Problem: Colleges have a lot of information about majors online, but it is very difficult to centralize that data without doing a lot of research. For Brown students specifically, we want to make a platform that allows students to input past classes they've taken, concentration interests, and workload preferences to help them better navigate concentration requirements and possible degrees.
- We will also take into account if a student wants a lighter workload, or is taking leave/study abroad for a particular semester and weight that semester with courses that require less time through scraping data from Critical Review.
- Personal preferences will weight classes by class structure (exams, projects, or essays) as well as the department a class is in. If a student prefers a particular type of class or if they want to take classes outside of their concentration.
- Teacher and class ratings from Critical Review, and merging a visual GUI with information from CAB including course name, course description, and professor.
- Returns a set of possible N-Semester concentration plans, based on past courses taken and a set of requirements for a list of possible concentration, including WRIT requirements (& 30 credits).
- Scraping information from Critical Review and CAB to get data for class weights and display on the user interface
- Not generic to other universities, only applicable for Brown Students (If we wanted to include RISD classes, we would also need to get that information as well)
- Logic on the algorithm: figuring out constraints on their preferences, requirements, workload, etc. Furthermore, there are a lot of constraints because of semesters a class is offered, time a course is offered, etc.
- Edge cases: e.g. 0.5'ers, independent concentrators, GISPs, capstone projects, personal projects, research, recruiting scheduling, TA schedules, Sports.
Team Members: Members: Melissa Cui, Natalie Rshaidat, Ifechi Ilozor, Nicholas Keirstead
None
Database Design:
We used the proxy pattern to abstract away the interactions between the sql database and any class that wants data from it. It also allowed me to use the Guava cache to store all the courses in the courses table when the cache is first initialized. This not only optimized times needed to get data from sql database, but it also allowed for us to populate each course's data when we get the courses associated with a concentration and avoid having the result set opened to query the concentration table and the courses table.
-
DatabaseInterface: Contains all the methods that the Database class and DatabaseCache share.
-
Database Class: The class that connects to the sql database, error checks concentration tables since we manually inputted their values, and convert what is returned from sql queries into Nodes for the pathway class to use.
-
DatabaseCache Class: The class that has the cache of all courses and calls on the Database class to query the sql database.
-PathwayProgram class: A class that contains all the information needed from the GUI and executes making pathways for the main class. It also serves as integrater class between the pathway class, main class, and database classes.
-GUI Design: In our GUI we have a home page that allows the user to select their university and once a university is selected it takes the user to a login page that allows the user to login, sign up, or login as a guest. After they have completed some sort of logging in then they are taken to a generate pathway page which asks the user for concentration information in order to build 3 pathways for them. They are then shown a low, median, and high pathway and can select on any of them to look at more stats and modify the courses listed.
Node class: This class represents courses in the database.
Semester class: This class represents a semester object with references to the courses in this semester and semester related stats.
Pathway class: This class builds the actuals pathway for the concentration.
Time optimizations were made by using the guava cache to load in all of the courses upon initialization of the databaseCache class. This optimization made queries runtime cut in half. Since there were only 1900 courses we felt it was not that immense of a burden on space compared to the runtime burden of each sql query. It takes 5 seconds on average to load in all of the courses which is done while the user is inputing their user preferences.
-
To build:
Use "mvn package" or "mvn site" (for more project details). Make sure to do the steps in the "How to run our tests" section before building, so that JUnit tests don't fail when building. -
To run Pathway via terminal:
./run open google chrome or safari and enter http://localhost:4567/login
- To run unit tests: "mvn test", or using one of the build commands (mvn package)
Since we do not have a REPL and only a GUI, I decided to make a Pathway Program class that handles all of the user input parsing and outputs data to gui. By having this class I was able to thuroughly test all outputs that would be displayed back to the user, which is why PathwayProgram has 100% coverage in jacoco. The testing suite for pathway program tests cornell extensibility, pathway outputs, pathway stats, pathway unique courses and not returning null pointers back to the main class. In addition, I also had to test all of database data and wrote 8 additional public methods to tests that the concentration tables are lining up, the data types are of the correct format, and to check that data was clean. Given that the unclean sql database has caused a lot of bugs and errors in classes that use that database's data, I had to use these error checking methods in my testing suite which is why I left those methods public.
-coursesDB.db: Brown Courses Database
-courses table: contains all courses from Fall 2020 and Spring 2020
-course_id: Deptartment code +space+ course number (CSCI 0150) from CAB
-course_name: Name of the course (Intro to Object Oriented Programming) from CAB
-prereqs: courseID's of the prereqs for this course where a ',' represents AND and '=' represents OR from CAB
-semester_offered: 0 is spring, 1 is fall, and 2 is both semesters from CAB
-professor: the course's professor from CAB
-courseRating: the most recent course rating from critical review (0-5)
-avg_hrs: the most recent record of average hours listed on critical review (hrs)
-max_hrs: the most recent record of maximum hours listed on critical review (hrs)
-CR_link: the critical review link for the course
-class_size: the most recent record of the class size from critical review
-department: the 3-4 letter code of the department the course is in (CSCI 0150-> CSCI)
-number: the course number listed after the department code in the course id (CSCI 0150-> 0150)
-ss_link: the self service link for the course
-concentrations table:
-concentration_id: the table name associated with the concentration name
-concentration_name: the name of the concentration for displaying on the GUI
-Each concentration has one table for its rules and one for the course options that satisfy credits for the concentration
-Naming of concentration courses tables: concentration name in all lowercase with no spaces with ba or bs in the end to specify if the courses listed are for the bachelor of sciences or just bachelors degree.
