test files

nilc/README.org

+#+TITLE: Needlet ILC tool
+#+STYLE: <link rel="stylesheet" type="text/css" href="org.css" />
+#+AUTHOR: M. Le Jeune
+#+DATE: 2011-01-05 13:58:59 CET
+#+EMAIL: lejeune@apc.univ-paris7.fr
+
+The Needlet ILC tool is simple binary aimed at performing Internal
+Linear Combination (ILC) of Healpix maps.
+'Needlet' stands for spherical wavelet, as it offers the possibility
+to perform ILC in the needlet analysis domain. 
+It can also be used in the simplest case of pixel domain analysis as
+described below. 
+
+* Installation
+
+The NILC tool relies on the *Healpix* C++ library
+ (http://healpix.jpl.nasa.gov/) and the *spherelib* library which
+ extends the Healpix package with needlet analysis and filtering
+ routines.
+
+Those two libraries are provided and automatically installed with the
+NILC source code.
+
+** Dependencies
+
+To build the binary, you will need the Scons construction tool.  Get
+it from your package manager or see http://www.scons.org
+
+It is also possible to use scons-local by downloading
+scons-local-xxxx.tar.gz, unpacking the tar ball in the current
+directory and then running:
+ =python scons.py -f SConstruct_local=
+after the file =SConstruct_local_example= has been edited into
+=SConstruct_local= to fit your system.
+
+** Installing the NILC tool
+
+Change directory to the NILC source code repository and run:  
+
+ =scons=
+
+To clean up after a build: 
+
+ =scons -c=
+
+To install the binary: 
+
+ =scons install --prefix=mydir=
+
+Default prefix is '/usr/local'.
+
+
+** Running the NILC tool
+
+A simple test can be run using the simulated data provided with the
+NILC package: 
+
+ =nilc fullsky.param=
+
+
+* Pixel based ILC
+
+The NILC tool can be used for pixel-based analysis performed:
++ on the full sky;
++ on a patch or using a galactic mask;
++ on a set of pixel zones.
+
+** Full sky analyis
+
+In this case, input parameters are minimum as in this simple example: 
+
+[[file:./fullsky.param]]
+
+** Patch analysis
+
+
+In order to exclude a zone of the analysis, one can define a binary
+mask which is an Healpix map with pixel equal to 0 in the excluded
+zone and 1 elsewhere:
+
+file:./patch.param
+ 
+
+** Localized analysis
+
+In order to performed a space localized ILC, this Healpix map can be
+used to define more pixel zones. 
+In this case, each pixel is set to its zone number starting from 1:
+
+file:./grid.param
+
+
+* Needlet ILC
+
+The Needlet ILC is based on the possibility to localized the
+statistics over both space and spectral domains.  
+
+The spectral localisation is made available through the use of spline
+window functions. Window function limits are set in a vector with the
+following convention: 
+ =bands=[lmin_0 lmin_1 lmin_2...lmin_J lmin_J-1 lmax_J]=
+with =lmax_j-2=lmin_j=. 
+
+The space localisation uses the same convention as described in the
+previous section, except that one has to define as many Healpix zone
+maps as needlet scales (J in this example). 
+
+file:./needlet.param
+
+
+* More options
+
++ Input files are Healpix map by default. It is possible to use alm
+  files instead by setting the option almopt to =true=. In this case,
+  one has to specify which resolution will be used for the output map
+  with the =nside= parameter.
++ Spherical harmonic transforms can be tuned with the =n_iter=
+  parameter.
++ One can use command line options instead of parameter file: 
+ + In the pixel based case: 
+   =nilc -p outfile -z zonefile N infile1 coeff1 ... infileN coeffN=
+ + In the needlet case:
+   =nilc outfile l0 J zonefile1 lmax1 ... zonefileJ lmaxJ N infile1 coeff1 ... infileN coeffN=