Semantic Result Formats

From EduTech Wiki
Jump to: navigation, search
Category:MediaWiki extension
Semantic Result Formats
Extension name Semantic Result Formats
About this article / disclaimer This is not an official page. In addition, it is just a stub. Lots of missing information....
Logo [[Image:|135px]]
Screenshot [[Image:|135px]]
Location of the main author Gent, Belgium
Coordinates of the main authors 51.054341666667, 3.717425
Developers Jeroen De Dauw, Frank Dengler, Steren Giannini, James Hong Kong, Fabian Howahl, Yaron Koren, Markus Krötzsch, David Loomer, Joel Natividad, Denny Vrandecic, Nathan Yergler, others
Licences GPL
Description Semantic Result Formats (SRF) is a MediaWiki extension, used in conjunction with the Semantic MediaWiki extension, that bundles a number of further result formats for SMW's inline queries. The individual formats can be added to the installation independently...
Mediawiki requirements Any recent version
Dependencies Semantic MediaWiki
Related extensions (documented here) Semantic Forms, Semantic Forms Inputs, Semantic Maps, Semantic MediaWiki
Related extensions
Discussion
Language support
Status stable
First release date 2010/12/10
Last release date (as of 2014/05/02!) 2014/04/04
Last version number 1.9.1 alpha (9482833)
Programming language PHP
Alternatives
Website home page
Publications
Support websites web site
Example websites
Last edited 2014/05/02


1 Introduction

Semantic Result Formats (SRF) is a MediaWiki extension, used in conjunction with the Semantic MediaWiki extension, that bundles a number of further result formats for SMW's inline queries. The individual formats can be added to the installation independently. (Mediawiki.org extension page, June 12 2013).

For installation tips, read the Semantic MediaWiki article, i.e. we suggest to use the composer dependency manager.

In order to understand result formats, you must understand how to formulate queries. The Semantic MediaWiki page includes a short introduction. Better, read the official documentation, in particular: Selecting pages. In a nutshell, you should know that queries return lists of pages or lists of pages with lists of property values.

Examples in this page use "real" data, i.e. all the properties/values defined in the pages of the MediaWiki extension category. By the way, in this wiki, most property values are defined through templates using Semantic Forms.

1.1 Ordinary SMW result formats

By default, SMW already includes a series of result formats as documented here.

Of particular interest for creating wiki pages are:

  • The table format. The default format for queries that include at least one property to display. See an example below.
  • The list format. The default format for queries that do not include properties to display. See an example below.
  • The ul and ol extend the list format to numbered and bulleted lists with additional output shown in parenthesis
  • The template format. Allows to use templates for custom formatting

In addition, there is a whole range of formats meant for download. For example

  • The CSV format could be useful for exporting a query result for further analysis

All or (most?) formats share the same base parameters:

Parameter Type Default Description
source text empty Alternative query source
limit whole number 50 The maximum number of results to return
offset whole number 0 The offset of the first result
link text all Show values as links
sort list of texts empty Property to sort the query by
order list of texts empty Order of the query sort
headers text show Display the headers/property names
mainlabel text no The label to give to the main page name
intro text empty The text to display before the query results, if there are any
outro text empty The text to display after the query results, if there are any
searchlabel text ... further results Text for continuing the search
default text empty The text to display if there are no query results

The most important parameter could be mainlabel. Use

mainlabel=-

to suppress printing the first result column which is the wiki page that includes a matching query

 mainlabel=Some sort of title

to add a title to the first column.

Supressing the first column is something that also applies to charts produced by Semantic Result Formats extensions ! Do not forget !

1.2 The table format

Below is a simple example of a table with three columns: Names of matched pages (i.e. all pages in the category) and values of the Is developed by and Is related to properties found in these pages.

This code:

{{#ask: [[Category:MediaWiki extension]]
| ?Is developed by
| ?Is related to
| limit=3
}}

... will produce this table:

 Is developed byIs related to
Maps (MediaWiki extension)Jeroen De Dauw
others
Mediawiki collection extension installationPediaPress.com
Siebrand Mazeland
Marcin Cieślak
Page FormsYaron Koren
Stephan Gambke
others
... further results

The table format can take additional parameters.

