Linux:

  1. Install IDL:
  2. First make sure Interactive Data Language (IDL) v5.4 or above is installed on your system. This is bloody expensive, so if you don't already have a full IDL licence, there are a few cheaper alternatives worth considering first. You should then be able to start IDL by typing "idl" from a prompt, or "idlde" for a fancier interface with editing and debugging tools. To view IDL's built-in manual, type "?"; to quit IDL, type "exit". Although the shapelets code requires IDL to run, some of the main routines can also be accessed directly from a Linux prompt via simple shell scripts that are provided along with the code.

  3. Download the shapelet routines:
  4. Download the latest versions of the shapelets source code and example data. The example file is rather large, but it will be needed in a moment for the example walkthrough of the main routines. Unpack the file into a new directory somewhere by typing "gunzip < shapelets_code+data.tar.gz | tar xfv -". I store mine inside ~/idl/. If you want to put them somewhere else, just be sure to specify their location during the next task.

  5. Tell IDL where the shapelet routines are stored:
  6. IDL only looks for routines inside a specified path, so you need to add your new "shapelets" directory. There are three equivalent ways to do this in Linux, so just pick the most convenient one for you.

  7. Tell the shapelet routines where your data is stored:
  8. I don't have very much space on my home disc, so I store data elsewhere. To avoid constantly typing in long file paths, the shapelet routines read and write within a directory specified in shapelets_paths.pro. (You can also disable to behaviour via the /FULL_PATH keyword). The data directory is initally set up to be ~/data/ but, if everything downloaded correctly, you should find the example data in .../shapelets/data/. Either move the data or edit the location specified in shapelets_paths.pro with your favourite text editor. Emacs is a good choice because it has a clever IDLWAVE add-on which can automatically highlight IDL syntax in a variety of colours.

    The path names are strings, contained in quotes. The base IDL directory (option 0) should return the directory where you uncompressed the zip files, containing shapelets_paths.pro. Options are 1, 2 and 3 point to directories containing your data. You could just set them to call the base directory and just append "data/". It would also be a good idea to personalise the string "Shapelets export version of path names!". This does nothing important, but is printed whenever a path is requested. It will be reassuring to know that IDL is looking in the right place, and will serve as a warning if you later install an updated version of the shapelet routines and overwrite this file. Once it is correct, you could also write-protect this file to avoid replacing it by typing "chmod -w shapelets_paths.pro". More information on this routine is contained in the shapelets_paths help page.

  9. Install SExtractor: (optional)
  10. For the object detection routines (only), you will also need to install SExtractor by Emmanuel Bertin. The book "Source Extractor for Dummies" by Benne Holwerda is often more useful and complete that the official manual. You will need to know the full path name to the SExtractor executable file on your system. If it was installed by the systems administrator, type "which sex" at a Linux prompt to find this path.

  11. Try the demonstration routine:
  12. You are now ready to try out the example walkthough of the most important shapelet routines.




Valid HTML 4.01!

Valid CSS!