python-sardana 1.2.0-2 / usr / bin / h5tospec

#! /usr/bin/python

##############################################################################
##
## This file is part of Sardana
##
## http://www.tango-controls.org/static/sardana/latest/doc/html/index.html
##
## Copyright 2011 CELLS / ALBA Synchrotron, Bellaterra, Spain
## 
## Sardana is free software: you can redistribute it and/or modify
## it under the terms of the GNU Lesser General Public License as published by
## the Free Software Foundation, either version 3 of the License, or
## (at your option) any later version.
## 
## Sardana is distributed in the hope that it will be useful,
## but WITHOUT ANY WARRANTY; without even the implied warranty of
## MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
## GNU Lesser General Public License for more details.
## 
## You should have received a copy of the GNU Lesser General Public License
## along with Sardana.  If not, see <http://www.gnu.org/licenses/>.
##
##############################################################################

'''
Script for extracting "scans" from Sardana NeXus files into Sardana Spec Files.

Usage:
h5spec <nexusfilename> [> <your_spec_file_name>] [2><your error/log file]

Note: only scalar values are extracted, Images and 1D (MCA..) are ignored (notified to stderr)   
'''

import nxs
import sys,os,re,time


def h52spectime(mydate):
    '''
    this function returns the time in format SPEC (from a hdf5 recorded by the sardanan recorder)
    '''
    #equivalent to ascii time
    #return time.strftime("%a %b %d %H:%M:%S %Y",_d)
    _a = time.asctime(time.strptime(mydate.split('.')[0],"%Y-%m-%dT%H:%M:%S"))
    return _a

def _getMovables(mystr):
    '''
    this internal function gets the movables from the "scan/mesh" command. 
    Movables should go first in SPEC columns
    '''
    #gets the movables from the command line (starting not with a number)
    _m = []
    _a = mystr.split(' ') 
    _a.pop(0)
    for i in _a:
        if re.match('[^-_0-9\.\+\\n]',i):
            _m.append(i)
    return _m 

def getMetadata(fd, entry):
    '''
    This gets the metatdata from the Sardana h5 file. Format is kind of free, 
    but nevertheless is very convenient the user/scientist makes sure all motors 
    (and motor positions) are there
    ''' 
    _mdata = {}
    try:
        fd.openpath("/%s/measurement/pre_scan_snapshot" % entry)
        _mdata = fd.getentries()
        for i in _mdata.keys():
            fd.openpath("/%s/measurement/pre_scan_snapshot/%s" % (entry, i))
            _mdata[i] = fd.getdata()
            #print >> sys.stderr, entry
    except:
        print >> sys.stderr, 'Cannot open hdf5 path "%s/pre_scan_snapshot". Skipping.'%entry
    return _mdata

def processEntry(fd, entry):

    '''
    process one entry from a hdf5 file
    it creates a string with all the data and prints it to the stdout in the end. 
    It could be evenually easily modified to return the values and to be introduced 
    in a class (to get it out from this simple script)

    '''

    #_myMovables = [] # This gets the motors in order to write them in the first columns
    mystring = "" # the string will construct the scan (and printed at the end)

    try:
        fd.openpath("/%s/entry_identifier" % entry)
        mystring = "#S %s " % str(fd.getdata())
        fd.openpath("/%s/title" % entry)
        _mycmd = str(fd.getdata())
        mystring += " %s \n" % _mycmd
        _myMovables= _getMovables(_mycmd)
        fd.openpath("/%s/start_time" % entry)
        mystring += "#D %s \n" % h52spectime(fd.getdata())
        # need to know here the timer for the counting time. In the mean time 
        # I use the last value of the cmd line (usually the ct time!!, not always ojo). 
        mystring += "#T %s  (seconds)\n" % _mycmd.split(' ').pop()
        #print mytime
    except:
        print >> sys.stderr,  'Error opening/processing hdf5 path "%s/...". Skipping.'% entry

    mystring += "#P0 " 
    _metadata =  getMetadata(fd, entry)
    for i in _metadata.values(): 
        mystring += ' ' + str(i)   
    mystring += '\n'

