DSpace REST API
This is a wiki page for DSpace REST API addon. The project is in development phase, but can be tested by the users. For the details please check this page.
Details
Project Title: |
DSpace REST API |
Author: |
Bojan Suzic |
Mentors (at GSoC): |
Aaron Zeckoski, Mark Diggory |
Contacting author: |
bojan.suzic AT gmail _DOT _com - using subject line DSpace REST |
SCM Location for Project: |
Project description
The REST approach promotes simplification and decoupling of software architecture, enabling further scalability, portability, granularity and simplified interaction of software systems and components.
The aim of this project is to provide DSpace with REST capable API and underlying component, which will enable developers and end-users to exploit the advantages of such approach.
Some of uses this module is intended to provide could be, for instance:
- interaction between DSpace systems and/or other repositories
- automation of different activities, e.g. submission of packages
- integrating repositories in process workflows of other applications or systems
- interaction with many kinds of systems or web applications, such as CMS, LMS, LCMS, VLS, AMS etc
- providing of other approaches to UI, such as client based/run UI
- crawling of repositories, exposing information in structural way
This project is continuation of GSoC 2009 and GSoC 2010 . In the first stage the basic support for REST for DSpace is provided, exposing many parts of DSpace functionality to the clients. Currently the module is being tested and improved in preparation for upcomign DSpace release.
Based on REST API for DSpace the new GSoC (2011) project is started. Vibhaj Rajan DSpace UI Client based on JavaScript and appropriate framework (JQuery). More information: Client UI REST API
Release plan and issues
It is expected to have working and tested code for DSpace 1.8 release. According to relevant discussions on DSpace developer meetings, there is possibility to have this code released asynchronously of DSpace, as independent module. After initial release the code will be actively maintained.
Currently (mid October 2010) there are several issues opened, of them I would notice the following:
- Performance issues during browsing bigger datasets. This is generally related to DSpace API. I have reported this problem and proposed solution at http://jira.dspace.org/jira/browse/DS-659. However as I am not sure whether it will be accepted and included in the upcoming release I will change the code and translate some functions used from DSpace API to DSpace REST API. This way some handling will be done directly at REST API level. Consequently some additional features related to sorting/ordering will be available.
- Multiple loops in listing Collections and some other entities. This issue is resolved as of end September. Additionally, option to fine-grain details level of the output is implemented (three levels).
- HTTP Basic Auth - this is easily to implement and will be done shortly
- Some Authorization related issues: the authorization handling is done on the level of DSpace API. However some functions translated to REST API are not directly available to users and thus do not provide Authorization mechanisms. In order to prevent misuse etc. this gap should be filled at REST API level.
- Not finished end-points. Working on them.
- Testing. Testing. I also need cooperation of you potential users. I need repository for testing containing at least several hundreds of items. If you can provide me with that please contact me via email.
- DSpace 1.5 and older versions support - planned to be implemented at the end of initial public release.
Recent changes
- June 10: XML input/digestion change, based on the input from Dhaivat Dave
- June 6: Added support for logo in communities and collections (issue reported by Vibhaj Rayan and Hayden Young)
- June 1: Applied Peter Dietz's patches (https://gist.github.com/952058) and pom.xml improvements discussed bellow on this page; fixed collection/items display reported by Hayden Young
Detailed activities
In the following sections main activities are elaborated in detail.
REST API Endpoints
In the following section listed are supported endpoints on the application level. The items marked with dot (in C column) are in phase of implementation, while other items are considered already working.
Please note that additional tests should be made in order to ensure proper stability of the whole application.
The sorting of the fields / output results is currently partially supported. This part of the application is implemented independently of the endpoints and will be worked on after the most of endpoints are completed.
Naming convention for endpoints
DSpace 1.x and 2.x are treating the resources on different way. 2.x is more generalized, suggesting the use of RDF-like interrelation notations.
