Publishing data using OMERO.web

The OMERO.web framework allows raw data to be published using built-in tools or supplied through web services to external web pages. Selected datasets can be made visible to a ‘public user’ using the standard OMERO permissions system, ensuring you always have control over how users can interact with your data.

There are several ways of publishing data using OMERO.web:

The sections below describe how you might use these features and how to set them up.

Configuring public user

The OMERO.web framework supports auto-login for a single username / password. This means that any public visitors to certain OMERO.web pages will be automatically logged in and will be able to access the data available to the defined ‘public user’.

To set this up on your OMERO.web installation:

  • Create a group with read-only permissions (the name can be anything e.g. “public-data”). We recommend read-only permissions so that the public user will not be able to modify, delete or annotate data belonging to other members.

  • Create a member of this group, noting the username and password (you will enter these below). Again, the First name, Last name, Username and Password can be anything you like.

    Note

    If you add this member to other groups, all data in these groups will also become publicly accessible for as long as this user remains in the group.

  • Enable the omero.web.public.enabled property and set omero.web.public.user and omero.web.public.password:

    $ omero config set omero.web.public.enabled True
    
    $ omero config set omero.web.public.user '<username>'
    
    $ omero config set omero.web.public.password '<password>'
    
  • By default the public user is only allowed to perform GET requests. This means that the public user will not be able to Create, Edit or Delete data, as these require POST requests. If you want to allow these actions from the public user, you can change the omero.web.public.get_only property:

    $ omero config set omero.web.public.get_only false
    
  • Set the omero.web.public.url_filter. This filter is a regular expression that will allow only matching URLs to be accessed by the public user. If this is not set, no URLs will be publicly available.

    You need to configure the url_filter to allow all URLs that are required for the pages you wish to be public but to block any URLs that you do not want public users to access.

    Some examples are listed below:

    • To allow all URLs from a single app, such as ‘webgateway’, use a filter for URLs that start with the app name. For example:

      $ omero config set omero.web.public.url_filter '^/webgateway'
      

      This filter permits all URLs needed for the full image viewer. If you wish to block webgateway URLs for downloading data, use:

      $ omero config set omero.web.public.url_filter '^/webgateway/(?!archived_files|download_as)'
      
    • You may need to allow access to additional URLs for some apps. For example, the OMERO.iviewer also uses some webgateway and api URLs:

      $ omero config set omero.web.public.url_filter '^/iviewer|webgateway|api'
      
    • You can use the full webclient UI for public browsing of images. Attempts by public user to create, edit or delete data will fail silently with the default omero.web.public.get_only setting above. You may also choose to disable various dialogs for these actions such as launching scripts or OME-TIFF export, for example:

      $ omero config set omero.web.public.url_filter '^/(webadmin/myphoto/|webclient/(?!(script_ui|ome_tiff|figure_script))|webgateway/(?!(archived_files|download_as))|iviewer|api)'
      
  • Set the omero.web.public.server_id which the public user will be automatically connected to. Default: 1 (the first server in the omero.web.server_list):

    $ omero config set omero.web.public.server_id 1
    

If you enable public access to the main webclient but still wish registered users to be able to log in, the login page can always be accessed using a link of the form https://your_host/webclient/login/.

Reusing OMERO sessions

As an alternative to granting permanent public access to the data, the OMERO.web framework supports password-less, OMERO session key-based authentication. For example a direct link to image will look as follows:

https://your_host/webgateway/img_detail/IMAGE_ID/?server=SERVER_ID&bsession=OMERO_SESSION_KEY

Note

The SERVER_ID should match the index from the list set using omero.web.server_list from the server session you created. If your list contains only one server, the index will be 1.

For more details about how to create an OMERO session see server-side session or use the command line interface to create one.

Full example of hosting data for a publication

Putting the pieces of this puzzle together, the following describes the steps of a complete workflow for using OMERO to host public data associated with a publication. It is illustrated using an example publication from the Swedlow lab in Dundee, Schleicher et al, 2017 with the data hosted at https://omero.lifesci.dundee.ac.uk/pub/schleicher-et-al-2017.

Ansible playbooks can be found describing how the production server in Dundee (“nightshade”) was configured in the prod-playbooks repository on GitHub.

Group setup

A group-per-publication allows the public user to be selectively added (or removed) from given publications to decide their visibility.

