The HEASARC Astrobrowse

Astrobrowse is a utility designed for the astronomical community, which enables them to query many astronomical databases at once for target data.
The great number of sites with information about astronomical objects makes it very difficult for astronomers to find out where there might be information of interest to their research. Astrobrowse is an effort to use modern Web technology to enable astronomers (and the public) to query many sites easily. Astrobrowse explodes a user's query to sites in its database, caches the results, and presents them to the user when desired. This document gives a brief overview of the user interfaces and inner workings of the Astrobrowse system.
The astronomer fills in just one form with a target name or coordinates, then Astrobrowse sends queries to all chosen databases in parallel and collects the results.

How Astrobrowse works

Astrobrowse emulates a user accessing the Web using a Web browser. On the Web there are many forms which allow a user to select astronomical information. Many of these forms allow the user to request information around some position but they have very heterogeneous interfaces. What Astrobrowse does is provide a very basic form where a user supplies the position and size of the region he or she is interested in, and then translates that request into the syntax needed at each of the services the user requests. In the future additional services may provide access to Web sites that need information other than a position in the sky.
Use the Quick Search box, or select one of the resource menus (quick overview, full listing, or outreach list), or create your own resource menu by searching for resources.

Starting the search

Currently there are three ways to start your search: use the Quick Search box to enter a target name but not choose specific data resources, or use the lists of resources (there are currently three - quick overview, full tree, and outreach, for non-astronomers) to choose which resources to serch for your target name. The third method is to use the search page to create a list of resources based on keywords: for example, a list of resources which contain radio data for galaxies.

The Quick Search box on the left side of the main page is the fastest way to start finding data. Simply enter an object name, click at least one catagory of data resources, and click 'go'. This method requires Javascript.

The quick resources overview is a single page which provides links to many resources. Users querying this page can quickly get a sense of what information is available on the Web for the specified object and position. This form is compressed as much as possible. Thus it does not provide as much information about a given resource as the other interfaces. In particular it does not indicate the site which is being queried though this is usually clear when looking at the results.

The HEASARC Astrobrowse tree is a set of Astrobrowse pages which describe all the resources linked by the HEASARC Astrobrowse agent as a hierarchical tree. Many commonly used resources are made available directly on the first page. Where an institution has put many resources on the Web, or where the resource is unlikely to be of interest to most users, it may be found by looking at the results of a higher level Astrobrowse call. For example, a few of the most popular SkyView surveys are accessible directly from the topmost page. However if a user wishes to access one of the other SkyView surveys he or she can select the All SkyView Surveys resource. This is linked to another Astrobrowse page which describes all of the SkyView surveys. When users display this page (see below) they can select any SkyView survey.

The keyword search interface provides a mechanism by which a user can specify some characteristics of the resources desired. Users can specify keywords that are to be found in the description of the resources. The search behavior can be modified by use of 'and', 'or', 'not', parentheses for grouping terms, and the '*' wildcard for truncating keywords. The indexing and searching engine currently used is SWISH-e 1.1.

  • The bandpass is simply the frequency/energy regime covered by the data. Some resources may support multiple bandpasses.
  • The data source type indicates whether the data returned are from heterogeneous observations, a more systematic survey, or whether only highly refined data (typically catalogs) are available.
  • The data available indicates the kind of data that are available either directly in the returned pages for this resource, or possibly through links from this page.
Astrobrowse sends the user's query to all selected resources in parallel.

Exploding the request

All the resource list interfaces lead to forms where the user can specify a position (or target name that is resolved into a position), and angular size for the search. Note that some sites use a radius (cone) search while others use a box search. The size is optional and is ignored by some services. Below this section is a list of services the user can query. The user checks as many as are of interest and then clicks on one of the submit buttons. If the keyword search was used, the sites will be selected by default and the user should de-select any resources he or she does not wish to query.

When the request is submitted, Astrobrowse converts the position coordinates into J2000 coordinates. If the user has specified a target name, this is translated using the NED and SIMBAD name resolvers.

Astrobrowse then loops through each of the services the user has selected and reads a configuration file for that service. The configuration file describes the URL to be used for the query and the form name/value pairs to be sent. It also includes information about how the RA, Dec and size fields are encoded in the form. For each form, Astrobrowse creates an HTML request and sends it off. Many requests may be submitted in parallel.

The left frame is an index of the individual results. Click to see results in the right frame.

Monitoring the progress of the request

Once all requests have been initiated, Astrobrowse returns a framed page to the user. On the left is a status bar giving the status of each request. On the right is a panel where the user can display the results of each query. At any time the user can click on the Update Status button to see the current status of each request.
Switching back and forth between results is fast since the results are cached at the HEASARC.

Viewing results

The user can click on the hyperlinks to the results to see data from a given page. Usually a user will wait until the request is completed, but for requests which take some time it may be possible to see intermediate results before the request is finished.

Each request is handled by a separate process. If too many requests have been submitted some requests may be queued until other requests have finished. The process queries the desired URL and filters the results so that any relative URLs are converted into absolute URLs. The results are then cached at the HEASARC. Thus a user can go back and forth amongst the results without reissuing a request. However, if the user clicks on an anchor in one of the results, the user will then typically move into the environment of the remote site. Cached results are kept for a reasonable period and then deleted.


Recursive requests

One of the kinds of pages that Astrobrowse may reference is another Astrobrowse page. When such a page is displayed the user can select one or more services from the page and click on the Start Search button.
=> show result in the full browser window (no frame).
=> delete the result from the status bar.

Managing the status bar

Each request on the status bar is followed by two small icons. One of these is a link to allow the user to display the returned results for that request in the top browser window. The user can get the status bar back by using the browsers Back button, but should note that when using Frames one does not always get back to the last state of a framed window.

The second icon deletes the request from the status bar. This doesn't delete the request. Using the browser Back button a user may be able to get back a status frame which will provide a link to the request. However, this request will not appear on any status bars generated after the deletion.

GLU is a client/server system which sends changes made at one site to all other registered sites, ensuring that everyone on the network has the same copy of resource descriptions.

Astrobrowse and GLU

The configuration data that Astrobrowse uses to describe each site is stored in a format compatible with the GLU (Générateur de Liens Uniformes) system developed at the Centre Donnees Astronomiques de Strasbourg (CDS). GLU provides a mechanism by which sites distributed over the Web can easily transmit information about changes to the Web pages supported at the site. E.g., suppose the syntax of a Web form used to query a database changes. If that site uses GLU to describe that Web page, the site can update their own GLU entry, and this change will automatically propagate to all other GLU sites which have registered an interest in that page.

In the context of Astrobrowse this can be extremely valuable, since it will permit local sites to change their own Web pages without breaking Astrobrowse links. We encourage other sites to bring up GLU engines to manage (and possibly add to) the Astrobrowse information for their site.

For more information on GLU, please see the GLU site or contact Pierre Fernique. For more information about how GLU and Astrobrowse interact, contact astrobrowse@athena.gsfc.nasa.gov.


Historical notes

The HEASARC Astrobrowse effort was inspired by a project initiated by Bob Hanisch (ST ScI) and Steve Murray (SAO). The current implementation was originally intended to provide a mechanism by which users could query high-energy archive sites, however it quickly became evident that it could be used in a wider context.

Further information

The HEASARC Astrobrowse team members are: For further information about Astrobrowse please contact us at astrobrowse@athena.gsfc.nasa.gov.