Repository browsing
Earlier Implementation Description - GSoC09
C |
Verb |
URL |
Description |
Mandatory parameters |
Optional parameters |
Sorting fields |
Response Data |
Formats |
Response codes |
---|---|---|---|---|---|---|---|---|---|
|
|
|
Returns a list of all communities on the system or return just top level communities. |
- |
The list of communities containing respective fields . |
|
|||
|
|
|
Return detailed information about |
|
|
- |
Fields describing community. |
|
|
|
|
|
Return a particular data field found in the community |
|
Respective field info |
|
|||
|
|
|
Return a list of all collections in the system. |
- |
The list of the collections containing respective fields. |
|
|||
|
|
|
Return detailed information about |
|
Fields of the collection entity. |
|
|||
|
|
|
Return a particular data field found in the collection |
|
|
Respective field info |
|
||
|
|
|
Return a list of the items in the system |
- |
- |
The list of the items containing related fields . |
|
|
|
|
|
|
Return detailed information about an item. |
|
|
Fields of the item entity. |
|
200, 204, 400, 500 |
|
|
|
|
Return a particular data field fould in the item |
|
- |
Respective field info |
|
200, 204, 400, 500 |
|
|
|
|
Return bitstream object - usually the library item file. |
|
- |
- |
Fields of the bitstream entity. |
|
200, 400, 401, 403, 404, 500 |
|
|
|
Return a particular data field found in bitstream |
|
- |
Respective field info |
|
200, 400, 401, 403, 404, 500 |
|
|
|
|
Return bitstream |
|
- |
- |
Return bitstream |
|
200, 400, 401, 403, 404, 500 |
|
|
|
Return a list of the groups in the system |
- |
- |
The list of the groups containing related fields . |
|
200, 204, 400, 500 |
|
|
|
|
Return a group object |
|
- |
Fields of the group entity. |
|
200, 204, 400, 500 |
|
|
|
|
Return a particular data field found in the group entity |
|
- |
Respective field info |
|
200, 204, 400, 500 |
|
|
|
|
Return a list of the users in the system |
- |
- |
The list of the users containing related fields . |
|
200,204,400,500 |
|
|
|
|
Return a user info |
|
- |
Fields of the user entity. |
|
200,204,400,500 |
|
|
|
|
Return a particular data field found in the user |
|
- |
Respective field info |
|
200,204,400,500 |
Note: modifier idOnly
is referred only to first layer of the results. For all other layers (e.g. nested results) only ids are returned in some cases, due to possible loops. Example: for community containing collections, on second level the response contains only ids for some elements where multiple loops may be created (community->has_collection->has_community....). Other data is modified according to idOnly
flag.
Optional parameters
Parameter |
Description |
---|---|
|
returns only top level communities |
|
if true return only the identifiers for the record |
|
return only direct parent community |
|
return only collections user has permission to work on |
|
return archived items for respective collection |
|
parameters: |
Sorting fields:
Not completed!
The sorting of the fields / output results is currently partially supported. This part of the application is implemented independently of the endpoints and will be worked on after the most of endpoints are completed.
Parameter |
Description |
Ordering supported |
---|---|---|
|
sort results by entity id |
|
|
sort results by entity name |
|
|
sort results by number of items contained |
|
|
sort results by date of last item modification |
|
|
sort results by submitter name |
|
|
sort results by submitter id |
|
Controlling results
Parameter |
Description |
Default |
Example |
---|---|---|---|
|
position of the first entity to return |
0 (first) |
|
|
page of data to display |
0 (first) |
|
|
number of results to show on each page |
0 (all) |
|
|
maximum number of entities to return |
0 (all) |
|
|
|
the sort order to return entities in |
|
Response codes
Code |
Description |
---|---|
200 |
OK |
201 |
Created |
204 |
No content |
400 |
Bad request |
401 |
Unauthorized |
403 |
Forbidden |
404 |
Not found |
405 |
Method not allowed |
500 |
Internal server error |
503 |
Service unavailable |
Authentication/Authorization
Currently only standard authentication is supported. The authentication data is provided in the request or header body.