Parameter Type Default Description
class text sortable wikitable smwtable An additional CSS class to set for the table
transpose yes/no no Display table headers vertically and results horizontally
sep text empty The separator for values

Now, semantic Result Formats adds other visualization types. Most of these are based on various JavaScript libraries, such as D3. Read on ....

1.3 List format

The list format takes the following extra parameters:

Parameter Type Default Description
sep text , The separator for values
template text empty The name of a template with which to display the printouts
named args yes/no no Name the arguments passed to the template
userparam text empty A value passed into each template call, if a template is used
introtemplate text empty The name of a template to display before the query results, if there are any
outrotemplate text empty The name of a template to display after the query results, if there are any
import-annotation yes/no no <smw-paramdesc-import-annotation>

The following query gets all pages in the MediaWiki extension and displays both the hit (i.e. the page) and the value of Is developed by property:

list of names: {{#ask: [[Category:MediaWiki extension]]
| ?Is developed by
| format=list
}}

... will produce this list of names: Maps (MediaWiki extension) (Is developed by Jeroen De Dauw, others), Mediawiki collection extension installation (Is developed by PediaPress.com, Siebrand Mazeland, Marcin Cieślak), Page Forms (Is developed by Yaron Koren, Stephan Gambke, others), Semantic Drilldown (Is developed by Yaron Koren, David Loomer), Semantic Forms Inputs (Is developed by Stephan Gambke, Yaron Koren, Jeroen De Dauw, Sanyam Goyal, Yury Katkov, others), Semantic Maps (Is developed by Jeroen De Dauw, others), Semantic MediaWiki (Is developed by Markus Krötzsch, Denny Vrandecic, Jeroen De Dauw, others), Semantic Result Formats (Is developed by Jeroen De Dauw, Frank Dengler, Steren Giannini, James Hong Kong, Fabian Howahl, Yaron Koren, Markus Krötzsch, David Loomer, Joel Natividad, Denny Vrandecic, Nathan Yergler, others), VisualEditor You can see that we have to require list format, else we would have gotten a table

The following query gets all pages in the MediaWiki extension that include a Is developed by property. It will only display the (useless) Is developed by values.

list of names: {{#ask: [[Is developed by::+]][[Category:MediaWiki extension]]
|?Is developed by
|mainlabel=-
}}

... will produce this list of names: Is developed by Jeroen De Dauw, others, Is developed by PediaPress.com, Siebrand Mazeland, Marcin Cieślak, Is developed by Yaron Koren, Stephan Gambke, others, Is developed by Yaron Koren, David Loomer, Is developed by Stephan Gambke, Yaron Koren, Jeroen De Dauw, Sanyam Goyal, Yury Katkov, others, Is developed by Jeroen De Dauw, others, Is developed by Markus Krötzsch, Denny Vrandecic, Jeroen De Dauw, others, Is developed by Jeroen De Dauw, Frank Dengler, Steren Giannini, James Hong Kong, Fabian Howahl, Yaron Koren, Markus Krötzsch, David Loomer, Joel Natividad, Denny Vrandecic, Nathan Yergler, others

List format on steroids

For custom formatting, use templates. Read the official Template format documentation.

1.4 The CSV format

In addition to the base parameters, the CSV format takes the following:

Parameter Type Default Description
sep text , The separator to use
showsep yes/no no Show separator in top of CSV file ("sep=<value>")
filename text result.csv The name for the output file

The following produces a downloadable CSV file using the default comma (,) separator:

Download: {{#ask: [[Category:MediaWiki extension]]
| ?Depends on
| ?Is developed by
| ?Has status
| ?Is related to
| format=CSV
}}

Download: CSV

1.5 Using templates

Template format can be used to taylor output, for example:

  • Changing the order in which output is displayed, or omitting or duplicating output;
  • Displaying images depending on query results;
  • Creating links for property values;
  • Using CSS styles to vary font-size, alignment, background-color, etc. by column in tables.

2 Examples of Semantic Result Formats

I.e. (some of) the ones we use here ...

Extensions to SMW can add other result formats. All (or most) of these display formats can be obtained with this extension. However, you will have to pay attention to the configuration parameters, i.e. by default, not all Semantic Result Formats are available if you just install and load this extension. Some JavaScript libraries have to be enabled.

Also, for many types of graphics, you will have to use to the distribution=yes parameter.

