Presentation link

Multilingual chat application

This is a application and workshop/exercise to extend a chat application with Tornado to tranlate user messages.

Environment setup

Clone the repository

git clone https://github.com/mei-li/multiling-chat.git

Python 3.5+ is required. Also install the requirements with:

pip install -r requirements.txt

Make sure the application runs:

cd step0/
python server.py

navigate in localhost:8889 in the browser. Try the app (add text in the box, try the /nick <yournickname> command, open another window - not tab and do the same)

Hands on session.

It is a self-paced hands on session. Work alone or in a group. Each step has an exercise. Every step is relevant to te corresponding directory. Try to implement each step, or move to the next step directory that includes an implementation of the step.

For beginners: Insist on the step 0, it is independent and has a test to help develop.

Step 0: Translate function

Go inside step0 directory. Check in translate.py file. Add the correct key in the API_KEY variable. Complete translate_text function code, by calling translate API so that this function translates the given text to the target_lang. Use requests library, POST method to call Google translate API endpoint. NO NEED to use the python library. For authentication use the query parameter key=API_KEY, aka add ?key-API_KEY after the url. Run tests to check correctness.

Test:

py.test translate.py

Resources:

• Translate API docs: https://cloud.google.com/translate/docs/translating-text
• requests library docs (POST method): http://docs.python-requests.org/en/master/user/quickstart/#make-a-request

Step 1: Set language command

Go to step1 directory. In server.py edit handle_message method to add a new /lang <language_code>. Store the <language_code> in an attribute called lang inside ClientSocket object.

Tip: Add a call to system_message to show the language change to all connections/users.

Run server: (run server and test manually)

python server.py
# navigate to localhost:8888 in the browser

Step 2: Translate to user language

Go to step2 directory. Replace the correct API_KEY in translate.py. Now open server.py and edit message_all to translate the text to each connection language before sending it. Use translate_text function that is already imported.

The end

Yei!! you should have a multilingual chat application running now, like in final directory.

cd final
TRANSLATE_API_TOKEN=... python server.py

If you want more you can look at the extra steps or checkout the original application of last PyLadies workshop (by Thomas Iorns) that is the base code of this session and supports additionally shared drawing!! (connect two users and then click in the white area of the page to draw)

git clone https://github.com/mesilliac/multitude.git

Extras

• Make sure the language that the user sets is valid, check https://cloud.google.com/translate/docs/discovering-supported-languages (tip: call the API only once)
• Sent also the original/untranslated text to all users as a command_message
• Click on deploy to heroku button on top to deploy a cloud version of the application

Next steps

• get your own API_KEY or service account key and try to connect it with another google API, eg. speech to text, to get multilingual autosubtitle talks (speech to text chnarges by hour) or connect to Natural Language API to get user mood when chating. Or go to a gaming direction by enhancing its javascript (like in the original app by Thomas, https://github.com/mesilliac/multitude.git)