Example:
/rest/communities.json?user=user@email.com&pass=userpassword
The same elements user
and pass
are used for header based authentication.
Authorization is done on underlying api level; in the case of error the proper message and error code are returned to the user.
Repository manipulation
C |
Verb |
URL |
Description |
Mandatory parameters |
Optional parameters |
Response Data |
Formats |
Response codes |
---|---|---|---|---|---|---|---|---|
|
|
|
Action to be done under community |
|
- |
|
|
200, 400, 401, 403, 500 |
|
|
/communities/{id}/{element} |
Update the field |
|
- |
Response code |
|
200, 400, 401, 403, 500 |
|
|
|
Set the logo for community |
|
- |
Response code |
binary |
200, 400, 401, 403, 500 |
|
|
|
Delete community from the system |
|
- |
Response code |
|
200, 400, 401, 403, 500 |
|
|
|
Remove attribute/value |
|
- |
Response code |
|
|
|
|
|
Action to be done under collection |
|
- |
Id ow newly created element |
|
200, 400, 401, 403, 500 |
|
|
|
Update field |
|
- |
Response code |
|
200, 400, 401, 403, 500 |
|
|
|
Delete collection from the system - CANNOT BE DONE DIRECTLY |
- |
- |
Response code |
|
200, 400, 401, 403, 500 |
|
|
|
Remove attribute/value |
|
|
Response code |
|
200,400,401,403,500 |
• |
|
|
Set the logo for collection |
|
|
Response code |
binary |
200,400,401,403,500 |
• |
|
|
Action to be done under item |
|
|
Id of newly created element |
|
200,400,401,403,500 |
• |
|
|
Update field |
|
- |
Response code |
|
200,400,401,403,500 |
• |
|
|
Delete item from the system |
|
- |
Response code |
|
200,400,401,403,500 |
• |
|
|
Delete element/attribute |
|
- |
Response code |
|
200,400,401,403,500 |
Content searching
C |
Verb |
URL |
Description |
Mandatory parameters |
Optional parameters |
Sorting fields |
Response Data |
Formats |
Response codes |
---|---|---|---|---|---|---|---|---|---|
|
|
|
Return a list of all objects found by searching criteria. |
- |
modifiers{{query= |
|
Item info with basic metadata for the search results. Additionally return only |
|
200, 204, 400, 500 |
|
|
|
Return a list of all objects that have been created, modified or withdrawn within specified time range. |
- |
|
- |
Contains item info including id, name, handle, metadata, bitstreams according to |
|
200, 204, 400, 500 |
Statistical info
C |
Verb |
URL |
Description |
Mandatory parameters |
Optional parameters |
Sorting fields |
Response Data |
Formats |
Response codes |
---|---|---|---|---|---|---|---|---|---|
|
|
|
Return general statistics. |
- |
- |
- |
Cummulative list of statistics data for the system currently available. |
|
200, 400, 500 |
Relationships interface
Experimental feature
This is considered as a experimental feature in the phase of being considered for compability with future versions of DSpace. Consider not important section; the status of the feature for upcoming release yet to be determined.
C |
Verb |
URL |
Description |
Mandatory parameters |
Optional parameters |
Sorting fields |
Response Data |
Formats |
Response codes |
---|---|---|---|---|---|---|---|---|---|
• |
GET |
/resource/{handle}/relations |
Return entities according to relation and parameters specified |
|
|
- |
ontains entities selected and sorted in conformance to request parameters. For more details see description of |
|
Mandatory parameters
Parameter |
Description |
Values |
Example |
---|---|---|---|
|
Return entities satisfying requested property relation |
Structural properties |
|
|
restriction on type - only entity with specifed type(s) would be returned |
|
|
rfield |
restriction on fields - return only selected fields; by default all fields are returned |
id |
rfield=id,name - return only entity id and name in response |
Note: incomplete/orientative properties, for more info check [Vocabularies|http://code.google.com/p/dspace-sandbox/source/browse/#svn/modules/dspace-rdf/tags/dspace-rdf-1.5.1/src/main/java/org/dspace/adapters/rdf/vocabularies].