Notice that the #smwdoc parser function allows for automatic parameter documentation generation for all result formats. E.g. if you are tired of looking up documentation at smenatic-mediawiki.org, you could insert something like {{#smwdoc:table}} in a wiki, page like we did here (look at the source of this page)

2.1 Simple lists

There are three variants: outline, ol and ul

Outline format

Parameter Type Default Description
outlineproperties list of texts empty The list of properties to be displayed as outline headers, separated by commas

ol format

Parameter Type Default Description
sep text , The separator for values
template text empty The name of a template with which to display the printouts
named args yes/no no Name the arguments passed to the template
columns whole number 1 The number of columns in which to display results (default is $1)
userparam text empty A value passed into each template call, if a template is used
introtemplate text empty The name of a template to display before the query results, if there are any
outrotemplate text empty The name of a template to display after the query results, if there are any
import-annotation yes/no no <smw-paramdesc-import-annotation>

This code

{{#ask: [[Category:Software information]]
| ?Is developed by
| format=outline
}}

produces this:


This code

{{#ask: [[Category:Software information]]
| ?Is developed by
| format=ol
}}

produces this:

  1. Mediawiki (Is developed by MediaWiki foundation and MediaWiki volunteers)
  2. Some software (Is developed by Many)

2.2 Tagcloud

Parameter Type Default Description
template text empty <srf-paramdesc-template>
userparam text empty <srf-paramdesc-userparam>
excludetags text empty Exclude tags (delimiter: ";")
includesubject yes/no no <srf-paramdesc-includesubject>
tagorder text alphabetical The order of the tags
increase text log How to increase the size of tags
widget text empty Available widget
class text empty Specify an additional cascading style sheet class
font text impact <srf-paramdesc-font>
height whole number 400 Height
width whole number 400 Width
mincount whole number 1 The minimum amount of times a value needs to occur to be listed
minsize whole number 77 The size of the smallest tags in percent
maxsize whole number 242 The size of the biggest tags in percent
maxtags whole number 1000 The maximum amount of tags in the cloud

The following:

{{#ask: [[Category:MediaWiki extension]]
| ?Is developed by
| format=tagcloud
}}

produces:


David Loomer Denny Vrandecic Fabian Howahl Frank Dengler James Hong Kong Jeroen De Dauw Joel Natividad Marcin Cieślak Markus Krötzsch Nathan Yergler others PediaPress.com Sanyam Goyal Siebrand Mazeland Stephan Gambke Steren Giannini Yaron Koren Yury Katkov

The following:

{{#ask: [[Category:MediaWiki extension]]
| ?Is developed by
| format=tagcloud
| widget=wordcloud
}}

produces:

Loading...
  • David Loomer
  • Denny Vrandecic
  • Fabian Howahl
  • Frank Dengler
  • James Hong Kong
  • Jeroen De Dauw
  • Joel Natividad
  • Marcin Cieślak
  • Markus Krötzsch
  • Nathan Yergler
  • others
  • PediaPress.com
  • Sanyam Goyal
  • Siebrand Mazeland
  • Stephan Gambke
  • Steren Giannini
  • Yaron Koren
  • Yury Katkov

2.3 D3 Bubble chart

Official documentation:

List of parameters:

Parameter Type Default Description
distribution yes/no no Instead of displaying all values, count their occurrences, and show these.
distributionsort text none Sort the value distribution by occurrence count.
distributionlimit whole number no Limit the value distribution to the count of only some values.
aggregation text subject <smw-paramdesc-aggregation>
min whole number no The minimum value to show on the Y-axis
charttype text treemap Available chart type
height whole number 400 Specify the height (in pixels) of a chart or graph
width text 100% Specify the width (in pixels or percent) of a chart or graph
charttitle text empty The title of the chart
charttext text empty Descriptive chart text
class text empty Specify an additional cascading style sheet class
datalabels text none Chart/graph data labels
colorscheme text empty Select a color scheme
chartcolor text empty Assign individual chart colors
{{#ask: [[Category:MediaWiki extension]]
| ?Is developed by
|format=d3chart
|distribution=yes
|mainlabel=-
|charttype=bubble
|headers=show
|link=all
|charttitle= SMW developers
|datalabels=chart
}}

Notice the use of the following parameters:

  • distribution=yes will count occurences
  • mainlabel=- will exclude the count of the pages themselves

We roughly get the same information as in the word cloud.

