Doc2Dial

Document-Grounded Dialog Composition Framework is an end-to-end framework for generating conversational data grounded in the documents via crowdsourcing.

NOTE: more code release in Auguest 2019.

Prerequisites

MongoDB: On MacOS, brew install mongodb-community@4.0. Please refer to MongoDB community edition installation instructions for other platforms.
On MacOS, homebrew will automatically start the mongodb service. If not, open terminal & run sudo mongod from home directory.

Installation

This package requires Python 3.6 or higher. We recommend creating a new virtual environment for this project (using virtualenv or conda).

Run the following commands:

pip install -r requirements.txt

Get Started

Prepare database

Please import the COLLNAME.json files to your local mongodb (use name "demodb") with commands such as,

mongoimport --db demodb --collection COLLNAME --drop --file COLLNAME.json

We are in process of acquiring appropriate licences for the data consist of sample documents and crowd-sourced annotations. We plan to release it next couple of weeks. We will be releasing code to push that data in to a MongoDB instance as well.

Launch

Run the following commands:

python3 run.py -d demo -p 8081

Explore

Go to http://localhost:8081
Login as an admin with password o replaced with 0 as password
Please make sure you select a task to explore before clicking Start

Live Demo

You can visit https://ibm.biz/doc2dial to see the Doc2Dial in action.

There is also a short video on YouTube demonstrating doc2dial's capabilities.

Troubleshoots and Contributing

If you have any questions, bug reports, and feature requests, please open an issue on Github.

We appreciate any kind of feedback or contribution. Feel free to proceed with small issues like bug fixes, documentation improvement. For major contributions and new features, please discuss with the collaborators in corresponding issues.

songfeng / doc2dial_pub