April

Goobi viewer Digest for April 2020

Coming soon

  • 🗺 Maps

  • 🛠 improved usabillity in the Admin Backend

  • 💻 REST API

General

This month, we will mainly report on the development itself. The new possibilities with the maps are a real highlight here!

It's important to note that the update to version 4.6 also requires an update to the Goobi viewer Indexer and the re-indexing of the dataset. Otherwise, due to a change in logic, not all the options will be displayed in the "Citation and reuse" widget.

Developments

Maps

In the backend there is a new section "Maps". There you can create and manage maps either manually or based on a Solr query. These are then available in the new CMS template "Text and map" or in all templates as widget in the sidebar.

Overview of the existing maps in the backend

When defining maps manually, any number of markers can be set. One title and one text can be entered per marker. In addition to the support of multilingualism, HTML can also be used in the text, for example to add links to other pages.

Markers and contents of the popup can be defined manually.

If the Solr search index contains records with indexed coordinates, a map can also be generated on this basis. A marker is displayed for each hit. If there are several hits for one point, they are automatically combined.

Maps can be automatically generated based on indexed coordinates.

With the new "Maps" widget for the sidebar, the previous "Geocoordinates" widget has been completely removed. A migration of existing configurations to the new map functionality does not exist. If the "Geocoordinates" widget was in use, the configured points must be recreated as a map after an update and added to the sidebar with the new widget. Since using the geocoordinates widget with manual input of latitude and longitude values was very cumbersome, we assume that only very few installations are affected.

Example for displaying a new map in the sidebar

Records without images

Support for records without images or structural data has been improved once again. True to the motto "Show only what makes sense" the existing behavior was evaluated and various changes were implemented.

Up to now, for example, a placeholder image was displayed in the search results with the hint that no image could be loaded. In installations that only contain records without images, this of course makes no sense. Instead, the thumbnail for the respective records is now simply hidden completely and more space is given to the metadata next to it.

If there are no images, the "Bibliographic data" page has always been called up. If there is no other page for the entire record, the title of the record is now displayed here instead of "Bibliographic Data".

Example for the display of a record consisting of metadata only

Browse

In the browse functionality only the indexed value could be offered so far. This is no problem with metadata like person or place names. However, if language codes according to ISO-639-2b such as ger, eng, fre or country codes according to ISO-3166-1-alpha2 such as DE, EN, FR are stored internally, browsing here made little sense.

With the new attribute translate="true" in the field definition for browsing, these fields can now be translated.

config_viewer.xml
<metadata>
<browsingMenu>
<luceneField translate="true">MD_LANGUAGE</luceneField>
</browsingMenu>
</metadata>

Attention: The translation requires computing time and should for performance reasons only be used for smaller data volumes, such as the authority data examples given above.

The functionality of the citation links for image sections, which was introduced in January, has been improved once again in its operation. After activating the functionality via the icon above the image display, the generated links are also displayed in the same popup and can be copied to the clipboard by clicking.

Additionally, the size of the frame can now be changed until the popover is closed.

Links are now displayed in popovers and the frames can be changed.

In the full screen display the functionality is now available in the sidebar:

Citation links for image sections can be selected in the full screen display in the sidebar.

PDF files

The Goobi viewer generates PDF files "on-the-fly" and accesses the content server to do so. For some time now, there has been criticism in the community that the PDF files generated can quickly become very large.

With version 4.6, the Goobi viewer uses a new version of ContentServer that includes two new optional switches to influence the size of the generated PDF files:

config_contentServer.xml
imageScale="0.5"
imageCompression="50"

The switch imageScale="" specifies how much the image should be scaled compared to its original size. The switch imageCompression="" defines how much the JPEG files are compressed. The values given above are the new default values.

In numbers, this means that our test record with 120 images and stored full text is previously 514.3MB and using the above switches, it is now only 88.7MB.

The size of the generated PDF file could be reduced by. 83 percent!

Grouping of results on CMS pages

Two users jointly commissioned the grouping of hits on CMS pages. This new functionality is available in the CMS templates "Text with search results" and "Accordion collection view". If a grouping field is selected, the contents are displayed summarized according to the values of the field, for example by year or by sub-theme. Fields where multiple values are possible are not available for grouping.

Grouping of collections by partner
Grouping of records by year of publication

Improved registration of user accounts

At the beginning of May, the mass automatic registration of SPAM user accounts was observed in many Goobi viewer installations. As a short-term solution, the sending of emails in the Goobi viewer had to be deactivated.

As a direct consequence of the incident, various improvements were integrated to prevent automatic registration in future.

The only visible change is a security question in the user registration form. The questions are deliberately trivial and can be answered very easily by a human. SPAM bots fail because of this, however. One advantage of this variant is the accessibility: Captchas with images are only accessible if they are also offered as audio files. We wanted to do without external services.

Invisible innovations are for example a so-called "honeypot field". Therefore an additional field was integrated into the form, which is hidden in the user interface via styling instructions. So people cannot fill out the field because they cannot see it. SPAM bots are based on the page source code. So if the field is filled out, it must be a SPAM bot at registration. In addition, further checks were added to the source code and logging was added, which would have helped with the diagnosis in early May.

The pool of security questions and allowed answers can be adjusted in the configuration file. Three questions and corresponding answers are delivered by default. Further suggestions for corresponding questions and an explanation of the configuration can be found in the documentation chapter 2.5.3.

In the form for the registration of new user accounts there is a new security question

CMS: Improved FAQ templates

The CMS templates for creating an FAQ have been revised. FAQ articles can now also contain images. The inline help was completely rewritten and the styling of the pages was improved.

Timeline

One of the entrances to the contents of the database is the timeline. The functionality was outdated in its current form. Operation was cumbersome, loading times were too long and users didn't know if and when something happened.

For this reason, the timeline was given a spring clean and now shines in new splendor. The technological basis is now modern, the display is much slimmer and contents are displayed much faster. With the possibility to set the maximum number of hits yourself, a new functionality has been added.

A new CMS template is also available for adapting the introductory text.

The timeline has been renewed and is much more comfortable to use.

Other

  • An initial implementation for SAML has been merged into the Core and can be configured as a new IdentityProvider, for example for Shibboleth. It may be necessary to make further adjustments to the code here during local commissioning.

  • It is now distinguished when displaying full texts whether the files are in text format or HTML. The display of HTML has been optimized.

  • In the widget "Cite and Reuse", an additional icon is now displayed in the badge for the PDF download in order to visually find this important functionality faster.

  • In the "Cite and Reuse" widget, the name of the structural element is now displayed instead of the heading "Structural element", for example "Chapter", "Title page", "Cover" or "Table of contents".

Goobi viewer Indexer

The Goobi viewer Indexer has been extended to include the option of indexing geocoordinates. See also chapter 3.7.19 in the documentation.

Version numbers

The versions that must be entered in the pom.xml of the theme in order to get the functions described in this digest are:

<dependency>
<groupId>io.goobi.viewer</groupId>
<artifactId>viewer-core</artifactId>
<version>4.6.1</version>
</dependency>
<dependency>
<groupId>io.goobi.viewer</groupId>
<artifactId>viewer-core-config</artifactId>
<version>4.6.3</version>
</dependency>

The Goobi viewer Indexer has the version number 4.6.0.

The Goobi viewer Connector has the version number 4.6.0.

The Goobi viewer Crowdsourcing Module has the version number 1.4.0.