Visitors' suggestions or wishes
Here the visitors and stakeholders can insert their suggestions or describe the needs for their applications in detail.
Comment: In this case it is not clear how to treat recent
part of endpoint. If we stick to semantic mapping, then it should look like /resource/id/mapping
, but recent
in this case obviously do not represent a mapping, but the property.
Comment #2: Semantic mapping presented in this case should be probably hardcoded for 1.x branch, but on abstraction level which enables easy replacement with some auto-discovery method prepared for 2.x and eventually backported to 1.x. This way we would be able to call something similar to /communities/id
or communities/id/capabilities
in order to get supported mappings (amongst other data).
Suggesting new options:
Instead of changing wiki contents visitors can enter their suggestions as a comments.
1) Kevin S. Clarke and Tim Donhue suggested adding of the new feature related to HTTP Basic Auth. Ok, I will investigate how it could be done and included here. More info comming.
Integration in the system
It is planned to consult two external subjects for cooperation and the assistance during integration process (LMS and national library internal automation process). More information coming soon - awaiting approval of other parties.
Documentation tasks
Although provided software module exposes basic documentation automatically to the end user, in order to make it easier for other developers and users the documentation in the following forms is additionaly to be provided:
- Confluence pages, current location
- integrated documentation in PDF form (manual)
- short slides containing technology overview, advocacy/facts, configuration and usage guideliens and examples
- code will be additionally commented
Example of usage
At the end of the current stage of this project as a bonus task (if time constraints allow) the examples of usage will be provided for several languages, the use-cases will be presented (example of integration in other software, e.g. LMS) and optionally simple client system demonstrating UI customization will be demonstrated (e.g. Flex or JavaFX like).
How to install and test the module?
Here I will show two ways to install and test this module.
Update: Current version includes this update (thanks Peter!). This explanation will be shortly removed.
Note: The code for the REST API from the Google Summer of Code 2010 may be out of date with the latest version of dspace. It may help to import the latest stable version of DSpace through the REST-API pom.
To do so, modify [dspace-rest-api-source]/pom.xml
Index: pom.xml =================================================================== --- pom.xml (revision 6356) +++ pom.xml (working copy) @@ -18,7 +18,7 @@ <parent> <artifactId>dspace-parent</artifactId> <groupId>org.dspace</groupId> - <version>1.6.0-SNAPSHOT</version><!--dspace2.version--> + <version>1.7.1</version><!--dspace2.version--> </parent> <repositories> @@ -100,7 +100,7 @@ <groupId>org.dspace</groupId> <artifactId>dspace-api</artifactId> <scope>compile</scope> - <version>1.6.2-SNAPSHOT</version> + <version>1.7.1</version> </dependency> <dependency> @@ -121,7 +121,7 @@ <groupId>org.dspace</groupId> <artifactId>dspace-jspui-api</artifactId> <scope>compile</scope> - <version>1.6.2-SNAPSHOT</version> + <version>1.7.1</version> <exclusions> <exclusion> <groupId>org.dspace</groupId> @@ -133,7 +133,7 @@ <groupId>org.dspace</groupId> <artifactId>dspace-api-lang</artifactId> <scope>compile</scope> - <version>1.5.2.2-SNAPSHOT</version> + <version>1.7.1.0</version> </dependency> <!--
For both approaches you should have Apache Maven installed. Then proceed using Subversion and check out the code from http://scm.dspace.org/svn/repo/modules/dspace-rest/trunk
1) This way assumes you are running DSpace under Tomcat. Locate src/main/webapp/WEB-INF/web.xml
(under directory you just downloaded DSpace REST API). Find variable named dspace-config
and alter it to point to current location of dspace.cfg
file of your DSpace instance. Position into the root directory of the REST API and type maven package
. If everything goes well, in target
directory will be packaged dspace.war
. Use this file and deploy it to your current Tomcat instance. The application will be available under http://localhost:8080/rest/ by default.
