Skip to content

An XBlock for posing questions to students

License

AGPL-3.0, Unknown licenses found

Licenses found

AGPL-3.0
LICENSE
Unknown
LICENSE.MIT
Notifications You must be signed in to change notification settings

mudassir-hafeez/problem-builder

 
 

Repository files navigation

Problem Builder and Step Builder

Circle CI

This repository provides two XBlocks: Problem Builder and Step Builder.

Both blocks allow to create questions of various types. They can be used to simulate the workflow of real-life mentoring, within an edX course.

Supported features include:

  • Free-form answers (textarea) which can be shared accross different XBlock instances (for example, to allow a student to review and edit an answer they gave before).
  • Self-assessment MCQs (multiple choice questions), to display predetermined feedback to a student based on his choices in the self-assessment. Supports rating scales and arbitrary answers.
  • MRQs (Multiple Response Questions), a type of multiple choice question that allows the student to select more than one choice.
  • Answer recaps that display a read-only summary of a user's answer to a free-form question asked earlier in the course.
  • Progression tracking, to require that the student has completed a particular step before allowing them to complete the next step. Provides a link to the next step to the student.
  • Tables, which allow to present answers from the student to free-form answers in a concise way. Supports custom headers.
  • Dashboards, for displaying a summary of the student's answers to multiple choice questions. Details

The following screenshot shows an example of a Problem Builder block containing a free-form question, two MCQs and one MRQ:

Problem Builder Example

Installation

You can install Problem Builder from PyPI using this command:

pip install xblock-problem-builder

For full details, see "Open edX Installation", below.

Usage

See Usage Instructions

Workbench Installation and Settings

Prerequisites

  • Python 3.5+
  • Compiler/build tool chain
  • Python headers
  • MySQL development libraries and headers
  • Virtualenv

On Ubuntu 16.04, these can be installed by running

sudo apt-get install build-essential libpython3-dev libmysqlclient-dev virtualenv

Developer Installation

For developers, you can install this XBlock into an XBlock SDK workbench's virtualenv.

First, create a Python3 virtualenv:

~/xblock_development $ virtualenv -p python3 venv
~/xblock_development $ . venv/bin/activate

Now run the following commands from the problem builder repo root to install the problem builder dependencies:

(venv) ~/xblock_development/problem-builder $ pip install -r requirements.txt
(venv) ~/xblock_development/problem-builder $ pip install -r requirements-dev.txt

Switch to the created XBlock SDK repository, install its dependencies, and create its migrations:

(venv) ~/xblock_development/problem-builder $ cd ../venv/src/xblock-sdk
(venv) ~/xblock_development/venv/src/xblock-sdk $ make requirements

Create the following configuration file in workbench/settings_pb.py:

from settings import *

INSTALLED_APPS += ('problem_builder',)

Testing problem-builder in the workbench requires MySQL instead of the standard SQLite configuration it uses. You can quickly spin up an instance of MySQL with Docker using the following command:

# -d for detach mode
docker run --rm -it -p 3307:3306 -e MYSQL_ROOT_PASSWORD=rootpw -e MYSQL_DATABASE=db -d mysql:5.6

By default, the xblock-sdk uses the SQLite database but MySQL can be used by specifying an environment variable WORKBENCH_DATABASES in the following format.

export WORKBENCH_DATABASES='{"default": {"ENGINE": "django.db.backends.mysql", "NAME": "db", "USER": "root", "PASSWORD": "rootpw", "HOST": "127.0.0.1", "PORT": "3306"}}'

Ensure that the database name and credentials match the ones configured in the docker container.

Run this from the XBlock repository root:

./manage.py migrate --settings=workbench.settings_pb

Running the workbench

./manage.py runserver 8000 --settings=workbench.settings_pb

Access it at http://localhost:8000/.

Running tests

The integration tests require a recent Firefox and geckodriver (CI uses Firefox 70 and geckodriver 0.26). These can be installed locally for testing if required. For example on Linux:

mkdir external
cd external
wget https://archive.mozilla.org/pub/firefox/releases/70.0.1/linux-x86_64/en-US/firefox-70.0.1.tar.bz2
tar jxf firefox-70.0.1.tar.bz2
wget https://github.com/mozilla/geckodriver/releases/download/v0.26.0/geckodriver-v0.26.0-linux64.tar.gz
tar xzf geckodriver-v0.26.0-linux64.tar.gz
export PATH="$(pwd):$(pwd)/firefox/:$PATH"

# now can run integration tests using this firefox version
cd ..
make test.integration

From the problem-builder repository root, run the tests with the following command:

make test

See also the following for more scoped tests:

make quality
make test.unit
make test.integration

Debugging CI Failures

Sometimes it can be hard to figure out why some tests fail in the CI. When Circle CI browser based tests fail for unknown reasons, it can be helpful to run them with VNC enabled so that you can observe the browser (or even interact with it) while the tests are running.

To enable VNC on Circle CI, first re-run the failing test with SSH enabled: in the Circle CI UI, click the "Rerun" dropdown and select "Rerun Job with SSH". The job will be re-run with SSH enabled. You can find the IP/port combination that lets you log into the VM with your github SSH key under the "Enable SSH" step in the pipeline UI.