Loading...

The result of not using mainlabel=- is below:

{{#ask: [[Category:MediaWiki extension]][[Is developed by::+]]
| ?Is developed by
|format=d3chart
|distribution=yes
|charttype=bubble
|headers=show
|link=all
|charttitle= SMW developers
|datalabels=chart
}}

Useless resulting chart:

Loading...

2.4 JqPlotChart

The Jpplotchart result format allows to create five different types of charts: bar, line, donut, bubble and scatter. Each of these requires a certain data model. Data sets can grouped depending of its source.

Read Help:Jqplotseries format

This extension has many parameters. Some only make sense with certain types of charts.

Parameter Type Default Description
distribution yes/no no Instead of displaying all values, count their occurrences, and show these.
distributionsort text none Sort the value distribution by occurrence count.
distributionlimit whole number no Limit the value distribution to the count of only some values.
aggregation text subject <smw-paramdesc-aggregation>
min whole number no The minimum value to show on the Y-axis
direction text vertical Specify the direction of a chart or graph
charttitle text empty The title of the chart
charttext text empty Descriptive chart text
numbersaxislabel text empty The label for the numbers axis
labelaxislabel text empty The label for the label axis
height whole number 400 Specify the height (in pixels) of a chart or graph
width text 100% Specify the width (in pixels or percent) of a chart or graph
smoothlines yes/no no Apply a smoothing algorithm on line charts
valueformat text %d Specify formatting rule for values
ticklabels yes/no yes Enable display of tick labels
highlighter yes/no no Display a data point highlighter
theme text empty Select a grid theme
filling yes/no yes Individual filling option
chartlegend text none Chart legend position
datalabels text none Chart/graph data labels
colorscheme text empty Select a color scheme
chartcolor text empty Assign individual chart colors
class text empty Specify an additional cascading style sheet class
charttype text bar Available chart type

The following example doesn't use numerical data values, but counts occurences using the parameter distribution=yes-

2.4.1 Donuts

{{#ask:[[Category:MediaWiki extension]][[Is related to::+]]
|?Is related to
|mainlabel=-
|format=jqplotchart
|charttype=donut
|chartlegend=e
|charttitle= Related extensions (%)
|distribution=yes
|width=33%
|datalabels=percent
}}
Loading...

2.4.2 Bar chart

{{#ask:[[Category:MediaWiki extension]][[Is related to::+]]
|?Is related to
|mainlabel=-
|format=jqplotchart
|charttype=bar
|charttitle= Related extensions
|distribution=yes
|width=33%
|datalabels=percent
}}
Loading...


To display more than one property in a chart, it's probably best to group (else you get all the data in the same ring or bar). However, you need numerical data for that (I believe ...)

2.5 Using templates

Achievement badges

The following code was used to dislay achievement badges for the Bases psychopédagogiques des technologies éducatives in the french sister site.

Query example that could be inserted in a user page:

{{#ask: [[REPORTING:+]] [[catégorie_principale::catégorie:Bases psychopédagogiques des technologies éducatives]] [[a intention de produire::vrai]] [[A taux_d'achèvement::100]] [[Agent::user:{{PAGENAME}}]]
|? agent
|? page wiki
|? A taux_d'achèvement
|? Promotion
| distribution=yes
| mainlabel=-
| format=template
| template=badge screencast completed
| userparam=show_page
| default=Vous n'avez pas encore gagné de badge(s) d'achèvement.
}}

Code for the badge screencast completed template

<includeonly>
<div class="badge">
[[image:badge_screencast_completed.png|128px]]
<div class="appear">
'''Badge d'achèvement'''<br/>
Carte conceptuelle et infographie
«{{{2|}}}» complétées
par {{{1|}}} - {{{3|}}} %

Délivré par [[user:Daniel K. Schneider|DKS]], EduTechWiki(fr), [http://tecfa.unige.ch TECFA], [http://www.unige.ch UNIGE], cours [[Bases psychopédagogiques des technologies éducatives|Bases]], promotion {{{4|}}}</div>
{{#ifeq:{{{userparam}}}|show_user|<span style="font-size:smaller">{{{1}}}</span>|}}{{#ifeq:{{{userparam}}}|show_page|<span style="font-size:smaller">{{{2}}}</span>|}}</div>
</includeonly>

