Message: Note about downloading datasets via code Not Logged In (login)
 Next-in-Thread Next-in-Thread
 Next-in-Forum Next-in-Forum

Note Note about downloading datasets via code 

Forum: Installation and Configuration
Date: Dec 18, 14:54
From: Tom Roberts <Tom Roberts>

The Geant4 datasets have moved from their original URLs (happened many moons ago). Along with the move came a change to Content-Encoding: x-gzip.

What this means is that if you use C++ or Python code to download a dataset, the underlying network library's get() will probably unzip it for you. That means "tar -xzf file.tar.gz" will probably fail. On a Mac, tar is forgiving and it works on an already-unzipped file, but on Linux it fails; on windows, bsdtar.exe fails. Note that unless you do something special, the gunzipped file will have the confusing name "file.tar.gz".

One solution is to omit the "z" from the options to tar; the solution that avoids the confusing name is to add the header "Accept-Encoding: gzip" to the get request, so the downloaded file will remain in gzip format.

Sadly, we are still saddled with the curious situation that each dataset has three different names: the name you use to download, the name of the environment variable, and the name of the directory to which it is unzipped. Experienced users can easily recognize them, but code cannot, so code to assist users in downloading datasets needs a detailed table to translate names.

 Add Message Add Message
to: "Note about downloading datasets via code"

 Subscribe Subscribe

This site runs SLAC HyperNews version 1.11-slac-98, derived from the original HyperNews


[ Geant 4 Home | Geant 4 HyperNews | Search | Request New Forum | Feedback ]