Overview of the Package correlations

General workflow

The package correlations provides some tools in order to compute all the correlations from a directory with a valid architecture (See requirement in the module correlations.correlationsMain):

General specification

1. Centralized case:

   Given:

   • A ‘LoadDirectory’.
   • A ‘SaveDirectory’.
   • A first component.
   • A second component.
   • A set of dates.
   • A set of stations.

   Make:

   1. Compute the correlations of all the couples of traces from the ‘LoadDirectory’ whose correspond to the FirstComponent, the SecondComponent, the set of dates and the set of stations.

   2. Save the correlations in the ‘SaveDirectory’ with architecture:

      SaveDirectory -> FirstComponent_SecondComponent -> Year -> JulianDay -> FirstStation -> FirstStation_SecondStation.npy

2. Decentralized case:

   Given:

   • A ‘LoadDirectory’.

   • A ‘SaveDirectory’.

   • A first component.

   • A second component.

   • For the dates:

     • A set of dates.
     • A number of sublist of dates.
     • An index of sublist of dates.

   • For the station:

     • A set of stations.
     • A number of sublist of stations.
     • Either one index of sublist of stations.
     • Or either two indexes for two disjoints sublists of stations.

   Make:

   1. Compute the correlations of all the couples of traces from the ‘LoadDirectory’ whose correspond to the FirstComponent, the SecondComponent, the sublists of dates and the one or two sublists of stations.

   2. Save the correlations in the ‘SaveDirectory’ with architecture:

      SaveDirectory -> FirstComponent_SecondComponent -> Year -> JulianDay -> FirstStation -> FirstStation_SecondStation.npy

General architecture

Architecture of modules (prefix “correlations” is omitted):

• CHANGE the modules correlations.correlationsParam in order to set the parameters.
• LAUNCH with python the module correlations.correlationsMain with the in line parameters for computing the correlations of a set of couples traces in a directory.
• REQUIRE python2.7, numpy, scipy (scientific libraries).

See also the tutorial