Skip to content

Latest commit

 

History

History
130 lines (86 loc) · 4.59 KB

README.md

File metadata and controls

130 lines (86 loc) · 4.59 KB
Raw

Knesset data pipelines

Data processing pipelines for loading, processing and visualizing data about the Knesset

We are in the process of migrating to airflow, see airflow/README.md for details.

Uses the datapackage pipelines and DataFlows frameworks.

Quickstart for data science

Follow this method to get started quickly with exploration, processing and testing of the knesset data.

Running using Docker

Docker is required to run the notebooks to provide a consistent environment.

Install Docker for Windows, Mac or Linux

Pull the latest Docker image

docker pull ghcr.io/hasadna/knesset-data-pipelines/knesset-data-pipelines-legacy

Run Jupyter Lab

Create a directory which will be shared between the host PC and the container:

sudo mkdir -p /opt/knesset-data-pipelines

Start the Jupyter lab server:

docker run -it -p 8888:8888 --entrypoint jupyter \
           -v /opt/knesset-data-pipelines:/pipelines \
           ghcr.io/hasadna/knesset-data-pipelines/knesset-data-pipelines-legacy lab --allow-root --ip 0.0.0.0 --no-browser \
                --NotebookApp.token= --NotebookApp.custom_display_url=http://localhost:8888/

Access the server at http://localhost:8888/

Open a terminal inside the Jupyter Lab web-ui, and clone the knesset-data-pipelines project:

git clone https://github.com/hasadna/knesset-data-pipelines.git .

You should now see the project files on the left sidebar.

Access the jupyter-notebooks directory and open one of the available notebooks.

You can now add or make modifications to the notebooks, then open a pull request with your changes.

You can also modify the pipelines code from the host machine and it will be reflected in the notebook environment.

Running from Local copy of knesset-data-pipelines

From your local PC, clone the repository into ./knesset-data-pipelines:

git clone https://github.com/hasadna/knesset-data-pipelines.git .

Change directory:

cd knesset-data-pipelines

Run with Docker, mounting the local directory

docker run -it -p 8888:8888 --entrypoint jupyter \
           -v `pwd`:/pipelines \
           ghcr.io/hasadna/knesset-data-pipelines/knesset-data-pipelines-legacy lab --allow-root --ip 0.0.0.0 --no-browser \
                --NotebookApp.token= --NotebookApp.custom_display_url=http://localhost:8888/

When running using this setup, you might have permission problems, fix it giving yourself ownership:

sudo chown -R $USER .

Running locally without Docker

Following instructions were tested with Ubuntu 18.04

Install system dependencies:

sudo apt-get install python3.6 python3.6-dev build-essential libxml2-dev libxslt1-dev libleveldb1v5 libleveldb-dev \
                     python3-pip bash jq git openssl antiword python3-venv

Install Python dependencies:

python3.6 -m venv env
source env/bin/activate
pip install 'https://github.com/OriHoch/datapackage-pipelines/archive/1.7.1-oh-2.zip#egg=datapackage-pipelines[speedup]'
pip install wheel
pip install psycopg2-binary knesset-data requests[socks] botocore boto3 python-dotenv google-cloud-storage sh
pip install datapackage-pipelines-metrics psutil crcmod jsonpickle tika kvfile pyquery dataflows==0.0.14 pymongo \
            tabulate jupyter jupyterlab
pip install -e .

Start environment (these steps are required each time before starting to run pipelines):

source env/bin/activate
export KNESSET_PIPELINES_DATA_PATH=`pwd`/data

Now you can run pipelines with dpp or start the notebook server with jupyter lab

Contributing

Looking to contribute? check out the Help Wanted Issues or the Noob Friendly Issues for some ideas.

Useful resources for getting acquainted:

  • DPP documentation
  • Code for the periodic execution component
  • Info on available data from the Knesset site
  • Living document with short list of ongoing project activities