AKSW / QuitStore

🖧 Quads in Git - Distributed Version Control for RDF Knowledge Bases
http://aksw.org/Projects/Quit
GNU General Public License v3.0
100 stars 22 forks source link
git graph quad rdf sparql sparql-endpoints version-control version-control-system versioning
The QuitStore Logo: A glass of quinch jam (German: Quittenmarmelade) with the Git logo on the lid. 'Graph jam in a git glass'

Quit Store

Build status of master branch:

Build Status Coverage Status

The Quit Store (stands for Quads in Git) provides a workspace for distributed collaborative Linked Data knowledge engineering. You are able to read and write RDF Datasets (aka. multiple Named Graphs) through a standard SPARQL 1.1 Query and Update interface. To collaborate you can create multiple branches of the Dataset and share your repository with your collaborators as you know it from Git.

If you want to read more about the Quit Store we can recommend our paper:

Decentralized Collaborative Knowledge Management using Git by Natanael Arndt, Patrick Naumann, Norman Radtke, Michael Martin, and Edgard Marx in Journal of Web Semantics, 2018 [@sciencedirect] [@arXiv]

Getting Started

To get the Quit Store you have three options:

Installation via pipx

$ pipx install git+https://github.com/AKSW/QuitStore.git
  installed package quit 0.25.4, installed using Python 3.12.1
  These apps are now globally available
    - quitstore
done! ✨ 🌟 ✨
$ quitstore --help

Installation from Source

Install poetry.

Get the Quit Store source code:

$ git clone https://github.com/AKSW/QuitStore.git
$ cd QuitStore

If you are using virtualenvwrapper:

$ poetry install
$ poetry run quitstore --help

Git configuration

Configure your name and email for Git. This information will be stored in each commit you are creating with Git and the Quit Store on your system. It is relevant so people know which contribution is coming from whom. Execute the following command if you haven't done that before.

$ git config --global user.name "Your Name"
$ git config --global user.email "you@e-mail-provider.org"

Start with Existing Data (Optional)

If you already have data which you want to use in the quit store follow these steps:

  1. Create a repository which will contain your RDF data.
$ git init /path/to/repo
  1. Put your RDF data formatted as N-Triples and sorted (e.g. using cat data-in.nt | LC_ALL=C sort -u > data-out.nt) into files like <graph>.nt into this directory.
  2. For each <graph>.nt file create a corresponding <graph>.nt.graph file which must contain the IRI for the respective graph. (These .graph files are also used by the Virtuoso bulk loading process).
  3. Add the data to the repository and create a commit.
$ git add …
$ git commit -m "init repository"

To ingest further versions of your data into the Quit Store you can add further commits by going through steps 2.-4.. Alternatively you are also able to execute SPARQL 1.1. Update operations to create new versions on the Quit Store.

Start the Quit Store

If you are using the binary:

$ chmod +x quit #
$ ./quit -t /path/to/repo

If you have it installed from the sources:

$ poetry run quitstore -t /path/to/repo

Open your browser and go to http://localhost:5000/.

Have a lot of fun!

For more command line options check out the section Command Line Options in the README.

Command Line Options

-b, --basepath

Specify a base path/application root. This will work with WSGI and docker only.

-t, --targetdir

Specify a target directory where the repository can be found or will be cloned (if remote is given) to.

-r, -repourl

Specify a link/URL to a remote repository.

-c, --configfile

Specify a path to a configuration file. (Defaults to ./config.ttl)

-nv, --disableversioning

Run Quit-Store without versioning activated

-f, --features

This option enables additional features of the store:

-v, --verbose and -vv, --verboseverbose

Set the log level for the standard output to verbose (INFO) respective extra verbose (DEBUG).

-l, --logfile

Write the log output to the given path. The path is interpreted relative to the current working directory. The log level for the logfile is always extra verbose (DEBUG).

Configuration File

deprecated (we plan to remove the configuration file feature)

If you want to work with configuration files you can create a config.ttl file. This configuration file consists of two parts, the store configuration and the graph configuration. The store configuration manages everything related to initializing the software, the graph configuration maps graph files to their graph IRIs. The graph configuration in the config.ttl is an alternative to using <graph>.nt.graph files next to the graphs. Make sure you put the correct path to your git repository ("../store") and the IRI of your graph (<http://example.org/>) and name of the file holding this graph ("example.nt").

conf:store a <YourQuitStore> ;
    <pathOfGitRepo> "../store" ; # Set the path to the repository that contains the files .
    <origin> "git:github.com/your/repository.git" . # Optional a git repo that will be cloned into dir given in line above on startup.

conf:example a <Graph> ; # Define a Graph resource for a named graph
    <graphUri> <http://example.org/> ; # Set the IRI of named graph
    <isVersioned> 1 ; # Defaults to True, future work
    <graphFile> "example.nt" . # Set the filename

API

The Quit-Store comes with three kinds of interfaces, a SPARQL update and query interface, a provenance interface, and a Git management interface.

