time series problem 3 (Bug!?)

Jeroen Ticheler Jeroen.Ticheler at FAO.ORG
Mon Oct 24 16:23:59 EDT 2005 

Hi!
I was wondering if anything was decided on the WMS/WCS time series  
support options?
Thanks in advance,
Jeroen

On 15 Oct 2005, at 21:53, Jeroen Ticheler wrote:

> Thanks! At least I now know I didn't just do things wrong and  
> understand why some things worked and others not. From the website  
> this is not at all clear as it gives the impression time series as  
> in point 2 can be served...
>
> It would be really great if option 2 is supported. Very much in  
> line with what Jeff wrote, I foresee problems with our map client  
> as well as with others when I start specifying the time series  
> following option 4, although it would be an elegant way of  
> describing the time series :-) Option 2 will be excellent for many  
> of our temporal data series.
>
> I'll be happy to test, but have to add that I'm traveling in the  
> coming two weeks, so its hard to reply directly.
> Looking forward to a solution!
>
> On one more note: I noted the way time series in WMS and WCS are  
> defined is different (according to the samples I used). WCS uses an  
> index file that is added as a separate LAYER, while WMS can  
> directly use the index shape file. I'm not sure (yet) a WCS time  
> series can be defined in the same way a WMS can be, but it might be  
> once solution 2 works for WMS!? Otherwise, would there be a way to  
> hide a LAYER completely from a WMS so one map file can be used for  
> both WCS and WMS?
> If my description is not clear, I can post a more detailed  
> explanation later when back in the office.
>
> Thanks very much in advance!
> Jeroen
>
> On 14 Oct 2005, at 04:43, Yewondwossen Assefa wrote:
>
>
>
>>
>>
>> Luis W. Sevilla wrote:
>>
>>
>>
>>> Hi,
>>> Yewondwossen Assefa wrote:
>>>
>>>
>>>
>>>> From the specs (wms 1.1 annex c c.3), the extent could be  
>>>> defined in 4 ways :
>>>>    1) value : A single value.  This is not directly supported in  
>>>> Mapserver but there is an easy workwound by specifying the same  
>>>> value as min and max
>>>>
>>>>    2) value1,value2,value3,... a A list of multiple values. This  
>>>> is not supported in Mapserver.
>>>>
>>>>    3) min/max/resolution An interval defined by its lower and upper
>>>>       bounds and its resolution.  This is supported in Mapserver  
>>>> (Note that the resolution is not supported)
>>>>
>>>>
>>>>
>>> It sounds quite strange: how do you specify the ammount of time- 
>>> slices, without giving the resolution?
>>>
>>>
>>>
>>
>>  You can specify the resolutions in your setting of the  
>> wms_timeextent parameter. It should be dumped as is in the  
>> capabilities document.  It just won't be used when validating the  
>> time value passed in the request to see if the value fits in the  
>> time extents defined. The validation will only use the min max  
>> values.
>>
>>
>>
>>
>>
>>>>
>>>>    4) min1/max1/res1,min2/max2/res2,... a A list of multiple  
>>>> intervals. Not supported
>>>>
>>>>
>>>>  I am willing to add the support for number 4) if you are  
>>>> willing to do additional testing.
>>>>
>>>>
>>>>
>>> We may do some testing too. we're on the way of developping a new  
>>> WMS cliente with time parameter support, and also interested in  
>>> publising multitemporal data on the same way.
>>> '
>>>
>>>
>>>
>>
>> I have entred a bug on this issue so discussions/testing can be  
>> done through this bug.
>>
>>   http://mapserver.gis.umn.edu/bugs/show_bug.cgi?id=1498
>>
>>
>>
>>
>>>>
>>>>  I am not sure if 2) is is a must have at this point.
>>>>
>>>>  Le me know what you think and then we can proceed with the  
>>>> impentation.
>>>>
>>>> Later,
>>>>
>>>>
>>>> Jeroen Ticheler wrote:
>>>>
>>>>
>>>>
>>>>
>>>>> Thanks for that! Is it a complex thing to add such support? I  
>>>>> had the  impression from the code that it was virtually there  
>>>>> already, just  that it first checked for a range and does not  
