brwsr

Lightweight Linked Data Browser

Features

• Browse all resources starting with a designated URI prefix in a SPARQL endpoint
• Implements content negotiation to serve representations of these resources both as HTML, and as RDF/XML, Turtle and JSON-LD
• HTML page asynchronously calls http://preflabel.org service to retrieve preferred labels for all resources.

In short, it is a very lightweight Python-based alternative to Pubby (with a slightly more attractive interface design)

Docker-based

Prerequisites

• Make sure you have Docker and docker-compose installed for your platform

Using

• To use the base setup, copy this docker-compose file to a directory: https://raw.githubusercontent.com/Data2Semantics/brwsr/master/docker-compose.yml.
• Run docker-compose up.
• This will run brwsr at http://localhost:5000 with the DBPedia SPARQL endpoint.
• To modify the configuration, change the environment variables in the docker-compose.yml file.
• For configuration parameters, see below.

As a Flask application

Installation

• Open up a terminal
• In a directory of your choice, run git clone https://github.com/Data2Semantics/brwsr.git
• (optional: setup a virtualenv and activate it)
• run pip install -r requirements.txt
• Rename config-template.py to config.py
• Make the appropriate settings in the file (see below)

Using

• Start it with python run.py if you're playing around, otherwise
• Adjust the gunicorn_config.py for your system, and start brwsr with gunicorn -c gunicorn_config.py app:app to run in daemon mode on port 5000 (behind e.g. an Apache or Nginx proxy)

Configuration Parameters

Parameter	Allowed Values	Default	Description	docker-compose.yml or config.py
LOCAL_STORE	True, False	False	Set LOCAL_STORE to True if you want brwsr to just load a (smallish) RDF file into server memory rather than operate on an external SPARQL store	both
LOCAL_FILE	File path	None	Set LOCAL_FILE to the relative or absolute path of the file you want brwsr to load when LOCAL_STORE is True. The brwsr application will just use RDFLib to guess the file format based on the extension. You can use UNIX file masks such as * and ? to load multiple files. When using Docker, make sure the files are on a filesystem that is accessible to Docker	both
SPARQL_ENDPOINT	URL	http://your.sparql.endpoint.here/sparql	Set this to the SPARQL endpoint uri of your triplestore e.g. http://dbpedia.org/sparql	both
SPARQL_ENDPOINT_MAPPING	Dictionary	None	If brwsr is backed by multiple separate triple stores, use SPARQL_ENDPOINT_MAPPING to make sure that each URI for which the LOCAL_NAME (i.e. the URI with the DEFAULT_BASE removed, if present) starts with a key of the SPARQL_ENDPOINT_MAPPING dictionary, the proper SPARQL endpoint is used. You can also use Python-style regular expressions in the prefix description (the keys of this dictionary) Note that brwsr will allways also query the default SPARQL_ENDPOINT (see example in config-template.py).	only config.py
DRUID_STATEMENTS_URL	URL	None	Set this to the statements URL of a Druid instance (http://triply.cc), e.g. <"http://druid.instance.url/_api/datasets/Username/Dataset/statements.triply">	both
LDF_STATEMENTS_URL	URL	None	Set this to the statements URL of a Linked Data Fragments service (http://linkeddatafragments.org), e.g. http://data.linkeddatafragments.org/dbpedia2014	both
DEFAULT_BASE	URI	http://your.base.uri.here	The DEFAULT_BASE is the prefix of the URI's in the triple store that can be browsed by brwsr. Requests to brwsr only include the local name (i.e. the the part after the third slash '/'), the DEFAULT_BASE is always prepended to this local name to make up the URI that's used to query the triple store: e.g. http://dbpedia.org (without the last slash!)	both
LOCAL_DOCUMENT_INFIX	String	doc	The LOCAL_DOCUMENT_INFIX is the infix used between the DEFAULT_BASE and the local name of the URI to denote the HTML representation of the RDF resource (see the Cool URI's specification)	both
LOCAL_SERVER_NAME	URI	http://your.server.name.here	The LOCAL_SERVER_NAME is the address brwsr listens to. It needs to know this to build proper requests when you click a URI in the brwsr page of a resource: e.g. "http://localhost:5000" if running Flask.	both
BEHIND_PROXY	True, False	False	By default brwsr assumes it is running at the root of the server. If you want to run brwsr under a directory (e.g. http://example.com/brwsr rather than http://example.com), you need to do this via a reverse proxy, and tell brwsr about it (set BEHIND_PROXY to True, and configure the proxy, see below)	both
START_LOCAL_NAME	String	some/local/name	The START_LOCAL_NAME is the local name of the first URI shown in brwsr if no URI is specified, e.g. "resource/Amsterdam" when using the DBPedia settings	both
START_URI	URI	DEFAULT_BASE + "/" + START_LOCAL_NAME	The START_URI is the URI that is shown when browsing to the SERVER_NAME URL. It is simply the combination of the DEFAULT_BASE and the START_LOCAL_NAME (i.e. there is no need to change this, usually) e.g. this will become "http://dbpedia.org/resource/Amsterdam"	both
QUERY_RESULTS_LIMIT	Integer	5000	Set query results limit because otherwise your browser might crash.	both
BROWSE_EXTERNAL_URIS	True, False	True	Browse URIs that do not match the DEFAULT_BASE. This allows for browsing resources from different namespaces within the same endpoint store.	both
DEREFERENCE_EXTERNAL_URIS	True, False	False	Dereference external URIs (i.e. retrieve RDF served at that location, and display the resource). This may be slow, depending on the responsiveness of the server at hand. Also, the resulting RDF is stored locally (in memory) which means that this is a potential memory hog for servers that are visited frequently.	both
PORT	Integer	5000	The port via which to run brwsr	both (but needs care when using Docker)
DEBUG	True, False	False	Switch on debug logging	both
SPARQL_METHOD	GET, POST	GET	Set the HTTP method to use for communicating with SPARQL endpoint.	both
CUSTOM_PARAMETERS	Dictionary	{'reasoning': True}	Set any custom parameters to be sent to the SPARQL endpoint, e.g. CUSTOM_PARAMETERS = {'reasoning': 'true'} for Stardog	only config.py
CACHE_TIMEOUT	Integer	300	Set the cache timeout in seconds	both
SUNBURST_DEPTH	Integer	1	Depth of the Sunburst visualization (default = 1). Warning: setting this to a value > 1 will really make the visualization a lot slower, also depending on the number of endpoints, or services you are calling.	both

