Skip to content

Latest commit

 

History

History
204 lines (140 loc) · 5.89 KB

README.md

File metadata and controls

204 lines (140 loc) · 5.89 KB
Raw

NU Course

Functions to get NU course information

This project is meant to get course data for ninest/husker.

Folders

  • constants/
    • Contains constants, like TERMS, stating which terms should be fetched
  • banner/
    • Contains all Banner-specific code for fetching terms, subjects, and courses
    • Only contains functions that call the API and return data as is (getAllSubjects, searchSections, getCourseCoreqs)
  • fetcher/
    • Contains functions to call Banner functions and write responses to JSON files
    • Use transformer functions to map API responses to required data types
    • Think of constants in constants/ as the "parameters" for the fetchers
  • transformers/
    • Contains functions to map Banner API responses to required types
  • read/
    • Contains utility functions to read data already fetched (data that's been fetched and is saved in one of the JSON files)
  • mappings/
    • Functions to read existing data and save it in another form
    • For example, terms to subject/courses: courses are fetched, and contain sections which contain terms. This makes it difficult to find out which courses are offered in which term. So, there is a mapping of terms to courses offered in that term.
  • util/
    • Dump of (complex) utility functions used in one or more locations
  • api/
  • .data/
    • subjects.json
    • courses/
      • Contains one JSON file per course
    • mappings/
    • all-courses.json: all courses in a single JSON file

Build Setup

First, install Deno. Clone or fork the repository, then do the following.

1. Constants

Then open constants/terms.ts, and enter the terms you want to fetch (TERMS).

For example, if you want to fetch courses from only Fall 2022, the file should look like this:

export const TERMS = [
  {
    code: "202310",
  },
];

By default, the output folder is the .data folder.

The description key in each term is optional, so [{ code: "202310" }] should also work.

You can find a list of available terms through NUBanner's API: nubanner.neu.edu/StudentRegistrationSsb/ssb/classSearch/getTerms?searchTerm=&offset=1&max=20.

Note that the output folder (.data/nu in this example) should contain a folder courses.

2. Subjects

To fetch all subjects, run

deno task fetch:subjects

This should create a file called subjects.json in the output folder path with a list of subjects containing a code and description:

[
  {
    "code": "ACCT",
    "description": "Accounting"
  },
  {
    "code": "AFAM",
    "description": "African American Studies"
  },
  ...
]

3. Courses

To fetch all courses under each subject, run

deno task fetch:courses

This will take a long time, so follow the progress through the logs.

This will create a file for each subject in the courses folder in the output path. For example, there will be courses/CS.json containing a list of courses.

Note that in fetching the courses, the data of the courses already fetched will be merged in with the "new" data from Banner's API. If the descriptions or requisites have already been fetched and deno task fetch:courses is run, deno task fetch:descriptions does not have to be run again.

4. Description

To fetch all descriptions for courses, run

deno task fetch:descriptions

This will check if a description exists for each course and fetch a description if not present.

In fetcher/constants.ts, populate the following with courses that may have different descriptions across sections. For example:

export const COURSES_WITH_ALT_SECTIONS = [
  { code: "CS", number: "2500" },
  { code: "CS", number: "2510" },
];

In this case, the descriptions of the first three sections are fetched, and the most common on is set as the main description. In the example for Fundies Accelerated, the accelerated section's description is not very useful, so the only the descriptions of the regular sections are saved.

See #1 for more information.

5. Requisites

a. Co-requisites

Co-requisites are lists of the following:

type Requisite = Pick<Course, "subject" | "number">;
// A list of coreqs is of type Requisite[]

Run

deno task fetch:coreqs

If a course already has the coreq key and a list (which can be empty), it will skip that course as it means that the co-requisites for that course have already been fetched.

b. Pre-requisites

Pre-requisites are lists of

export type Prerequisite = "Or" | "And" | "(" | ")" | Requisite | string;

See #7 for details on this format.

Run

deno task fetch:prereqs

Similar to co-requisites, lists of pre-requisites already fetched won't be re-fetched.

Mappings

Mappings are meant to make the data easier to use in applications. After all, these are just JSON files, not a relational database where we can nicely get the data we want.

Combine courses

First, merge all courses into a single JSON file by running

deno task combine-courses

Terms to subject/courses (#24)

Map each term code to courses offered in that term. Each term has its own JSON file (202340.json for Summer I 2023) with a key for each subject code. The value is a list of courses (containing course number and a list string[] of CRNs)

export interface TermSubjectCourseMapping {
  [subjectCode: string]: { number: string; crns: string[] }[];
}

Run

deno task mapping:term-courses

Format

Finally, run

deno task fmt:data

Support

If you have any suggestions, please open an issue or pull request with your changes.

License

MIT