SPARQL Update and Query Interface

The SPARQL interface support update and select queries and is meant to adhere to the SPARQL 1.1 Protocol. You can find the interface to query the current HEAD of your repository under http://your-quit-host/sparql. To access any branch or commit on the repository you can query the endpoints under http://your-quit-host/sparql/<branchname> resp. http://your-quit-host/sparql/<commitid>. Since the software is still under development there might be some missing features or strange behavior. If you are sure that the store does not follow the W3C recommendation please file an issue.

Examples

Execute a select query with curl

curl -d "select ?s ?p ?o ?g where { graph ?g { ?s ?p ?o} }" -H "Content-Type: application/sparql-query" http://your-quit-host/sparql
curl -d "select ?s ?p ?o ?g where { graph ?g { ?s ?p ?o} }" -H "Content-Type: application/sparql-query" http://your-quit-host/sparql/develop

If you are interested in a specific result mime type you can use the content negotiation feature of the interface:

curl -d "select ?s ?p ?o ?g where { graph ?g { ?s ?p ?o} }" -H "Content-Type: application/sparql-query" -H "Accept: application/sparql-results+json" http://your-quit-host/sparql

Execute an update query with curl

curl -d "insert data { graph <http://example.org/> { <urn:a> <urn:b> <urn:c> } }" -H "Content-Type: application/sparql-update"  http://your-quit-host/sparql

Provenance Interface

To use the provenance browsing feature you have to enable it with the argument --feature=provenance. The provenance browsing feature extracts provenance meta data for the revisions and makes it available through a SPARQL endpoint and the blame interface. The provenance interface is available under the following two URLs:

Git Management Interface

The git management interface allows access to some operations of quit in conjunction with the underlying git repository. You can access them with your browser at the following paths.

Docker

We provide a Docker image for the Quit Store on the public docker hub as well as on the github docker registry. The image exposes port 8080 by default. The default user within the image is the user quit with the user id 1000. For this user a git configuration with user.name QuitStore and user.email quitstore@example.org is preset. Without any further configuration, a git repository is initialized within the container in the /data directory (owned by the default user quit).

To store the data on the host a local directory or volume is required to store the git repository. An host directory or volume can be linked to the directory /data. Make sure the quit process running with the user id 1000 within the docker container has write access to this directory.

Alternatively the user id within the container can be set using the docker run --user $UID … option. In this case you have to make sure a user.name a user.email is configure using git config within the repository (.git/config) or a git config file is mounted to /.gitconfig (to /usr/src/app/.gitconfig if you are running it with user id 1000).

Example setup with the default user:

mkdir /store/repo
sudo chown 1000 /store/repo
sudo chmod u+w /store/repo

To run the image execute the following command (maybe you have to replace docker with sudo docker):

docker run -it --name containername -p 8080:8080 -v /store/repo:/data aksw/quitstore

The following example will start the quit store in the background in the detached mode.

docker run -d --name containername -p 8080:8080 -v /store/repo:/data aksw/quitstore

Now you should be able to access the quit web interface under http://localhost:8080 and the SPARQL 1.1 interface under http://localhost:8080/sparql.

The default configuration is located in /etc/quit/config.ttl, which can also be overwritten using a respective volume or by setting the QUIT_CONFIGFILE environment variable.

Further options which can be set are:

Run the Tests

You need to have the quitstore installed from source, see section Installation from Source.

poetry run pytest

Troubleshooting

Use on Windows with restricted permissions

On Windows you might not be able to download the .exe file directly. If so, use the curl command in the power shell.

When you start the QuitStore (e.g. with quit.exe -t .) it will try to open a port that is available from outside, which will require permission by the administrator user. To open the port only locally you should start the QuitStore with:

quit.exe -t . -h localhost

The default port is 5000 (http://localhost:5000/).

Migrate from old Versions

Update to 2018-11-20 from 2018-10-29 and older

If you are migrating from an NQuads based repository, as used in older versions of the QuitStore (release 2018-10-29 and older), to an NTriples based repository (release 2018-11-20 and newer) you can use the following commands to migrate the graphs. You should know that it is possible to have multiple graphs in one NQuads file, which is not possible for NTriples files. Thus, you should make sure to have only one graph per file. You may execute the steps for each NQuads file and replace graphfile.nq according to your filenames.

sed "s/<[^<>]*> .$/./g" graphfile.nq | LC_ALL=C sort -u > graphfile.nt
mv graphfile.nq.graph graphfile.nt.graph
git rm graphfile.nq
git add graphfile.nq.graph graphfile.nt graphfile.nt.graph
git commit -m "Migrate from nq to nt"

License

Copyright (C) 2017-2022 Norman Radtke http://aksw.org/NormanRadtke, Natanael Arndt http://aksw.org/NatanaelArndt, and contributors

This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 3 of the License, or (at your option) any later version.

This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.

You should have received a copy of the GNU General Public License along with this program; if not, see http://www.gnu.org/licenses. Please see LICENSE for further information.