    tmp = "/%s/measurement/point_nb" % entry
    tmp2= "/%s/measurement/Pt_No" % entry

    try:
        fd.openpath(tmp)
    except:
        #print 'Cannot open hdf5 path "%s". Skipping.'%tmp
        try:
            fd.openpath(tmp2)
        except:
            #print 'Cannot open hdf5 path "%s". Skipping.'%tmp
            return False 
    ptnb = fd.getdata()
    fd.closedata()

    headers = [] 
    data = {} 

    for name,nxclass in fd.entries():
        if name == 'point_nb' or name == "#Pt No": continue
        if name == 'pre_scan_snapshot' : continue
        dshape,dtype = fd.getinfo()

        if tuple(dshape) != ptnb.shape: 
            print >> sys.stderr, "Skipping (no data or not a scalar ???) %s in %s" % (name, entry),
            print >> sys.stderr, ptnb.shape
            continue #not a scalar (incompatible shape)

        #insert data and headers in the rigth order (movables first)
        mydata = fd.getdata().flatten()
        if name in _myMovables:
            headers.insert(_myMovables.index(name), name)
        else:
            headers.append(name)
        data[name] = mydata 
    
    mystring += "#N  %d \n" % len(headers)
    mystring += "#L " 
    for i in headers:
        mystring += '  ' + i
    mystring += "\n" 

    #print >>sys.stderr, headers
    #print >>sys.stderr, _myMovables
    #print >>sys.stderr, ptnb.size
    #print >>sys.stderr, data
    #print >>sys.stderr, ptnb

    for i in range(ptnb.size):
        # Movables (motors) come first. Then the other columns  (exp channels)
        for n in [channel for channel in _myMovables if channel in headers]: 
            mystring += ' ' + str(data[n][i])
        for n in [channel for channel in headers if channel not in _myMovables]: 
            mystring += ' ' + str(data[n][i])
        mystring += '\n'

    try:
        fd.openpath("/%s/end_time" % entry)
        mystring += "#C Scan ended at %s \n" % h52spectime(fd.getdata())
    except:
        print >> sys.stderr, 'Cannot open hdf5 path "%s/end_time". Skipping.'% entry

    print mystring
    

def measurement2spec(fd):
    """
    Takes a file (hdf5) as an inputProcess a hdf5 file and outputs the result to the stdout. 
    One can redirect the output to create the file.
    Information and error messages are output to stderr, in order not to pollute the file 

    """

    mystring = ""
    mystring += "#F %s\n" % fd.filename
    outFileAttrs =  "#CCC original Nexus hdf5 file: %s\n" % fd.getattrs()
    outCurrentDate = "#C conversion date: %s\n" % time.asctime()

    entrynames = [name for name,c in fd.entries() if c=='NXentry']
    
    mystring += "#E %s\n" % fd.getattrs()['epoch']
    mystring += "#D %s\n" % time.asctime(time.localtime(fd.getattrs()['epoch']))
    mystring += "#C User \n" 

    # Gets metadata headers (motor names usually) from the first entry (asuming all the same!)
    _metadata = getMetadata(fd,entrynames[0])
    mystring += "#O0" 
    for i in _metadata.keys(): 
        mystring += '   ' + i 

    mystring += '\n' + outFileAttrs 
    mystring += outCurrentDate 
    print mystring 
  
    #Sort entry2 before entry10 (otherwhise, it was strictly alphabetic) 
    for entry in sorted(entrynames, key = len):
        processEntry(fd, entry)

def main():
    '''
    process the entire file (todo: many things. Do only the selected entries, convert it to a class)
    '''
    if len(sys.argv) > 1:
        fname = sys.argv[1]
    else:
        print "Usage:\npython h5tospec <nexusfilename> [> <your_spec_file_(or_to_console)> ] [2> log/error file]"
        sys.exit(1)
        
    fd = nxs.open(fname,'r')
    
    measurement2spec(fd)
    
    fd.close()

if __name__ == "__main__":
    main()