Skip to content

Latest commit

 

History

History
72 lines (62 loc) · 2.52 KB

README.md

File metadata and controls

72 lines (62 loc) · 2.52 KB

Question Answering over Multiple and Heterogeneous Knowledge Bases

                                        Docker Python Task License

The MuHeQA (Multiple and Heterogeneous Question-Answering) system creates natural language answers from natural language questions using knowledge base from both structured (KG) and unstructured (documents) data sources.

Quick Start!

Install the muheqa package:

pip install muheqa

Create a new connection to Wikidata, or DBpedia, or D4C (Drugs4Covid). The first time it may take a few minutes to download the required models:

import muheqa.connector as mhqa

wikidata = mhqa.connect(wikidata=True)

And finally, make a question in natural language!:

response = wikidata.query("Who is the father of Barack Obama")
print("Response:",response)

Preparation

  1. Prepare a Python 3 environment and install the Conda framework.
  2. Clone this repo: 
    git clone https://github.com/librairy/MuHeQA.git
  3. Move into the root directory: 
    cd MuHeQA
  4. Create an environment (if it does not already exist): 
    conda create --name .muheqa python=3.9
  5. Activate the environment: 
    conda activate .muheqa
  6. Download the answer classifier and unzip into the root project directory. The folder resources_dir/ is created. 
    wget -O resources.zip https://delicias.dia.fi.upm.es/nextcloud/index.php/s/Jp5FeoBn57c8k4M/download
unzip resources.zip
  7. Install dependencies 
    pip install -r requirements.txt

M1 Environments (only for Apple's M1 devices)

  1. Install TensorFlow dependencies 
    conda install -c apple tensorflow-deps
  2. Install base TensorFlow 
    pip install tensorflow-macos
  3. Install tensorflow-metal plugin 
    pip install tensorflow-metal