SSH into the VM, forwarding the VNC port:

ssh -p <port> <ip-address> -L 5900:localhost:5900

Install the required packages:

sudo apt-get install -yq xvfb x11vnc fluxbox

Start up xvfb and the VNC server:

rm -f /tmp/.X$(echo ${DISPLAY:-:0} | cut -b2-)-lock
Xvfb ${DISPLAY:-:0} -ac -listen tcp -screen 0 1440x900x24 &
/usr/bin/fluxbox -display ${DISPLAY:-:0} -screen 0 &
x11vnc -display ${DISPLAY:-:0} -forever -noxdamage -rfbport 5900 -quiet -passwd pass &

You should now be able to connect to the server via VNC. On macOS, you can use the built-in VNC viewer that you can launch by opening Finder and choosing the "Go -> Connect to Server.." from the menu. Type in localhost:5900 and enter pass when asked for the password.

You are all set up to run integration tests with screen sharing enabled. For some reason Firefox does not want to start in foreground mode when run as non-root, so you'll have to run the tests as root.

unset MOZ_HEADLESS
cd /home/circleci/project
source venv/bin/activate
make test

Working with Translations

For information about working with translations, see the Internationalization Support section of the Open edX XBlock Tutorial.

Prepare your virtualenv and ensure that the Transifex authentication file (~/.transifexrc) is properly set up.

Push new strings to Transifex:

$ make extract_translations
$ make push_translations

To get the latest translations from Transifex:

$ make pull_translations
$ make compile_translations

For testing purposes it's faster to avoid Transifex and work on dummy Esperanto translations:

$ make build_dummy_translations

The Transifex configuration is stored in .tx. For more information read transifex's documentation

If you want to add a new language:

  1. Add language to problem_builder/translations/config.yaml
  2. Make sure all tagged strings have been extracted and push to Transifex as described above.
  3. Go to Transifex and translate both of the Problem Builder and the Problem Builder JS resources.
  4. When you're done with the translations pull from Transifex as described above.

Adding custom scenarios to the workbench

Within the xblock-problem-builder repository, create the templates/xml and add XML scenarios to it - all files with the *.xml extension will be automatically loaded by the workbench:

$ mkdir templates/xml
$ cat > templates/xml/my_pb_scenario.xml

Restart the workbench to take the new scenarios into account.

Upgrading from Version 1

To upgrade a course from xblock-mentoring ("v1") to xblock-problem-builder ("v2"), run the following command on a system with edx-platform, xblock-mentoring, and xblock-problem-builder installed:

$ SERVICE_VARIANT=cms DJANGO_SETTINGS_MODULE="cms.envs.devstack" python -m problem_builder.v1.upgrade "Org/Course/Run"

Where "Org/Course/Run" is replaced with the ID of the course to upgrade.

Open edX Installation

Problem Builder releases are tagged with a version number, e.g. v2.6.0, v2.6.5. We recommend installing the most recently tagged version, with the exception of the following compatibility issues:

  • edx-platform version open-release/ironwood.2 and earlier must use versions < 4.0.0. See PR 262 for details.
  • edx-platform version open-release/eucalyptus.2 and earlier must use ≤v2.6.0. See PR 128 for details.
  • edx-platform version named-release/dogwood.3 and earlier must use v2.0.0.

The edx-platform master branch will generally always be compatible with the most recent Problem Builder tag. See the EDXAPP_PRIVATE_REQUIREMENTS setting for the version currently installed on edx.org.

To install new versions of Problem Builder (v3.1.3+), use pip install xblock-problem-builder or specify a version using e.g. pip install xblock-problem-builder==3.1.3. To do this on Open edX could look like:

$ sudo -Hu edxapp bash
edxapp $ cd && . edxapp_env  && . ./venvs/edxapp/bin/activate && cd edx-platform/
edxapp $ pip install xblock-problem-builder
edxapp $ ./manage.py lms migrate --settings=aws  # or openstack, as appropriate

Then, restart the edxapp services:

$ sudo /edx/bin/supervisorctl restart edxapp:
$ sudo /edx/bin/supervisorctl restart edxapp_workers:

To install old verions of Problem Builder (< v3.1.3) on an Open edX installation, choose the tag you wish to install, follow the above instructions but instead of the pip install xblock-problem-builder command, use:

TAG='v2.6.5' pip install "git+https://github.com/open-craft/problem-builder.git@$TAG#egg=xblock-problem-builder==$TAG"

Note that Problem Builder requires xblock-utils. If you are installing it into a virtualenv used by edx-platform, xblock-utils should already be installed. But if you are installing it into another virtualenv, you may need to first install xblock-utils manually (recent versions of it are not available on PyPI so will not be automatically installed).

See Usage Instructions for how to enable in Studio.

Publishing to PyPI

Whenever we tag a new version, e.g. v3.1.3 and push it to GitHub, CircleCI will build it and deploy it to PyPI automatically. For details on how this works, see this pull request.

License

This XBlock is available under the GNU Affero General Public License (AGPLv3).

About

An XBlock for posing questions to students

Resources

License

AGPL-3.0, Unknown licenses found

Licenses found

AGPL-3.0
LICENSE
Unknown
LICENSE.MIT

Stars

Watchers

Forks

Packages

No packages published

Languages

  • Python 72.9%
  • JavaScript 19.5%
  • HTML 4.5%
  • CSS 2.8%
  • Makefile 0.3%