#!/usr/bin/python
#-*- coding: UTF-8 -*-

#************************************************
# showDep.py
#
# This script process the output of the perl
# script used to retrieve the dependencies information
# from the bioconductor SVN.
# It aims to get the list of dependencies for a 
# specified list of libraries.
#
#
# Made by Pierre-Yves chibon
#
# Version 1.1 -- 28th July 2008
#   * Add the number of libraries to package at the end
#   * Small layout fix
#
# Version 1.0 -- 25th July 2008
#   * Gives for a list of libraries all their 
#       dependencies and sub-dependencies
#
# Distributed under License GPLv3 or later
# You can find a copy of this license on the website
# http://www.gnu.org/licenses/gpl.html
#
#***********************************************

import getopt, sys, re#, shutil, os

## Here is defined list of libraries of interest
Rlib = ['affy', 'affydata', 'affyPLM', 'annaffy', 'annotate', 'Biobase', 'Biostrings', 'DynDoc', 'gcrma', 'genefilter', 'geneplotter', 'hgu95av2.db', 'limma', 'marray', 'matchprobes', 'multtest', 'ROC', 'vsn', 'xtable', 'affyQCReport']
# Empty list that will contain the list of library 
#   investigated by the program
libProcessed=[]

#################################################
USAGE ="""
This program is designed to parse the output generated by the perl script that retrieve the dependencies information from the Bioconductor SVN.
Usage:
  python probe2seq.py -i file
  python probe2seq.py --inpuFile=file

The list of library investigate is defined line 29 of the script.
"""
#################################################

#################################################
### argument
# retrieve the arguments given and show the help when needed
def argument():
        try:
		opts, args = getopt.getopt(sys.argv[1:], "i:", ["inputFile="])
        except getopt.GetoptError:
		print USAGE
		sys.exit(1)
        inputfile = None

        for option, argument in opts:
                if option in ("-i", "--inputFile"):
                    inputfile = argument
                		
	return (inputfile)
################################################

#################################################
### search
# search in the file given the dependencies of the library
def search(File,lib, Rlib, libProcessed):
    # Open the input file
    sF=open(File, 'r')
    # Print the library for which the dependancies are looked for
    print lib
    # Read the file
    for line in sF.readlines():
	# Clean each line of the file
	line = line.replace(";\n","").replace("\"","").replace("\t","")
	# Check wether the line has the '->'
	versionMotif = re.compile('->')
	tmp = versionMotif.findall(line)
	if tmp:
	    # Split the line in two part among the '->'
	    (rlib,libDep) = line.split( ' -> ' )
	    # If the library in the first part of the line == the library
	    #   of interest
	    #if (rlib == lib and not libDep in libProcessed ): #reduced the number of libraries shown, does not reflect really the reality
	    if (rlib == lib ):
		# Add this library to the list and print it
		Rlib.append(libDep)
		print "\t\t",libDep
    # Add the library looked for to the list of already processed library
    libProcessed.append(lib)
    
    # Close the file and return the values
    sF.close()
    return(Rlib, libProcessed)
    #sys.exit(0)
	

######### Main #############
inputfile=argument()
# Show all the argument given #Debugging purpose
#print  "\ninputFile from sys.argv ==> ", inputfile,"\n"

# Return the usage message if one arguments is missing
if (inputfile == None):
	print USAGE
	sys.exit(0)

# Layout
print "The script starts for the libraries :",Rlib,"\n"
print "Library \t Dependencies"

# Process all the libraries
for lib in Rlib:
    if not lib in libProcessed:
	(Rlib,libProcessed) = search(inputfile,lib, Rlib,libProcessed)

# print Rlib # For debugging purpose
# print '# Number of R libraries',len(Rlib)
# print libProcessed # For debugging purpose
print '# Number of libraries to package',len(libProcessed)