Skip to content

Latest commit

 

History

History
226 lines (146 loc) · 10.5 KB

README.md

File metadata and controls

226 lines (146 loc) · 10.5 KB
Raw

Learning Management System

by Ari Gilmore, Thomas Clark, Meet Patel, and Zachary Selk

Modeled from https://developer.ibm.com/tutorials/cfc-starter-kit-speech-to-text-app-example/

Prerequisites

Instructions

1. Clone the repository

git clone https://github.com/zacharyselk/Interns-be-Hacking.git
cd Interns-be-Hacking

2. Setup environment variables

Create a .env file in the root project directory containing the following environment variables. Note - these will be replaced by your IBM Cloud service credentials in the next step.

touch .env
STT_API_KEY=<api key for speech to text service>
STT_URL=<URL for speech to text service>
TRANSLATE_API_KEY=<api key for translator service>
TRANSLATE_URL=<URL for translator service>
COS_API_KEY=<cloud object storage api key>
COS_IAM_ROLE_CRN=<cloud object storage IAM role crn. e.g. crn:v1:bluemix:public:iam::::serviceRole:Writer>
COS_ENDPOINT=<cloud object storage endpoint. e.g. s3.eu-gb.cloud-object-storage.appdomain.cloud>
COS_BUCKET_NAME=<cloud object storage bucket name>

3. Create IBM Cloud services and obtain service credentials

Register/Login to IBM Cloud and create the following services:

  • IBM Watson Speech to Text

    • Copy the apikey and url values within the service credentials to the STT_API_KEY and STT_URL environment variables in the .env file you created in step 2.

  • IBM Watson Language Translator

    • Copy the apikey and url values within the service credentials to the TRANSLATE_API_KEY and TRANSLATE_URL environment variables in the .env file you created in step 2.

  • IBM Cloud Object Storage.

    • Create a standard bucket with a given name. Copy this name to the COS_BUCKET_NAME environment variable in the .env file you created in step 2. For the purposes of this starter kit, the bucket you create in Cloud Object Storage requires public access.
    • Copy the apikey and iam_role_crn values within the service credentials to the COS_API_KEY and COS_IAM_ROLE_CRN environment variables in the .env file you created in step 2.
    • Navigate to the endpoints URL within the service credentials (e.g. https://control.cloud-object-storage.cloud.ibm.com/v2/endpoints) and choose a public service-endpoint that is close to your location. Copy your chosen endpoint to the COS_ENDPOINT environment variable in the .env file you created in step 2.

4. Install dependencies and run the applications

Docker

docker build --tag video-transcriber:1.0 .
docker run --publish 3000:3000 --detach --name demo video-transcriber:1.0

Server

  1. This tutorial uses pipenv. If you are using another python distribution or package manager, you will need to install the dependencies located in the Pipfile. Alternatively, using pipenv, from the root project directory, create a pipenv virtual environment.

    pipenv --python <path to python executable>

    Note. If python 3.6 is installed in the default location, this can be specified as:

    pipenv --python 3.6

  2. Activate the pipenv shell:

    pipenv shell

  3. Install the project dependencies:

    pipenv install

  4. To run your application locally, use:

    python manage.py start

    The manage.py utility offers a variety of different run commands to match your situation:

    • start: Starts a server in a production setting using gunicorn.
    • run: Starts a native Flask development server. This includes backend reloading upon file saves and the Werkzeug stack-trace debugger for diagnosing runtime failures in-browser.
    • livereload: Starts a development server using the livereload package. This includes backend reloading as well as dynamic frontend browser reloading. The Werkzeug stack-trace debugger will be disabled, so this is only recommended when working on frontend development.
    • debug: Starts a native Flask development server, but with the native reloader/tracer disabled. This leaves the debug port exposed to be attached to an IDE (such as PyCharm's Attach to Local Process).

    There are also a few utility commands:

    • build: Compiles .py files within the project directory into .pyc files.
    • test: Runs all unit tests inside of the project's test directory.

    The server is running at: http://localhost:3000/ in your browser.

Frontend UI Development

  1. If you have not done so already, install Node.js and Yarn.

  2. In a new terminal, change to the frontend directory from the project root and install the dependencies:

    cd frontend
yarn install

  3. Launch the frontend application:
    Compiles and hot-reloads for development

    yarn serve

    Compiles and minifies for production

    yarn build

    Lints and fixes files

    yarn lint

    The frontend UI is now running at http://localhost:8080/ in your browser.

5. Language Translator Extension

This tutorial shows you how to create a Watson Language Translator service and write the necessary server side code to translate video transcriptions. The front-end UI implementation is left as an extension for you to implement yourself. Hint - inspecting the upload_video function in server/routes/index.py, you can see that the server side expects a source and a target language as part of the POST request form data to /upload_video. Supported language models are provided at https://localhost:3000/language_models once your server is running.

6. Deploy the app

The following instructions apply to deploying the Python Flask server. To deploy the frontend UI, follow the Node.js build and deploy tutorial.

Deploying to IBM Cloud

You can deploy this application to IBM Cloud or build it locally by cloning the repo first. Once your app is live, you can access the /health endpoint to build out your cloud native application.

Use the button below to deploy this same application to IBM Cloud. This option creates a deployment pipeline, complete with a hosted GitLab project and DevOps toolchain. You will have the option of deploying to either Cloud Foundry or a Kubernetes cluster. IBM Cloud DevOps services provide toolchains as a set of tool integrations that support development, deployment, and operations tasks inside IBM Cloud.

Deploy to IBM Cloud

Building locally

To get started building this application locally, you can either run the application natively or use the IBM Cloud Developer Tools for containerization and easy deployment to IBM Cloud.

Native application development

Native application development was covered in step 4 above when you installed and ran the app. Your server is running at: http://localhost:3000/ in your browser.

There are two different options for debugging a Flask project:

  1. Run python manage.py runserver to start a native Flask development server. This comes with the Werkzeug stack-trace debugger, which will present runtime failure stack-traces in-browser with the ability to inspect objects at any point in the trace. For more information, see Werkzeug documentation.

  2. Run python manage.py debug to run a Flask development server with debug exposed, but the native debugger/reloader turned off. This grants access for an IDE to attach itself to the process (that is, in PyCharm, use Run -> Attach to Local Process).

You can also verify the state of your locally running application using the Selenium UI test script included in the scripts directory.

Note for Windows users: gunicorn is not supported on Windows. You can start the server with python manage.py run on your local machine or build and start the Dockerfile.

IBM Cloud Developer Tools

Install IBM Cloud Developer Tools on your machine by running the following command:

curl -sL https://ibm.biz/idt-installer | bash

Create an application on IBM Cloud by running:

ibmcloud dev create

This will create and download a starter application with the necessary files needed for local development and deployment.

Your application will be compiled with Docker containers. To compile and run your app, run:

ibmcloud dev build
ibmcloud dev run

This will launch your application locally. When you are ready to deploy to IBM Cloud on Cloud Foundry or Kubernetes, run one of the following commands:

ibmcloud dev deploy -t buildpack // to Cloud Foundry
ibmcloud dev deploy -t container // to K8s cluster

You can build and debug your app locally with:

ibmcloud dev build --debug
ibmcloud dev debug

Summary

This tutorial has shown you how to build and deploy an app that uses Watson Speech to Text to transcribe video files. This tutorial also covered deploying a Watson Language Translator and IBM Cloud Object Storage. You can use this simple app as a base to add more complex functionality and create a robust learning app that will help instructors and students improve their online learning experience.