Data providers and other community stakeholders must currently invest operational and development resources producing search clients to serve a segment of the community. Earthdata Search portals reduce the overhead of providing a search interface similar to Earthdata Search to targeted users. It provides several useful features, including:
- Running a maintained, up-to-date instance of Earthdata Search without incurring the cost of hardware or administration
- The ability to present a limited set of collections to users, scoped to only those relevant to the target user community
- Presentation of additional logos, navigation, and page titles
- A simple URL for the portal (https://search.earthdata.nasa.gov/portal/your-portal-name) with bookmark-able links within the portal
- Interoperability with the full Earthdata Search site, allowing users to optionally and easily search the full list of collections, and see their saved projects and past retrievals from the portal in Earthdata Search
The remainder of this document will walk through the process of creating and configuring a portal.
Step 1: Reach out!
We're here to help, and we want to understand community needs so we can refine our features. Email Valerie Dixon (email@example.com) to get started with your portal and get in touch with developers.
Step 2: Fork our GitHub repository
Important: The workflow for configuring portals is geared toward teams who have familiarity with basic development tools such as git. If this is not the case for you, please reach out to us and let us know; we may be able to assist you with basic configuration.
Because portal configurations are powerful enough to modify Earthdata Search's functionality in arbitrary ways, we must handle portal configuration through our Github project (https://github.com/nasa/earthdata-search/). This allows us to understand incoming changes, test existing portals against new features, and prevent insecure or malicious code from entering our operational environment.
To provide us configuration most efficiently, you will need to fork our GitHub repository.
Step 3: Startup a Local instance of Earthdata Search
Clone the code from your fork and follow the README to get a local instance of Earthdata Search set up.
Step 4: Configure your portal
To create your portal you will need to create a new directory in "/portals" with a unique name you want as your portal identifier, for this example we will use "example". You then will create the config.json file ("/portals/example/config.json").
A significant amount of configuration can be done through this config.json file. You can view example configurations at https://github.com/nasa/earthdata-search/tree/master/portals/simple/config.json and https://github.com/nasa/earthdata-search/tree/master/portals/complex/config.json. Here is an explanation of the configuration:
|advancedSearch||Boolean||Enable/disable the advanced search feature|
|authentication||Boolean||Enable/disable all features that rely on authentication (logging in/projects)|
|featureFacets||Object||Feature Facet configuration|
|showAvailableFromAwsCloud||Boolean||Show/hide the Available from AWS Cloud feature facet|
|showCustomizable||Boolean||Show/hide the Customizable feature facet|
|showMapImagery||Boolean||Show/hide the Map Imagery feature facet|
|displayVersion||Boolean||Display the current application version in the footer|
|attributionText||String||Text to display in the footer, example "NASA Official: ..."|
Array of links to display on the left side of the footer
This array is made up of objects with two values:
|secondaryLinks||Array||Array of links to display on the right side of the footer. This uses the same objects as `primaryLinks`|
This tells EDSC if it needs to load a "styles.scss" file, see Advanced Configuration.
|Boolean||This tells EDSC if it needs to load a "scripts.js" file, see Advanced Configuration.|
|Object (Optional)||Configuration for a logo to appear to the right of the NASA logo in the upper left of the page. If omitted, no logo will be added.|
|String (Optional)||This is the id attribute of the <a> tag that houses the logo image, see Advanced Configuration.|
|String (Optional)||Path to logo image. This is not recommended, please see Advanced Configuration for the preferred method of adding the logo image.|
|String (Optional)||This is the URL that users will be sent to when clicking on the logo image. If no value is provided clicking on the logo image will return the user to the EDSC portal home (/portal/example/search).|
|String (Optional)||This text is set as the title attribute in the <a> tag that houses the logo image, it will be displayed when you user hovers their mouse over the logo|
|String (Optional)||A short string representing the organization that replaces "Earthdata" in the top-left of the site toolbar. If omitted, it will just say "Earthdata."|
|pageTitle||String||This will appear after the default portal's `org` and `title` in the HTML page title (see in the tab of the browser). For example if EDSC is the default portal then the page title will be "Earthdata Search :: <pageTitle>"|
|parentConfig||String||The parent config file for this portal. If your portal will be deployed to EOSDIS Earthdata Search this value should be "edsc". If not, set this to "default".|
These parameters are used to filter the list of collections exposed in the portal.