2) You can run REST API under Jetty container. Proceed with the same steps as under #1. Then instead to deploy .war file to Tomcat web server, from the root of REST API source tree issue command mvn jetty:run-war
. This will run REST support under Jetty and the web point will be available at http://localhost:8080/dspace-rest/ by default.
(Option 3) Install as a DSpace Module
If you have an existing instance of DSpace that you are developing, you can connect the rest api module to your existing code base by adding it as a module.
Modify [dspace-source]/dspace/pom.xml by adding the path to the checked out rest code.
--- a/dspace/pom.xml +++ b/dspace/pom.xml @@ -505,6 +505,7 @@ --> <modules> <module>modules</module> + <module>../../../dspace-rest-gsoc10</module> </modules> <build>
Once you rebuild your dspace-src code with mvn package and ant update, you will additionally need to copy the compiled .war file produced in the dspace-rest-gsoc10 target directory to tomcat's webapps directory.
cp /path/to/dspace-rest-gsoc10/target/dspace.war /var/lib/tomcat6/webapps/rest.war
Afterwards you can restart tomcat and visit the rest api in action at: http://localhost:8080/rest
(Option 3a) Install as a DSpace Module in a source code installation
Its unlikely you will want to do this unless you are a committer or just nosey like me and like to play around with the code.
1) Create a new directory for the REST module source code - dspace-src/dspace-rest.
2) Checkout the source code from http://scm.dspace.org/svn/repo/modules/dspace-rest/trunk/ into the new directory.
3) Incorporate the new module into your project by adding a new <module> element for dspace-rest to the 'all' profile in dspace-src/pom.xml.
4) Tell Maven to use your new local module by adding a new <profile> to dspace-src/dspace/pom.xml. If you don't do this the project will build okay but won't be using your local source code for that module.
5) Create a new directory dspace-src/dspace/modules/rest.
5a) Add a sub-directory src/main/webapp and a pom.xml to the directory created in 5. (Copy the pom from any other modules/xxxx module).
6) Add a <profile> to dspace-src/dspace/modules/pom.xml.
7) Rebuild your project.
Possible problems:
- If you have troubles starting the application, try to check dspace-config
variable and make it to point to location dspace.cfg
using absolute addressing (take a look at commented line in configuration file).
- If you already run some other application on port 8080 try instead to start Jetty container with the following line: mvn jetty:run-war -Djetty.port=9090
for port 9090.
Please take a note that this is still experimental module so there may be bugs/errors in processing. You use it at your own risk.
I would highly appreciate user input. If you have comments or feature requests or anything else you can post it on this wiki in comments section. Additionally for the bugs/errors/issues found you can use JIRA at http://jira.dspace.org/jira/browse/GSOC?report=select to report or contact me directly via emal.
Information for developers
In this section the main sections of the software will be briefly explained in order to ease update or extension of the components.
The REST API for DSpace uses Aaron Zeckoski's EntityBus and DSpace standard libraries, as dspace-api.
There are two main packages: org.dspace.rest.providers and org.dspace.rest.entities. Providers are responsible for serving content/feeds to the users. They usually prepare entities or particular entity and/or handle update/delete/create functions. The main class there is AbstractBaseProvider, which is extended by other providers.
In the providers, at the constructor level created are mappings between particular endpoints (e.g. /rest/items/1/collections) and related functions in entities (org.dspace.entities.items.getCollections). Thus, for each GET, PUT, POST or DELETE function in the entity provider's constructor defined are such mappings between URL endpoints and functions. After the client makes specific call, the provider prepares answer and in this phase calls mapped function and prepare results for display.
So, if you want to extend currently available endpoints for already present providers and entities, it is necessary to define mapping at provider level and prepare corresponding function at the entity level (based on the template). The system then calls this function and provides necessary arguments for its successfully handling.
If you want to develop a new provider, it is usually necessary to create new provider class in org.dspace.rest.providers and then create related entity in org.dspace.rest.entities. The currently available providers are good example how to do that.