  1. Create a dedicated read-only group to host the raw data underlying the publication (see User/group management).

  2. Add all the authors of the paper to this new group.

  3. Once you have configured OMERO.web to create a public user (see below), add the public user as a member of the newly created read-only group.

Configuring OMERO.web

If you wish to have an automatically logged-in public user while still giving your existing OMERO users an unchanged user experience (i.e. not automatically logging them in as the public user), a dedicated, separate web server for servicing the public workflows can be added and configured to point at your existing OMERO.server. This is the workflow adopted here by adding a public OMERO.web at https://omero.lifesci.dundee.ac.uk, without changing the existing internal OMERO.web.

  1. Follow the steps in Configuring public user above on the chosen OMERO.web.

  2. Also configure the filter on the public user on the chosen OMERO.web by setting omero.web.public.url_filter to allow ‘webclient’ so that the full webclient is visible for the public user, and thus the Data tree with Projects and Datasets is also browsable, as well as the Tags tab and the full image viewer.

Data migration

The data to be made public will need to be in the publication group to be considered “published”.

  1. Move the original images into the dedicated group using OMERO.web or OMERO.cli. The CLI is best used where Images or Datasets are cross-linked to other Datasets or Projects in the original group. The command omero chgrp Project:$ID --include Dataset,Image cuts the cross-links in the original group and preserves the Project/Dataset/Image hierarchy prepared for the move by the author.

  2. If you have used OMERO.figure to create your figures for publication, you can always find the original data by using the ‘info’ tab, as shown in the OMERO.figure Help guide (OMERO.figure supports a complete figure creation workflow, including exporting figures into image processing applications for final adjustments - see the OMERO.figure Help guide for full details).

  3. Having all the data belong to one user simplifies the UI experience for public users. If necessary, ownership of data can be transferred using the ‘Chown’ privilege (see Administrators with restricted privileges and Changing ownership of objects).

Data layout

Once the data is in the dedicated read-only group, it can be reorganized and renamed to reflect the publication e.g. Projects can be renamed according to the corresponding figure panels in the manuscript while the names of the Datasets could be retained corresponding to different treatment conditions represented in each figure panel. For example, Project Schleicher_etal_figure7_c contains images underlying the publication Figure panel 7(c). Some Projects underlie two publication figure panels, such as Project Schleicher_etal_figure2_a_c where representative images are shown in panel (a) and the corresponding quantification is shown in panel (c) of Figure 2. This makes clear which original images are underlying which figure panels in the publication.

Data can also be tagged with OMERO tags to enhance the browsing possibilities through these data for any user with basic knowledge of OMERO. For example, see Tag:Schleicher_etal_figure1_a. The tags are highlighting the images displayed in the publication figures as images. The other, non-tagged images in the group are the ones used for analysis which produced the published numerical data.

Key-Value pairs can be used to add more detailed information about the study and publication. For example, go to Schleicher_etal_figure1_a and expand the ‘Key-Value Pairs’ section in the right-hand pane to display the content (see the Managing data guide for information on using Key-Value pairs).

Configuring URLs

The URL of the first Project (corresponding to the first figure in the publication) can be used for a DOI and data landing page. For example, Project ‘Schleicher_etal_figure1_a’ https://omero.lifesci.dundee.ac.uk/webclient/?show=project-27936 corresponds to http://dx.doi.org/10.17867/10000109.

Optionally, you can decide on a set pattern of URLs for this and future publications. For example, in Dundee we have established a pattern which supposes every new publication from our institution will be in a separate group, and this group will be directly navigable by the public user using the syntax: “server-address/pub/publication-identifier”. This means for example, https://omero.lifesci.dundee.ac.uk/pub/schleicher-et-al-2017 is the link where “omero.lifesci.dundee.ac.uk” is the server address, and “schleicher-et-al-2017” is the publication-identifier.

This makes use of redirects allowing https://omero.lifesci.dundee.ac.uk/pub/schleicher-et-al-2017 to link to the correct group and Project in OMERO, just as the DOI above does. Redirects need to be set in the NGINX component of the OMERO.web installation dedicated to publication workflows. You can find our configuration for this example here on GitHub:

location /pub/schleicher-et-al-2017 {
    return 307 /webclient/?show=project-27936;
}