Discourse Annotation

This applet is meant to be the prototype platform that works locally for the annotators to perform a specific annotation task that involves Reddit conversations.

Installation

Usage (First Time) :

Step 1: Create a virtual environment and activate it.

It is better to have a virtual env set up so that the other versions on your machine don't lead to nasty errors when working with this repo. You can do so using : Option 1 : Conda - Follow the steps listed here Option 2 : Terminal - Follow the stpes listed here

Step 2: Install the dependencies.

pip install requirements.txt

Step 3: Download the working dataset.

Please delete all the sample data from the data folder first before downloading the new data.Make sure that you download the data in the data folder and not elsewhere.

Sagar Sanskar

Step 4: Run the app.

streamlit run app/intro.py

Please be careful in selecting the kind of annotation you would be performing.

1. The sub_problem page in the app is dedicated to the sub-problem identification
2. The nodes_edge page in the app is dedicated to the node and edge identification
3. The info page contains all the guidlines needed for the annotation process. It is recommned to refer to this tab every once in a while after each conversation so that you get acquainted with the schema quickly.

Step 5: Update your progress in the Gsheet

Color code the ones that are done as "Green" in your specific column for the specific conversation. Link : Sagar Sanskar

Resuming your previous work :

These are the steps for when you have already done the steps above but are already how you to utilize the app but you had to take a pause and now you're back to it.

Step 6: Always activate the virtual environment

code

Step 7: Use the app as intended.

Follow the same convention i.e. Step 4 followed by Step 5.

What next, when done with one batch?

Upload it to the drive - Here Use the nomenclature based on the scheme from other members - <Your initials>_<Batch_Code>. For example, Arpit Patel would have a file name AP_A32.

Follow the annotation scheme to pick up the next batch. - Here

Roadmap

TBA

Contributing

Pull requests are welcome. For major changes, please open an issue first to discuss what you would like to change. I will soon update some pointers on how we could refine this platform for better accessibility and scalability.