Tables with pictures

{{#ask: [[Category:data mining and learning analytics tools]] [[Analysis orientation::General analysis]]
| ?Has logo #-
| ?Has name = Name
| ?Has last revision number=
| ?Was last released on#iso
| ?License type = License
| ?Data tool type = Tool type
| ?Data manipulation type = 
| ?Has website=
| ?Last edition=
| mainlabel=General purpose analysis
| format=template
| template= Table with image
| introtemplate= Table with image/intro
| outrotemplate= CS Table with img/outro
| limit=3
|}}

produces:

DM and LA tools license Tool type Data manipulation Access Page edition
No image.png
Free&Closed source Data analysis http://www.laurenceanthony.net/software.html#antconc 14 November 2014
Mahout-logo-brudman.png
Free&Open source Framework/Library/API Data analysis http://mahout.apache.org/ 2 October 2014
OpenNLP.png
Apache OpenNLP 1.5.3 (2013-04-13)
Free&Open source Framework/Library/API Data extraction, Data transformation, Data analysis http://opennlp.apache.org/index.html 21 March 2014

... further results

3 Tips

3.1 Getting it to work with MW 1.24

  1. Install like for 1.22+ as explained in the official doc. Make sure to install recent GraphViz and Image Map extensions.
  2. Fix a JQuery problem: As explained in the release notes for MW 1.24, jQuery was upgraded from 1.8.3 to 1.11.1. This will break the jqplot library. To fix the problem, do the two following:
  • Add $wgIncludejQueryMigrate = true; to LocalSettings.php

3.2 Doesn't seem to take all data

By default an inline query made with #ask only returns 50 results. This number can be misleading. If you do a query that aggregates pages (using distribution, sum, count, etc.), "limit" concerns the number of the pages you will retrieve before passing their data to a result format. Limit does not specify the number of pages (or other elements) that are on display !

This behavior can be changed by setting smwgQDefaultLimit to higher value in LocalSettings.php

Alternatively, use limit= in the query. In the example below, we will see one chart element for each page, but there is a huge multiple of pages found. If limit is too low, results will simply be wrong !

{{#ask: [[REPORTING:+]] [[agent::+]] [[catégorie_principale::catégorie:Bases psychopédagogiques des technologies éducatives]] 
|?agent
|mainlabel=-
|distribution=yes
|format=d3chart
|charttype=treemap
|height=800
|width=100%
|headers=show
|link=all
|min=1
|limit=500
|charttitle= N pages annotées
|datalabels=chart

4 Links

Official

Tutorials

Depends onSemantic MediaWiki +
Has about meta informationThis is not an official page. In addition, it is just a stub. Lots of missing information.... +
Has addressGent, Belgium +
Has descriptionSemantic Result Formats (SRF) is a MediaWi
Semantic Result Formats (SRF) is a MediaWiki extension, used in conjunction with the Semantic MediaWiki extension, that bundles a number of further result formats for SMW's inline queries. The individual formats can be added to the installation independently...
added to the installation independently... +
Has last revision number1.9.1 alpha (9482833) +
Has licensesGPL +
Has location51° 3' 16", 3° 43' 3"Latitude: 51.054341666667
Longitude: 3.717425
+
Has nameSemantic Result Formats +
Has statusstable +
Has support websitehttp://www.semantic-mediawiki.org/wiki/Semantic_Result_Formats +
Has websitehttp://www.semantic-mediawiki.org/wiki/Semantic_Result_Formats +
Is developed byJeroen De Dauw +, Frank Dengler +, Steren Giannini +, James Hong Kong +, Fabian Howahl +, Yaron Koren +, Markus Krötzsch +, David Loomer +, Joel Natividad +, Denny Vrandecic +, Nathan Yergler + and others +
Is related toPage Forms +, Semantic Forms Inputs +, Semantic Maps + and Semantic MediaWiki +
Needs MediaWiki versionsAny recent version +
Uses programming languagePHP +
Was first released onDecember 10, 2010 +
Was last editedMay 2, 2014 +
Was last released onApril 4, 2014 +