>>>>> proceed to check for  ranges?
>>>>> The problem is that a range doesn't work for this series. Data  
>>>>> comes  in at three 10 daily periods a month, which is not  
>>>>> exactly true as a  month does not always have 30 days. Still  
>>>>> the images are always dated  with a 01, 11 and 21 for the day  
>>>>> (or 1,2 or 3 for a dekad).
>>>>>
>>>>> Ciao,
>>>>> Jeroen
>>>>> _______________________
>>>>> Jeroen Ticheler
>>>>> FAO-UN
>>>>> Tel: +39 06 57056041
>>>>> http://www.fao.org/geonetwork
>>>>> 42.07436°N 12.34327°E
>>>>>
>>>>>
>>>>> On 13 Oct 2005, at 20:03, Yewondwossen Assefa wrote:
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>> Hi There,
>>>>>>
>>>>>>  Sorry for the late answer.
>>>>>>
>>>>>>  From what I can see in your examples, you are running into a   
>>>>>> limitation of mapserver and it's support for the time extent   
>>>>>> values. Currently the time extent can only be defined as a  
>>>>>> range  value and only one range value is valid (so something  
>>>>>> like is the  only thing valid 2005-09-11/2005-09-21)
>>>>>>
>>>>>>  You should update your "wms_timeextent"  and try it with only  
>>>>>> one  range. I will enter documentation bug so that this  
>>>>>> limitation is  clearly defined. Give it a try and let me know.
>>>>>>
>>>>>> Later,
>>>>>>
>>>>>> Jeroen Ticheler wrote:
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>>> Hi, sorry for continuing this story, but I really want to  
>>>>>>> solve  this  and hope my additional info helps to sort out  
>>>>>>> the problems I  face.
>>>>>>> I am using a shapefile with the imgdate column added to it. I  
>>>>>>> use   exactly the same index to work with a WCS on the same  
>>>>>>> data, that  one  works perfectly fine.
>>>>>>> I have browsed through the mapServer sourcecode to see if I  
>>>>>>> could   discover something/ understand the problem better.  
>>>>>>> I'm not  familiar  with C coding, so forgive my ignorance :-)  
>>>>>>> What I could  understand is  that my request fails when the  
>>>>>>> time string  submitted in the request  is compared to the  
>>>>>>> timeextent given in  the capabilities document.
>>>>>>> The exact section of the code is in  between line 545 and  
>>>>>>> line  569,  apparently the msValidateTimeValue() procedure  
>>>>>>> fails to find  the  requested date in the extent. I do not  
>>>>>>> get an error when I  change my  extent to a range like   
>>>>>>> "2005-09-11,2005-09-21/2005-09-21" or    
>>>>>>> "2005-09-11,2005-09-11/2005-09-21", but it will only render  
>>>>>>> the  very  last image, even when an earlier one is requested.
>>>>>>> So I guess I face a bug!?
>>>>>>> Thanks again,
>>>>>>> Jeroen
>>>>>>> Begin forwarded message:
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>> From: Jeroen Ticheler <Jeroen.Ticheler at FAO.ORG>
>>>>>>>> Date: 12 October 2005 09:54:26 GMT+02:00
>>>>>>>> To: MAPSERVER-USERS at LISTS.UMN.EDU
>>>>>>>> Subject: [UMN_MAPSERVER-USERS] time series problem 2
>>>>>>>> Reply-To: Jeroen Ticheler <Jeroen.Ticheler at FAO.ORG>
>>>>>>>>
>>>>>>>>
>>>>>>>> Further to my previous message, the exception I get is the    
>>>>>>>> following one:
>>>>>>>>
>>>>>>>> <ServiceExceptionReport version="1.1.1">
>>>>>>>>     <ServiceException code="InvalidDimensionValue">
>>>>>>>> msWMSApplyTime: WMS server error. Time value(s) 2005-09-21  
>>>>>>>> given  is  invalid or outside the time extent defined    
>>>>>>>> (2005-03-11,2005-03-21,2005-04-01,2005-04-11,2005-04-21,2005-05 
>>>>>>>> -01,2 00   
>>>>>>>> 5-05-11,2005-05-21,2005-06-01,2005-06-11,2005-06-21,2005-07-01, 
>>>>>>>> 2005- 07   
>>>>>>>> -11,2005-07-21,2005-08-01,2005-08-11,2005-08-21,2005-09-01,2005 
>>>>>>>> -09-1 1, 2005-09-21), and default time set is invalid  
>>>>>>>> (2005-09-21)
>>>>>>>>     </ServiceException>
>>>>>>>> </ServiceExceptionReport>
>>>>>>>>
>>>>>>>> And the URL's I use are:
>>>>>>>>
>>>>>>>> For Capabilities:
>>>>>>>>
>>>>>>>> http://193.43.36.137/cgi-bin/spotveg?   
