 is the main focus now since a lot of majors has errors relating to this
• Unknown catalog type - some majors just don't contain the word "major", "minor" or "concentration" in their name
• Potentially removing the Section Info type - This type is basically just XOM but at the beginning of a section, so we can move the tokenize logic to the grammar instead

[AlpacaFur]

Looks awesome, thank you so much for working on this!