[Zoo-discuss] Handling reference input files for processing Zoo-kernel

Fri May 6 03:01:44 PDT 2016

Dear Gerald,
I think I will go for the first option treating input file URL as string instead of file reference. This should work as I don't need to change the legacy processor code which downloads the file as and when needed instead of downloading all to start processing.Your answer is clear and definitely helpful.   
Thanks

> Subject: Re: [Zoo-discuss] Handling reference input files for processing Zoo-kernel
> From: gerald.fenoy at geolabs.fr
> Date: Fri, 6 May 2016 11:53:04 +0200
> CC: zoo-discuss at lists.osgeo.org
> To: rakeshuk at outlook.com
> 
> Dear Rakesh,
> sorry for late reply.
> 
> By now, there is no existing attribute for input to specify that you don’t want the server to download the data for an input given as reference, so no way to specify that the ZOO-Kernel should not download a reference. Nevertheless, in you specific use case, I suppose that you may define your inputs in a different way. Indeed, if you need to have access to the reference itself rather than the downloaded value, the easiest way to do so is to use input string rather than a reference to a file to download. This way, your service will have access to the reference url as the value of the corresponding input, then the service code can handle the download on its own.
> 
> Obviously if it is required for your use case (because, for example, you already have a client interface which is forcing you to use real input reference), then we can imagine a specific attribute to set in the ZCFG corresponding to your service specifying that a reference should not be downloaded by the ZOO-Kernel. This way, you can choose which reference should be downloaded and which one should not be.
> 
> I hope my answer was clear and to hear back from you.
> 
> Best regards,
> 
> > Le 3 mai 2016 à 23:33, rakesh prithiviraj <rakeshuk at outlook.com> a écrit :
> > 
> > Thanks Gerald. 
> > 
> > In one of my use cases, there are nearly hundreds of input files (all are reference inputs) and it the responsibility of the processor to download and check the validity of these files in a particular sequence. In this case, I do not want the zoo-kernel to download the files but leave it to the processor to handle the download. Because if there is an error in the first few files, it is not necessary to download the rest of the input files. 
> > 
> > Also we need to support certificate based authentication to download data from data sources. So in some cases, it is better to let the processor decide when to download using the appropriate authentication mechanism. 
> > 
> > It is nice to have the default behaviour of kernel downloading the reference data. But is there any way to tell the kernel not to download the referenced input? 
> > 
> > Thanks in advance.
> > Rakesh
> > 
> > > Subject: Re: [Zoo-discuss] Handling reference input files for processing Zoo-kernel
> > > From: gerald.fenoy at geolabs.fr
> > > Date: Tue, 3 May 2016 17:22:01 +0200
> > > CC: zoo-discuss at lists.osgeo.org
> > > To: rakeshuk at outlook.com
> > > 
> > > Dear Rakesh,
> > > you should be able to access data protected by basic authentication for example. You can try accessing a simple web page requiring to authenticate and use its value from the service to make sure, but there should not be any particular problem about this.
> > > 
> > > Note also that in some software, like MapMint for example, we use our own method to authenticate the users, we use COOKIE handled by the ZOO-Kernel and so the associated session file (very similar to the main.cfg except that there is only one section in it, named sess). Once the user authenticate himself from the application, it has a COOKIE returned by the server, then the cookie pass in the input reference header (as we are mainly using the POST method) to make sure the user is authorized to access the specified resource or process.
> > > 
> > > So, to summarize what I explained, the ZOO-Kernel already support different simple authentification methods.
> > > 
> > > The ZOO-Kernel should be able to provide access to both HTTP and FTP ressources.
> > > 
> > > Best regards,
> > > 
> > > 
> > > > Le 3 mai 2016 à 16:39, rakesh prithiviraj <rakeshuk at outlook.com> a écrit :
> > > > 
> > > > Can I assume that if the data (referenced by input) is protected by some authentication, it cannot be fetched by zoo-kernel. Is that right? or does zoo-kernel supports some of the authentication configuration to fetch data from protected end points? 
> > > > 
> > > > I also believe that zoo-kernel supports HTTP and FTP endpoints for fetching the data. Is this correct?
> > > 
> > > 
> > > 
> > > Gérald Fenoy
> > > http://wiki.osgeo.org/wiki/User:Djay
> > > 
> 
> 
> 
> Gérald Fenoy
> http://wiki.osgeo.org/wiki/User:Djay
> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/zoo-discuss/attachments/20160506/652efd4f/attachment.html>