Current supported values:
One common pattern is to show all the datasets in a dataCenter. This example shows the ORNL DAAC portal configuration:
Likewise, for Science Project-focused portals, here is the AIRMOSS project portal:
A more nuanced approach can be applied using tags in CMR. Here is an example from the Suborbital portal using the tagKey parameter (note that the tagKey can contain simple wildcards):
If you need support of another CMR query parameter, please contact us.
Refer to the CMR documentation for information on parameters:
A short string to be used when referring to the portal. This will appear verbatim where the word "Search" is in the top-left of the site toolbar and will appear in various other places as "Example Portal" when referring to the portal. If omitted, a capitalized version of the portal ID will be used, i.e. "Example."
|showOnlyGranulesCheckbox||Boolean||Show/hide the "|
Inheriting Configuration Values
Portals now will inherit configuration values from a parent configuration. See the "parentConfig" value above. This allows you to use all of a parent's config values and keep your portal configuration minimal if your portal needs most of the same features as the parent configuration.
If you are authoring something complex or if you have needs that you feel would be better addressed at the basic configuration level, please reach out to our developers, who Chris Lynnes (firstname.lastname@example.org) can put you in touch with.
Setting up a Custom Portal Logo:
This is the preferred way to add a new logo image, which are displayed next to the NASA meatball, centered vertically. You can reference the existing portals in the ./portals/... directory for examples of how this works in practice.
titlein the logo section of the config.json
- Save your image(s) into the ../portals/myportal/images/ directory
- If in In the logo section of your config.json file, you set the "id" field to "example-logo", then you would use this SASS rule to load that image:
height: 25px; // set the correct height for your logo
width: 54px // set the correct width for your logo ;
background-image: url('./images/<portal name>-logo.[png|jpeg]');
Custom Logo Best Practices:
- Images will be displayed at a maximum of 25px tall by 57px wide. Logos should be trimmed to size as to fit will within those maximum dimensions.
- Images can be provided at twice their displayed size (50px x 114px maximum) for better display on high resolution screens.
- Images should be named using the following format: <portal name>-logo.[png|jpeg].
- If available, versions of the logo with transparent backgrounds (typically png's), meant to be displayed on dark backgrounds are preferred.
In order for this file to be executed, hasScripts must be set to true in config.json
Step 5: Incorporating Changes into Earthdata Search
When you have made and tested your changes, push them to your forked copy of the Earthdata Search repository and issue a pull request into the main repository using the pull request template. Our developers will review and incorporate the changes as soon as possible.
Our normal progression to operations is as follows:
- When a pull request is merged, it starts a build. Within a few minutes of the build becoming green, that changes gets pushed to our SIT environment for testing: https://search.sit.earthdata.nasa.gov
- Every other Wednesday, our current UAT environment (https://search.uat.earthdata.nasa.gov) is deployed to our operational environment (https://search.earthdata.nasa.gov) and our SIT environment from the previous Monday is deployed to our UAT environment.
So the progression from a pull request being accepted to it becoming an operational change takes about a month, during which it is available in testing environments.
While we would prefer to stick to this schedule wherever possible, we understand you may have differing schedule needs. If you need it to reach operations sooner, please indicate the desired date on the pull request and we will do our best to accommodate.
(1) How does a curator tag the actual data collections?
(2) Can we write out the fork and pull request procedures as a step by step recipe that a scientist could understand?
As far as I understand, CMR has not opened up tagging to users because the permissions needed would allow users to tag any collection, not just their own collections. They would also be able to edit/delete tags from other collections as well.
I think that CMR does have stories to improve the permissions granularity, but that would need to be prioritized.
I'm not thinking of end users. I am thinking of DAAC support scientists, NASA applications scientists, etc. who are on the front lines of dataset curation for EOSDIS' various communities.
Patrick Quinn please add a table of contents so I can link to Subsections. Thank you.
Should include here a policy of review and retirement for Portals that fall into disuse. For example, if usage metrics show a Portal falls below (TBD threshold of hits/year) we will reach out to the Portal's GITHub branch owner and notify them that due to disuse the Portal will be retired by (TBD time period - 3 months?). Then put a tkt in for next PI to close out the Portal. I'm sure there are sticking points and devilish details, but we don't want to become a forever repo for stale portals. And folks reading this guide should be made aware of that.