#!/usr/bin/env python
"""Usage: FootnoteSorter.py [-a] [--test] [< input.txt]
    -a     Number footnotes in order of apperance in text
    --test Self-testing using the SAMPLE in the script source.
"""

SAMPLE = '''A great brown fox [13] jumped of a pile of lorem ipsum [4], [7]. He
met with a silver penguin, browsing the Linux Kernel Mailinglist [3]. They
debated other the question whether to start a C-program with "main (int argc,
char **argv)" or with "main (int argc, char *argv[])". Square brackets annoyed
them [9999]. Multiple references may exist to same targets [4].

@footnote:

[13] Al Fabetus: "On characters and animals", 1888, self published.
[4] Lorem Ipsum, <a href="http://en.wikipedia.org/wiki/Lorem_ipsum">Web Link</a>
[9999] Annoying Link.
[7] B. Fox: "More on Blind Text".
[3] Linux Kernel Maintainers: LKML
'''

import sys, re, fileinput, tempfile, os

FOOTNOTE_PATTERN = re.compile(r'\[\d+\]')

def checkOption(args, option):
    '''Check if an option is contained in the array args.
    If it is in the array, it will be deleted.'''
    try:
        index = args.index (option)
        del args[index]
        return True
    except ValueError:
        return False

def fileFilter(input, old2New):
    '''Filter every line of input.'''
    
    def op(m, old2New=old2New):
        '''This gets called for every match of the FOOTNOTE_PATTERN.'''
        key = m.group(0)
        # Check if we already have a mapping for this reference
        newValue = old2New.get(key)
        if newValue is None:
            # If not, create a new one
            newValue = '[%d]' % (len(old2New) + 1)
            old2New[key] = newValue
        return newValue

    # Read every line of input and replace the footnote references in it
    # using the operator defined above.
    for line in input:
        line = FOOTNOTE_PATTERN.sub(op, line)
        sys.stdout.write(line)

if __name__ == '__main__':
    isTest = checkOption (sys.argv, '--test')
    byFirstOccurrence = checkOption (sys.argv, '-a')

    if isTest:
        # Create a text input from the sample
        input = SAMPLE.split('\n')
        for i in range(0,len(input)):
            input[i] = '%s\n' % input[i]
    else:
        input = fileinput.FileInput(openhook=fileinput.hook_compressed)
    
    if byFirstOccurrence:
        # This is a simple case. Just read the file
        # and build the map as we encounter references.
        fileFilter (input, {})
    else:
        # Here, we need a temporary file because we have to do
        # two passes. In the first pass, we figure out the
        # correct order by reading the footnote block.
        #
        # In the second pass, we use the mapping to filter
        # the file just as in the simple case.
        #
        # Since the filter will find existing mappings, it won't
        # create them by itself and thus, we get the desired result. 
        tmpFD, tmpFile = tempfile.mkstemp(text=True)
        try:
            fh = os.fdopen(tmpFD, 'w+')
            
            old2New = {}
            isText = True
            for line in input:
                # Make a copy of the input
                fh.write(line)
                if isText:
                    if line.startswith('@footnote:'):
                        isText = False
                    continue
                else:
                    # If we are in the footnote part, create the mapping
                    # by adding new references as we encounter them.
                    pos = line.find(']')
                    if pos == -1:
                        continue
                    oldValue = line[0:pos+1]
                    newValue = '[%d]' % (len(old2New) + 1)
                    old2New[oldValue] = newValue
            
            # Read the temporary file again from the start
            fh.seek(0)
            # and filter it just as above.
            fileFilter (fh, old2New)
            fh.close()
        finally:
            # Always delete the temporary file
            os.remove(tmpFile)