yfiles-jupyter-graphs-for-sparql 1.0.0

The open-source adapter for working with RDF databases and SPARQL queries in Jupyter notebooks leveraging the yFiles Graphs for Jupyter plugin.

yFiles Jupyter Graphs for SPARQL

Easily visualize a SPARQL query for RDF graphs in a Jupyter Notebook.

This packages provides an easy-to-use interface to the yFiles Graphs for Jupyter widget to directly visualize queries.

Installation

Just install it from the Python Package Index

pip install yfiles_jupyter_graphs_for_sparql==0.9.0rc1

or see README_DEV.md to build it yourself.

Usage

from SPARQLWrapper import SPARQLWrapper
from yfiles_jupyter_graphs_for_sparql import SparqlGraphWidget

g = SparqlGraphWidget(wrapper=SPARQLWrapper("http://dbpedia.org/sparql"))

q = """
    SELECT ?sub ?p ?ob
    WHERE {
        ?sub ?p ?ob .
    }
    """
g.show_query(q)

See examples

Supported Environments

The widget uses yFiles Graphs for Jupyter at its core, and therefore runs in any environment that is supported by it, see supported environments.

Documentation

The main class SparqlGraphWidget provides the following API:

Constructor

• SparqlGraphWidget: Creates a new class instance with the following arguments

Argument	Description	Default
limit	The node limit which is added to all queries	50
wrapper	A SPARQL wrapper, that is used to send queries to	None
layout	Can be used to specify a general default node and edge layout. Available algorithms are: "circular", "hierarchic", "organic", "interactive_organic_layout", "orthogonal", "radial", "tree", "map", "orthogonal_edge_router", "organic_edge_router"	organic

For all arguments, there is a set_[arg] and get_[arg] method.

Methods

[!IMPORTANT]
If you want to use SELECT query types, ensure you select all three triple components—subject, predicate, and object. Otherwise, a graph cannot be constructed from the selected data. For an example look at the Getting Started notebook

• show_query(query, layout: Optional[str] = None)
  • query: The query that should be visualized.
  • layout (Optional[str]): The graph layout that is used. This overwrites the general layout in this specific graph instance. The following arguments are supported:
    • hierarchic
    • organic
    • interactive_organic_layout
    • circular
    • circular_straight_line
    • orthogonal
    • tree
    • radial
    • map
    • orthogonal_edge_router
    • organic_edge_router

To get an overview of the data structure, you can use the following function. The output is constrained by the limit property, meaning only a partial schema may be displayed depending on the dataset.

• show_schema()

The graph visualization can be adjusted by adding configurations to each node label or edge type with the following functions:

• add_subject_configuration(predicate: Union[str, list[str]], **kwargs: Dict[str, Any])

  • predicate: The predicate of the subject this configuration should be used for.
  • **kwargs: Visualization configuration for the given node label. The following arguments are supported:
    • text: The text that displayed at the node. By default, the node's label is used.
    • color: A convenience color binding for the node (see also styles argument).
    • size: The size of the node.
    • styles: A dictionary that may contain the following attributes color, shape (one of 'ellipse', ' hexagon', 'hexagon2', 'octagon', 'pill', 'rectangle', 'round-rectangle' or 'triangle'), image.
    • property: Allows to specify additional properties on the node, which may be bound by other bindings.
    • type: Defines a specific "type" for the node as described in yFiles Graphs for Jupyter which affects the automatic positioning of nodes (same "type"s are preferred to be placed next to each other).
    • parent_configuration: Configure grouping for this node label. See configurations_example.ipynb for examples.
    • heat: A heat value in between 0 and 1.

• add_object_configuration(predicate: Union[str, list[str]], **kwargs: Dict[str, Any])

  • predicate: The predicate of the object this configuration should be used for.
  • **kwargs: Visualization configuration for the given node label. The following arguments are supported:
    • text: The text that displayed at the node. By default, the node's label is used.
    • color: A convenience color binding for the node (see also styles argument).
    • size: The size of the node.
    • styles: A dictionary that may contain the following attributes color, shape (one of 'ellipse', ' hexagon', 'hexagon2', 'octagon', 'pill', 'rectangle', 'round-rectangle' or 'triangle'), image.
    • property: Allows to specify additional properties on the node, which may be bound by other bindings.
    • type: Defines a specific "type" for the node as described in yFiles Graphs for Jupyter which affects the automatic positioning of nodes (same "type"s are preferred to be placed next to each other).
    • parent_configuration: Configure grouping for this node label. See configurations_example.ipynb for examples.
    • heat: A heat value in between 0 and 1.

• add_predicate_configuration(type: Union[str, list[str]], **kwargs: Dict[str, Any])

  • type: The predicate type(s) for which this configuration should be used. Supports * to address all types.
  • **kwargs: Visualization configuration for the given predicate type. The following arguments are supported:
    • text: The text that displayed at the edge. By default, the predicate's type is used.
    • color: The edge's color.
    • thickness_factor: The edge's stroke thickness factor. By default, 1.
    • property: Allows to specify additional properties on the edge, which may be bound by other bindings.
    • heat: A heat value in between 0 and 1.

• add_parent_relationship_configuration(type: Union[str, list[str]], reverse: Optional[bool] = False) -> None

  • type: The predicate type that should be visualized as node grouping hierarchy instead of the actual relationship.
  • reverse: By default the target node is considered as parent. This can be reverted with this argument.

For a detailed documentation look at the core widget

To remove a configuration use the following functions:

• del_object_configuration(type): Deletes configuration for the given object predicate type.
• del_subject_configuration(type): Deletes configuration for the given subject predicate type.
• del_edge_configurations(type): Deletes configuration for the given predicate type.
• del_parent_predicate_configuration(type: Union[str, list[str]]) -> None: Deletes configuration for the given parent predicate type(s).

How configuration bindings are resolved

The configuration bindings (see add_object_configuration, add_subject_configuration or add_predicate_configuration) are resolved as follows:

If the configuration binding is a string, the package first tries to resolve it against the item's properties and uses the property value if available. If there is no property with the given key, the string value itself is used as a constant binding.

In case you want to create a constant string value as binding, which also happens to be a property key, use a binding function with a constant string as return value instead.

If the configuration binding is a function, the return value of the function is used as value for the respective configuration.

yFiles Graphs for Jupyter

The graph visualization is provided by yFiles Graphs for Jupyter, a versatile graph visualization widget for Jupyter Notebooks.

It can import and visualize graphs from various popular Python packages (e.g. NetworkX, PyGraphviz, igraph) or just structured node and edge lists.

And provides a rich set of visualization options to bring your data to life (see the example notebooks).

Feature Highlights

Mapping visualization	Schema visualization
Grouping visualization	Heat Mapping

For a detailed feature guide, check out the main widget example notebooks

Code of Conduct

This project and everyone participating in it is governed by the Code of Conduct. By participating, you are expected to uphold this code. Please report unacceptable behavior to [e-mail].

Feedback

This widget is by no means perfect. If you find something is not working as expected we are glad to receive an issue report from you. Please make sure to search for existing issues first and check if the issue is not an unsupported feature or known issue. If you did not find anything related, report a new issue with necessary information. Please also provide a clear and descriptive title and stick to the issue templates. See issues.

Dependencies

• yFiles Graphs for Jupyter

License

See LICENSE file.