Monoidal data sequences.


Keywords
data, monoids, sugarcube, transformation, data-mining, data-preservation, data-team, human-rights, javascript
License
GPL-3.0
Install
npm install @sugarcube/core@0.42.0

Documentation

SugarCube

License: GPL v3 Build Status Coverage Status

Synopsis

SugarCube is a framework to fetch, transform and publish data. Data processes are described using plugins, which are chained in sequence to transform any sort of data. It is used to support data based investigations.

The best way to get started is the tutorial. See the glossary for an explanation of terms and concepts of SugarCube. If you want to contribute, please take a look here or write an email. This code is licensed under the GPL 3.

Development

This is a Lerna mono-repo.

To setup the project, run yarn setup. This will execute yarn install and lerna bootstrap.

To link all packages into the local run lerna exec yarn install.

To build all packages run yarn build. It's possible to build only specific packages using yarn build -- cli core.

All packages are linted using yarn lint. The coding standard is enforced by Prettier. Run yarn fix to rewrite your code to follow the coding standard.

Run the tests with yarn test.

Note: Sometimes a test is failing. This often happens in the tests regarding the state. This is related to jsverify. The error thrown is TypeError: str.split is not a function. It doesn't happen too often though.

When committing to master, follow the conventional commits guidelines.

To prepare a new release, login first to NPM with npm adduser --scope=@sugarcube. This will write ~/.npmrc. I usually copy the file into the project root. In the end my .npmrc looks somthing like that:

access=public
@sugarcube:registry=https://registry.npmjs.org/
//registry.npmjs.org/:_authToken=<YOUR AUTH TOKEN>

To actually release a new version, simply run yarn release.

To develop the following host dependencies are required when running yarn setup.

  • Java JDK (7 or 8) for plugin-tika.

To install those dependencies run:

  • On Archlinux: pacman -S jdk8-openjdk
  • On Debian: apt install openjdk-7-jdk
  • Using Homebrew: brew cask install java

Compilation

There is a compile run target that compiles all packages. Provide the name of one or more package names to only compile those.

yarn compile
yarn compile plugin-tika cli core

Using the watch target a single package can be watched for file changes that trigger a compilation.

yarn watch plugin-tika

Testing

This repository contains an example project, to quickly try pipelines during development. Use yarn over npm to avoid an error about not found paths. jq parses the JSON file to extract all dependencies and outputs them as a text stream. To set it up with the current development version:

yarn setup
cd project
mkdir -p node_modules/.bin
cd node_modules
ln -sf ../../packages @sugarcube
cd .bin
ln -sf ../@sugarcube/cli/bin/sugarcube .

The example project provides a Vagrantfile as well. It boots a VM with MongoDB, Elasticsearch and NodeJS installed.

cd project
vagrant box add debian/stretch64
vagrant up

Test the DB connections:

mongo mongodb://localhost:27007
curl -X GET http://localhost:9200/_cluster/health?pretty

Scripts

The ./scripts folder contains some programs I used in the past. Install all dependencies.

To use a program run it from the project root.

./scripts/count-imports.js dashp
  • count-imports.js :: Count the usage of a dependency in the SuagrCube code base. The program takes the name of the dependency as it's arguments and prints a count of the imports, e.g. ./scripts/count-imports lodash/fp. This script is currently not working.
  • make-docs.js :: Generate the documentation from the different packages and place them in the docs directory.