CSC/ECE 517 Fall 2015 E1576 Refactoring submitted content (hyperlinks and files)

Introduction to Expertiza

Expertiza<ref>https://github.com/expertiza/expertiza</ref> is a project developed using Ruby on Rails<ref>http://guides.rubyonrails.org/getting_started.html</ref>. It provides features like peer review, team assignments and submission of projects. This can be achieved by submitting code base, URL of hosted code on remote server and Wiki submissions. It is an open source application and the code can be cloned from GitHub. This application provides an efficient way to manage assignments, grades and reviews. This makes the process easier and faster when the class strength is large.

Problem Statement

Files involved

• assignment_participant.rb
• assignment_team.rb and
• views related to submitted hyperlinks and files (submitted_content/_hyperlink.html.erb, submitted_content/_submitted_files.html.erb)

What it does

It handles the display, submission, deletion of hyperlinks and files by the teams.

What is wrong with it

The submitted hyperlinks and files are stored in participants table, which means they are associated with individuals instead of team. This is not a good design because in Expertiza, students are always grouped in teams, even when there is only one person in each team (individual assignments). Currently, when a team member wants to see all the submitted hyperlinks of his team, Expertiza needs to load all the submitted hyperlinks from each participant in this team and merge them all before display. Similar overhead happens when a team member wants to delete a submitted hyperlink. Also, the directory_num field needs to be moved from the participants table to the teams table. It signifies the sub-directory which stores all the files submitted by this team. Intuitively it is clear that each team should ideally have only one common copy of directory_num.

What needs to be done

• Create a new column hyperlinks in the teams table.
• Write db migration, move all the participants’ submitted hyperlinks from participants table into teams table. Duplicate hyperlinks need to be removed.
• Write db migration, remove the hyperlinks field from participants table.
• Rewrite the hyperlink-related methods in both assignment_participant.rb and assignment_team.rb to make sure the new code works with the new db design.
• Make sure when doing the peer-review, the reviewer can see the submitted content from the reviewees.
• Make sure when instructor see the peer-review records, they can see the submitted content from the reviewees.
• Write db migration, move the directory_num field (also the content, of course) to the teams table.
• Move the set_student_directory_num method from participants_controller to teams_controller, then refactor this method into smaller methods.
  • This method no longer needs to test all the participants.
  • Check if any of them have directory_num. If so, instead, it should just check the directory_num from teams table.
• Make change to the submitted-file-related code to make it work with the new design.
• Write test cases to test student-submitted hyperlinks and files.

Scope

This project can be divided into four major work items:

1. Moving the hyperlinks field from the participants table to the teams table. After this change, relevant changes are needed in the code to support this database change.
2. Moving the directory_num field from the participants table to the teams table. After this change, relevant changes are needed in the code to support this database change.
3. Moving the set_student_directory_num method from the participants_controller to the teams_controller. It just needs to check directory_num from the team table.
4. Write test cases to verify student-submitted hyperlinks and files.

Implementation

Expertiza files that will be modified for each work item along with the description:

Work Item #	Files to be modified	Description
1	assignment_participant.rb assignment_team.rb db:participants table db:teams table	- Create a new column hyperlinks in teams table - Write db migration, move all the participants’ submitted hyperlinks from participants table into teams table. All duplicated hyperlinks have to be removed. - Write db migration, remove the hyperlinks field from participants table. - hyperlink-related methods in both assignment_participant.rb and assignment_team.rb have to be rewritten to make sure the new code works with the new db design.
2	participants_controller.rb teams_controller.rb db:participants table db:teams table	- Write db migration, move the directory_num field (also the content) to teams table. - Move the set_student_directory_num from participant_controller to team_controller, then refactor this method into smaller methods. - This method no longer needs to test all the participants and see if any of them have directory_num, instead, it just check the directory_num from teams table. - Make change to the submitted-file-related code to make sure it works with the new design.
3	New files to be added	- Write test cases to test student-submitted hyperlinks and files.

Design

Consider a simple use case where we need to retrieve the submitted hyperlinks and files corresponding to team_id, say, 23841. Since there is no existing correlation between teams and submitted content, we will first need to look up the participants in the team using the teams_users table. And then, using those user_id values one at a time, we need to look up the participants table.

Finally, we need to merge the results before displaying it.

def hyperlinks
  links = Array.new
  self.participants.each { |team_member|
    links.concat(team_member.hyperlinks_array) if team_member.hyperlinks_array}
  links
end

So, it becomes a tedious 3-step process. It is evident from this example that the current design is inefficient.

Instead, if the hyperlinks data were available in teams table, we could directly access all the submitted content for a team in one shot, making it much more streamlined.

As a part of this change, we’ll have to move the following routines in assignment_participants.rb, which deal hyperlink handling to assignment_team.rb.

