#!/bin/python
###############################################################################
#
# Diff two licenses.xml files. There are two cases:
# * Two arguments
#    In this case, the two arguments are the two files to compare
#
# * No arguments
#     In this casem the license files are specified by the following
#     parameters in pardoc-distro.xml:
#     - prev_baseline
#     - prev_lic_file
#     - new_lic_file
#
# The result is presented as three sets:
# 1) Removed
#   Packages present in the previous file, but not in the new.
#
# 2) Added
#   Packages present in the new file, but not in the previous.
#
# 3) Changed
#   Packages present in both files, but with different versions. If more than
#   one version of a package is included, then all difference in the version
#   set causes the package to be listed as changed.
#   E.g.
#    (v1) -> (v2)
#    (v1, v2) -> (v2, v3)
#
# Note that packages with the unchanged version is not listed.
#
# The output is presented as XML code printed to stdout. A summary is printed
# to STDERR at the end.
#
###############################################################################

import os
import sys
import subprocess as sp
import xml.etree.ElementTree as ET
import re

repo_root   = sp.check_output(["git", "rev-parse", "--show-toplevel"]).rstrip()
script_root = os.path.dirname(os.path.realpath(__file__))
param_file  = os.path.join(script_root, "docsrc_common/pardoc-distro.xml")

def get_param(param_name):
    pat = re.compile("%s.*>([^<>]+)</" % param_name)

    with open(param_file) as f:
        for line in f:
            m = pat.search(line)
            if m:
                return m.group(1)

    return None

def get_pkgs(file_spec):
    if file_spec.find(":") >= 0:
        s = sp.check_output(("git", "show", file_spec))
    else:
        f = open(file_spec)
        s = f.read()
        f.close()
        del f

    # ET can't handle some special quotes
    for old, new in (("&rdquor;", "&quot;"), ("&rdquo;", "&quot;")):
        s = s.replace(old, new)

    root = ET.fromstring(s)

    for node in root.iter("section"):
        if "id" in node.attrib:
            if node.attrib["id"] == "licenses_packages":
                break

    else:
        return None

    for node in node:
        if node.tag == "informaltable":
            break
    else:
        return None

    tab = node[0][-1]

    plist = dict()
    for row in tab:
        pname = row[0].text
        pver  = row[1].text
        if not pname in plist:
            plist[pname] = set()

        plist[pname].add(pver)

    return set(plist), plist

#----------------------------------------


if len(sys.argv) == 3:
    new_file, prev_file = sys.argv[1:3]

elif len(sys.argv) == 1:
    prev_baseline = get_param("prev_baseline")
    prev_lic_file = get_param("prev_lic_file")
    new_lic_file  = get_param("new_lic_file")

    if not (prev_baseline and prev_lic_file and new_lic_file):
        print '<?xml version="1.0" encoding="ISO-8859-1"?>'
        print '<!DOCTYPE chapter PUBLIC "-//OASIS//DTD DocBook XML V4.2//EN"'
        print '"http://www.oasis-open.org/docbook/xml/4.2/docbookx.dtd">'
        print '<section id="relinfo-package-set-changes">'
        print '  <title>Changes in the Set of Provided Package</title>'
        print '  <para>'
        print '    N/A. No previous baseline defined.'
        print '  </para>'
        print '</section>'
        exit(0)

    new_file = os.path.relpath(os.path.join(repo_root, new_lic_file))
    prev_file = "%s:%s" % (prev_baseline, prev_lic_file)

else:
    sys.stderr.write("Usage:\n")
    sys.stderr.write("  1) %s\n" % sys.argv[0])
    sys.stderr.write("  2) %s " % sys.argv[0])
    sys.stderr.write("<new license file> <old license file>\n")
    sys.stderr.write("\n")
    sys.stderr.write("In case 1, the files are specified using the following\n")
    sys.stderr.write("parameters in pardoc-distro.xml:\n")
    sys.stderr.write("  - prev_baseline\n")
    sys.stderr.write("  - prev_lic_file\n")
    sys.stderr.write("  - new_lic_file\n")
    exit()

sys.stderr.write("New license file  : %s\n" % new_file)
sys.stderr.write("Prev license file : %s\n" % prev_file)

old_pset, old_pdict = get_pkgs(prev_file)
new_pset, new_pdict = get_pkgs(new_file)

added   = new_pset - old_pset # Set subtraction
removed = old_pset - new_pset # Set subtraction
common  = old_pset & new_pset
changed = [ p for p in common if old_pdict[p] != new_pdict[p] ]

print '<?xml version="1.0" encoding="ISO-8859-1"?>'
print '<!DOCTYPE chapter PUBLIC "-//OASIS//DTD DocBook XML V4.2//EN"'
print '"http://www.oasis-open.org/docbook/xml/4.2/docbookx.dtd">'
print '<section id="relinfo-package-set-changes">'
print '  <title>Changes in the Set of Provided Package</title>'
print '  <para>'
print '    This section describes changes in the provided packages.'
print '  </para>'
print '  <section id="relinfo-added-packages">'
print '    <title>Added Packages</title>'
print '    <informaltable>'
print '      <tgroup cols="2">'
print '        <thead>'
print '          <row>'
print '            <entry>Package</entry>'
print '            <entry>Version(s)</entry>'
print '          </row>'
print '        </thead>'
print '        <tbody>'

for p in sorted(list(added)):
    print '          <row>'
    print '            <entry>%s</entry>' % p
    print '            <entry>%s</entry>' % ", ".join(sorted(new_pdict[p]))
    print '          </row>'

print '        </tbody>'
print '      </tgroup>'
print '    </informaltable>'
print '  </section>'

print '  <section id="relinfo-removed-packages">'
print '    <title>Removed Packages</title>'
print '    <informaltable>'
print '      <tgroup cols="2">'
print '        <thead>'
print '          <row>'
print '            <entry>Package</entry>'
print '            <entry>Version(s)</entry>'
print '          </row>'
print '        </thead>'
print '        <tbody>'
for p in sorted(list(removed)):
    print '          <row>'
    print '            <entry>%s</entry>' % p
    print '            <entry>%s</entry>' % ", ".join(sorted(old_pdict[p]))
    print '          </row>'
print '        </tbody>'
print '      </tgroup>'
print '    </informaltable>'
print '  </section>'

print '  <section id="relinfo-changed-packages">'
print '    <title>Changed Package Versions</title>'
print '    <informaltable>'
print '      <tgroup cols="3">'
print '        <thead>'
print '          <row>'
print '            <entry>Package</entry>'
print '            <entry>Previous Version(s)</entry>'
print '            <entry>New Version(s)</entry>'
print '          </row>'
print '        </thead>'
print '        <tbody>'
for p in sorted(list(changed)):
    print '          <row>'
    print '            <entry>%s</entry>' % p
    print '            <entry>%s</entry>' % ", ".join(sorted(old_pdict[p]))
    print '            <entry>%s</entry>' % ", ".join(sorted(new_pdict[p]))
    print '          </row>'

print '        </tbody>'
print '      </tgroup>'
print '    </informaltable>'
print '  </section>'
print '</section>'

sys.stderr.write("Package Summary:\n")
sys.stderr.write("  Prev file : %3d\n" % len(old_pset))
sys.stderr.write("  New file  : %3d\n" % len(new_pset))
sys.stderr.write("  Added     : %3d\n" % len(added))
sys.stderr.write("  Removed   : %3d\n" % len(removed))
sys.stderr.write("  Changed   : %3d\n" % len(changed))
sys.stderr.write("  Unchanged : %3d\n" % (len(common) - len(changed)))
sys.stderr.write("Done\n")