Example Nginx Reverse Proxy configuration for use with Gunicorn:

server {
    listen 80;
    server_name your.server.name.here;
    access_log /var/log/nginx/brwsr_access.log;
    error_log /var/log/nginx/brswr_error.log;


    location /socket.io {
        proxy_pass http://127.0.0.1:5000/socket.io;
        proxy_redirect off;
        proxy_buffering off;

        proxy_set_header Host $host;
        proxy_set_header X-Real-IP $remote_addr;
        proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;

        proxy_http_version 1.1;
        proxy_set_header Upgrade $http_upgrade;
        proxy_set_header Connection "Upgrade";
    }

    location / {
        add_header Access-Control-Allow-Origin *;

        proxy_pass http://127.0.0.1:5000;
        proxy_redirect off;

        proxy_set_header Host $host;
        proxy_set_header X-Real-IP $remote_addr;
        proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
    }
}

Running brwsr at a relative location other than '/'

If for some reason you want or need to run brwsr at a relative path other than '/', you should do two things:

1. Set the BEHIND_PROXY parameter in your config.py or docker-compose.yaml to True
2. Setup a Nginx proxy along the lines of the below example (adapted from http://flask.pocoo.org/snippets/35/):

location /myprefix {
  proxy_pass http://localhost:5000;
  proxy_set_header Host $host;
  proxy_set_header Upgrade $http_upgrade;
  proxy_set_header Connection "upgrade";
  proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
  proxy_set_header X-Scheme $scheme;
  proxy_set_header X-Script-Name /myprefix;
}

Where myprefix should be set to the location you want to be running brwsr under

The proxy_pass setting should point to the address and port you are running brwsr at (default is localhost port 5000).

Acknowledgements

This work was funded by the Dutch national programme COMMIT/ and the NWO-funded CLARIAH project.