  # Appends the hyperlink to a list that is stored in YAML format in the DB
  # @exception  If is hyperlink was already there
  #             If it is an invalid URL
  def submit_hyperlink(hyperlink)
    hyperlink.strip!
    raise "The hyperlink cannot be empty" if hyperlink.empty?
    url = URI.parse(hyperlink)
    # If not a valid URL, it will throw an exception
    Net::HTTP.start(url.host, url.port)
    hyperlinks = self.hyperlinks_array
    hyperlinks << hyperlink
    self.submitted_hyperlinks = YAML::dump(hyperlinks)
    self.save
  end

  # Note: This method is not used yet. It is here in the case it will be
    needed.
  # @exception  If the index does not exist in the array
  def remove_hyperlink(hyperlink_to_delete)
    hyperlinks = self.hyperlinks_array
    hyperlinks.delete(hyperlink_to_delete)
    self.submitted_hyperlinks = YAML::dump(hyperlinks)
    self.save
  end

  def hyperlinks
    team.try(:hyperlinks) || []
  end
 
  def hyperlinks_array
    self.submitted_hyperlinks.blank? ? [] :YAML::load(self.submitted_hyperlinks)
  end

Similar refactoring work needs to be done for the directory_num field as well. Finally, the refactoring needs to be thoroughly tested by writing test cases for student-submitted hyperlinks and files by making sure that:

• the submitted hyperlinks and files are correctly displayed on the submitted content page.
• the reviewer can see the submitted content from the reviewees.
• the instructor can see the submitted content from the reviewees.

Use Cases

Use Case 1: Submit hyperlink for an assignment by student.

Description: A student who is an assignment participant should be able to submit hyperlink on the “Your work” section of an assignment.

Preconditions: Student should be a participant for the assignment he is trying to submit the hyperlink.

Postconditions: The submitted hyperlink should be stored in the teams table of the database.

Success Scenario:

1. The submitted hyperlink is visible every time the student views the “Your work” section for the Assignment.
2. All other team members should be able to view the submitted hyperlink.
3. A reviewer should be able to see the submitted hyperlink.
4. An instructor should be able to see the submitted hyperlink.

Use Case 2: Submit file for an assignment by student.

Description: A student who is an assignment participant should be able to submit a file on the “Your work” section of an assignment.

Preconditions: Student should be a participant for the assignment he is trying to submit the file.

Postconditions: The directory_num field which stores the directory number of the directory storing all the submitted files for a team should be stored in the teams table of the database.

Success Scenario:

1. The submitted file is visible every time the student views the “Your work” section for the Assignment.
2. All other team members should be able to view the submitted file.
3. A reviewer should be able to see the submitted file.
4. An instructor should be able to see the submitted file.

Use Case 3: Delete hyperlink for an assignment by student.

Description: A student who is an assignment participant should be able to delete submitted hyperlink on the “Your work” section of an assignment.

Preconditions: Student should be a participant for the assignment he is trying to delete the hyperlink.

Postconditions: The deleted hyperlink should be removed from the teams table in DB.

Success Scenario:

1. The deleted hyperlink should not be visible the next time the student clicks on “Your work” section.
2. All other team members should not see the deleted hyperlink.
3. A reviewer should not be able to see the deleted hyperlink.
4. An instructor should not be able to see the deleted hyperlink.

Accomplishments

The project was divided into 4 categories:

1. Database migrations
2. Refactor submitted_hyperlink related methods
3. Refactor directory_num related methods
4. Testing

We followed the steps listed below in the process of refactoring the submitted content controller:

• Writing database migrations to move the submitted_hyperlinks and directory_num attributes from the Participants table to the Teams table.
• Rewriting submitted_hyperlinks related methods in the Participants model to read from and write to the Teams table.
• Modifying the corresponding views to read hyperlinks_array instead of hyperlinks. hyperlinks_array stores the the submissions of all the team members of a team.
• Removing directory_num related methods from the participant model.
• Rewriting directory_num related methods in the team model.
• Rewriting the corresponding views to read directory_num from the teams table instead of the participants table.
• For testing, we first created fixtures for all required tables.
• We tested our changes by writing functional tests to test the SubmittedContentController.
• Our test suite included the following tests:
  • Testing submission of a hyperlink by one student - Check the count of submmited hyperlinks, verify that the hyperlink is stored properly in the database and displayed correctly from the UI.
  • Testing submission of two hyperlinks by two students in a team - Check the count of submmited hyperlinks, verify that the hyperlink is stored properly in the database and displayed correctly from the UI.
  • Testing submission of the same hyperlink by two students in a team - Check that the duplicate hyperlinks does not get stored.
  • Testing deletion of a hyperlink - Check the count of submmited hyperlinks and the updated list is displayed correctly from the UI.
  • Testing submission of one file by one student - Check the submission of a file and check the updated count of submitted files.
  • Testing deletion of one file by one student - Check the deletion of a submitted file and check the updated count of submitted files.
  • Testing submission of multiple files by multiple students - Check the submission of multiple files and check the updated count of submitted files.
  • Testing deletion of multiple files by multiple students - Check the deletion of multiple submitted files and check the updated count of submitted files.

We can conclusively say that, after refactoring the submitted content controller, the query time for retrieving submissions will improve significantly as they are stored together on a team-basis rather than for each participant.

Video demo of our work can be found at https://www.youtube.com/watch?v=VHVWF_UKgOc&feature=youtu.be
GitHub repository containing the refactoring changes can be found at https://github.com/nitinnagaraja/expertiza/
Link to the pull request to Expertiza is https://github.com/expertiza/expertiza/pull/629

Future Work

A proposal for further improvement would be to possibly create a new submitted_content table in the database. This table could include attributes like user_id, team_id, can_submit, hyperlink, directory_num, uploaded_at, updated_at. This approach would resolve the existing hyperlinks issue and more importantly, aid in recording a history of the submitted hyperlinks. There could be a “Show previous submissions” view for a team. So, if ever there’s a scenario where a student or an instructor needs to revert or go back and check a previous submission, he/she can easily access it from this table.

References

<references/>