Andrea Aime wrote:
> Alessio Fabiani ha scritto:
>> Hi guys,
>>
>> I made few improvements to the GeoServer RESTConfig and sldService 
>> stuff which I would like to briefly discuss here, and in case of 
>> interest I can commit it:
>>
>> 1. DataStoreFileResource: this class originally was intended to let 
>> GeoServer get a shapefile and create a new datastore. I corrected 
>> some errors and changed a bit the behavior. Originally this resource 
>> was checking just for the arrival of a single .shp file, which is not 
>> sufficient for my needs; so I changed it in order to let it accept a 
>> ZIP file containing the whole Shapefile stuff, decompress it into 
>> GeoServer data dir, create the new datastore and contextually create 
>> the FeatureType also.
>
> This seems sensible to me. David?
This sounds a bit off to me, the DataStoreFileResource that was already 
there should have been accepting a ZIP file with the shapefile and its 
companions.  I will look over the code in SVN and figure out what the 
confusion is here.  (Hunch: I put the ZIP stuff in trunk after the 1.6.x 
branch of RESTconfig was created.)  Anyway, I am all for having the 
ZIPped fileset instead of just the shapefile.
>> 2. FeatureTypeResource: this class allows a user to insert/update a 
>> FeatureType. Originally the resource constrained the user to send the 
>> whole feature type properties everytime. I changed a bit in order to 
>> let a user to update also just some of the properties of the 
>> FeatureType if the latter is already present in the catalog.
>
> Hum, how do you do that? Are you exposing the properties as 
> sub-resources? In REST there is no partial update of a resource afaik,
> usually that is obtained by exposing explicitly what you want to make
> modifiable singularly.
aaime++
>> 3. StyleResource: this resource was incomplete, in particoular it 
>> didn't save the style on the catalog and also didn't check for the 
>> style existence in the case of a replacement. I changed it to perform 
>> those tasks.
>
> Sounds sensible again.
Hm, sounds like another mismatch between trunk and 1.6.x.  Again, all 
for the improvements if they're not already there.
>
>> 4. SimpleClient: I put this simple utility class in the test package 
>> in order to allow users to perform simple GET/POST/PUT/DELETE 
>> requests to the REST interfaces. It is not fully complete though.
>
> Is it any better than using curl on the command line? (did not look,
> just wondering).
Other clients I've used for quick REST requests include 
http://restclient.org/ and the firefox extension RESTTest 
(http://addons.mozilla.org/en-US/firefox/addon/5946)
>> 5. on sldService I added few Resources to let users to easily 
>> generate color classifications and optionally color Ramps, without 
>> the need to create and save a new SLD. Maybe they are redoundant, but 
>> very usefull to me. On a POST request, the user can specify the 
>> datasotre, the property, the type of color quantization to apply and 
>> the number of classes. 
>
> POST == creation of new resource. If your code is just returning
> a style without storing it on the server, the request should be a GET.
> Moreover, why specify a datastore? Wouldn't it be better to have
> a GET request on the FeatureType node that looks like
>
> .../topp:states?request=classify&attribute=myAtt&method=quantile&intervals=5&ramp=red
>  
>
>
>> The user can also send as optional parameters of the form the type of 
>> color ramp to generate and in case of custom color ramp also the 
>> start, end and mid colors. The result is a JSON string representing 
>> the quantization rules. Each rule consists basically of a title, a 
>> filter and if the color ramp was specificed a symbolizer describing 
>> the fill to apply.
>>
>> 6. I added to the sldService a UserStyleResource which allows a user 
>> to modify the default SLD of a wms layer, independetly if it is 
>> generated from a featuretype or a coverage. Not yet fully complete 
>> though.
>
> I thought that was already covered by RESTConfig?

Yes, me too.  What does this accomplish that isn't met by the existing code?
> Cheers
> Andrea
>
> !DSPAM:4040,482d45ed41092085621377!
>
-david

-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft 
Defy all challenges. Microsoft(R) Visual Studio 2008. 
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________
Geoserver-devel mailing list
Geoserver-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/geoserver-devel

Reply via email to