>>>>>>>> version=1.1.1&service=WMS&request=GetCapabilities
>>>>>>>>
>>>>>>>> For a working map without specifying the time:
>>>>>>>>
>>>>>>>> http://193.43.36.137/cgi-bin/spotveg?   
>>>>>>>> version=1.1.1&service=WMS&request=GetMap&srs=EPSG:   
>>>>>>>> 4326&bbox=-180,-90,180,90&format=image/   
>>>>>>>> png&layers=spotndvi&transparent=true&width=400&height=200
>>>>>>>>
>>>>>>>> For a request with a time, the exception is returned:
>>>>>>>>
>>>>>>>> http://193.43.36.137/cgi-bin/spotveg?   
>>>>>>>> version=1.1.1&service=WMS&request=GetMap&srs=EPSG:   
>>>>>>>> 4326&bbox=-180,-90,180,90&format=image/   
>>>>>>>> png&layers=spotndvi&transparent=true&width=400&height=200&time= 
>>>>>>>> 2005- 09 -21
>>>>>>>>
>>>>>>>> These URL's can be reached from outside, so you can try  
>>>>>>>> this  yourself.
>>>>>>>> Thanks in advance for any help to a solution,
>>>>>>>> Jeroen
>>>>>>>>
>>>>>>>> On 11 Oct 2005, at 19:20, Jeroen Ticheler wrote:
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>>> Hi,
>>>>>>>>> I have a problem with time series for WMS. I have used  
>>>>>>>>> the   following WMS attributes in my map file and I get the  
>>>>>>>>> default   image the first WMS map request. The moment I  
>>>>>>>>> start requesting   maps with the time variable set, I get  
>>>>>>>>> an exception indicating   that the time format used is not  
>>>>>>>>> correct.
>>>>>>>>>
>>>>>>>>> The data is GeoTIFF
>>>>>>>>> I have an index shapefile with an imgdate column containing  
>>>>>>>>> the   dates in the same format
>>>>>>>>>
>>>>>>>>> Anyone an idea, I can't find an example that has a whole  
>>>>>>>>> range  of  images with a specific time set for each image.
>>>>>>>>>
>>>>>>>>> Map file extract:
>>>>>>>>>
>>>>>>>>> At the map level (although I don't think this is supported/  
>>>>>>>>> needed  for GeoTIFF data!?)
>>>>>>>>> ____________________________
>>>>>>>>>       "wms_timeformat"                 "YYYY-MM-DD"
>>>>>>>>> ____________________________
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> At the layer level
>>>>>>>>> ____________________________
>>>>>>>>>
>>>>>>>>>      "wms_style_default_title"              "default"
>>>>>>>>>      "wms_format"                           "image/png"
>>>>>>>>>      "wms_timedefault"                  "2005-09-11"
>>>>>>>>>      "wms_timeextent"                     
>>>>>>>>> "2005-08-21,2005-09-01,2005-09-11"          # also tried  
>>>>>>>>> with   spaces after each comma
>>>>>>>>>      "wms_timeitem"                         "imgdate"
>>>>>>>>>
>>>>>>>>>     END
>>>>>>>>>     DUMP TRUE
>>>>>>>>>     TILEINDEX 'ndvi_idx'
>>>>>>>>>     TILEITEM "location"
>>>>>>>>>     EXTENT -180.0 -90.0 180.0 90.0
>>>>>>>>>     PROJECTION  "init=epsg:4326" END
>>>>>>>>> ____________________________
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> Thanks in advance,
>>>>>>>>> Jeroen
>>>>>>>>> _______________________
>>>>>>>>> Jeroen Ticheler
>>>>>>>>> FAO-UN
>>>>>>>>> Tel: +39 06 57056041
>>>>>>>>> http://www.fao.org/geonetwork
>>>>>>>>> 12.34327°N 12.34327°E
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>> ----------------------------------------------------------------
>>>>>> Assefa Yewondwossen
>>>>>> Software Analyst
>>>>>>
>>>>>> Email: assefa at dmsolutions.ca
>>>>>> http://www.dmsolutions.ca/
>>>>>>
>>>>>> Phone: (613) 565-5056 (ext 14)
>>>>>> Fax:   (613) 565-0925
>>>>>> ----------------------------------------------------------------
>>>>>>
>>>>>>
>>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>> gvSIG development Team
>>> www.gvsig.gva.es
>>>
>>>
>>>
>>
>> -- 
>> ----------------------------------------------------------------
>> Assefa Yewondwossen
>> Software Analyst
>>
>> Email: assefa at dmsolutions.ca
>> http://www.dmsolutions.ca/
>>
>> Phone: (613) 565-5056 (ext 14)
>> Fax:   (613) 565-0925
>> ----------------------------------------------------------------
>>
>>
>>
>

More information about the mapserver-users mailing list