-Naming of concentration rules tables: same table name as teh concentrations course list but with '_rules' on the end
-cognitiveneuroscienceba table: sample concentration courses table
-course_id: the course id used in the courses table
-next: course id of the next course in a sequence (CS15 would have CS16 as its next)
-category: the category that the course is credit for
-cognitiveneuroscienceba_rules table: sample concentration rules table
-category: 0-20+ we organized concentration rules into categories where one has to take a specified number of courses from one category to satisfy a certain number of credits for the concentration
-num_credits: the number of course one must take to satisfy the categories requierements
-cornellcoursesDB.db: Cornell Courses Database
-courses table: contains all courses from Fall 2020 and Spring 2020
-course_id: Deptartment code +space+ course number (CSCI 0150) from ss_link
-course_name: Name of the course (Intro to Object Oriented Programming) from cornell course ss_link
-prereqs: courseID's of the prereqs for this course where a ',' represents AND and '=' represents OR from ss_link
-semester_offered: 0 is spring, 1 is fall, and 2 is both semesters from ss_link
-professor: the course's professor from ss_link
-courseRating: the most recent course rating from cure reviews (0-5)
-avg_hrs: the most recent record of average hours listed on cure reviews (hrs)
-max_hrs: the most recent record of maximum hours listed on cure reviews (hrs)
-CR_link: the https://www.cureviews.org link equivalent to brown's critical review
-class_size: the most recent record of the class size from CR_link
-department: the 3-4 letter code of the department the course is in (CSCI 0150-> CSCI)
-number: the course number listed after the department code in the course id (CSCI 0150-> 0150)
-ss_link: the https://classes.cornell.edu link for each course
-concentrations table:
-concentration_id: the table name associated with the concentration name
-concentration_name: the name of the concentration for displaying on the GUI
-economicsba table: sample concentration courses table
-course_id: the course id used in the courses table
-next: course id of the next course in a sequence (CS15 would have CS16 as its next)
-category: the category that the course is credit for
-economicsba_rules table: sample concentration rules table
-category: 0-20+ we organized concentration rules into categories where one has to take a specified number of courses from one category to satisfy a certain number of credits for the concentration
-num_credits: the number of course one must take to satisfy the categories requierements
-coursesDBColErr.db:has wrong column names and number of columns for a multitude of tables for testing
-coursesDBempty.db: has an empty courses table for testing
-coursesDBInvalid.db: has wrong type fields in multiple tables for testing
None.
Natalie:
-Database and comprehensive unit tests
-Integrating CAB scraping to main sql database
-Manually double checking, fixing inaccurate prereqs, and entering missing prereqs
for all courses in MATH, APMA, BIOL, CSCI.
-Figured shibboleth authentication for all scraping
-Cleaning CAB parsing data (removing nulls, extra commas, numbers), debugging cab
scraping, and figuring out CAB / Self Service scraping
-Routing and various handling of user input
-Navigation bar functionality
-Extensive debugging of all parts (db, algo & frontend)
-Setting up login, login as guest, and sign up routing correctly
-Setting up university extensibility
-Setting up PathwayProgram and calculating stats for pathways
-All of PathwayProgram JUnit tests (unique courses, cornell, brown) and optimizing their runtime
-Making unique courses hover info and making the the 3 pathway containers clickable in 3 pathway page
-Making GUI flexible to changes in course size
Mel:
-Critical Review scraping
-Manually double checking, fixing inaccurate prereqs, and entering missing prereqs for
all courses in ECON
-Concentrations table in the database
-Connecting Spark to our Java code
-Setting up the entire front end framework
-Routing, alerting and comprehensive handling of user input
-Styling and adding functionality to front end (login, faq, main form, 3-pathway display)
-Transitioned styling to semantic ui to improve user experience
-Solving Flash of Unstyled Content (when the HTML loads before the CSS)
-Setting up cornell db and integrating cornell into GUI
-Setting up university extensibility
Ifechi:
-Developing and optimizing the fundamental makePathway algorithm
-Layering complexity onto the algorithm: weighting courses based on different Bayesian stastical
measures, fast weighted shuffle using binary search, and tracking lag for pacing of pathways
-Pathway package
-JUnit testing for makePathway algorithm (by-hand and with the database, covering all scenarios)
-Manually double checking, fixing inaccurate prereqs, and entering missing prereqs
for all courses in CHEM, PHYS, HIAA.
-Solving deep copy of the course set to make multiple unique and valid pathways
-mypath.ftl & mypath.css
-Implementing and styling customization functionality on the front end (add, move, swap)
-Implementing and styling hover to show course details + linking to Self-Service
-Adjusting style according to pathway data (grouping by year and fall/spring)
-Styling semester boxes, arrows, and pathway statistics (everything responsive)
-Styling all navigation bars
Nick:
-CAB/self service scraping
-Manually fixing prereqs for CHEM
-Unit testing CAB database outputs
-Username/password fields on sign up page, guest button in the login page
-Unique courses for 3-pathway display and displaying pathway statistics
-Sorting pathways by workload to display correctly
Mentor TA: Julianne Rudner julianne_rudner@brown.edu