#! /usr/bin/env python3

import argparse
import datetime
import hashlib                  # md5
import lxml.etree as ET
import os
import re
import shutil
import subprocess
import time
import sgprops
import sys
import catalogTags
import catalog
from catalog import make_aircraft_node, make_aircraft_zip, parse_config_file, parse_template_file


CATALOG_VERSION = 4

# The Python version.
PY_VERSION = sys.version_info[0]

parser = argparse.ArgumentParser()
parser.add_argument("--update", help="Update/pull SCM source",
                    action="store_true")
parser.add_argument("--no-update",
                    help="Disable updating from SCM source",
                    action="store_true")
parser.add_argument("--clean", help="Force regeneration of all zip files",
                    action="store_true")
parser.add_argument("--quiet", help="Only print warnings and errors",
                    action="store_true")
parser.add_argument("dir", help="Catalog directory")
args = parser.parse_args()

includes = []
mirrors = [] # mirror base URLs

# xml node (robust) get text helper
def get_xml_text(e):
    if e != None and e.text != None:
        return e.text
    else:
        return ''

# use svn commands to report the last change date within dir
def last_change_date_svn(dir):
    command = [ 'svn', 'info', dir ]
    result = subprocess.check_output( command )

    # Python 3 compatibility.
    if PY_VERSION == 3:
        result = result.decode('utf8')

    match = re.search('Last Changed Date: (\d+)\-(\d+)\-(\d+)', result)
    if match:
        rev_str = match.group(1) + match.group(2) + match.group(3)
        return int(rev_str)

# find the most recent mtime within a directory subtree
def scan_dir_for_change_date_mtime(path):
    maxsec = 0
    names = os.listdir(path)
    for name in names:
        fullname = os.path.join(path, name)
        if name == '.' or name == '..':
            pass
        elif os.path.isdir( fullname ):
            mtime = scan_dir_for_change_date_mtime( fullname )
            if mtime > maxsec:
                maxsec = mtime
        else:
            mtime = os.path.getmtime( fullname )
            if mtime > maxsec:
                maxsec = mtime
    return maxsec


def get_md5sum(file):
    f = open(file, 'rb')
    md5sum = hashlib.md5(f.read()).hexdigest()
    f.close()
    return md5sum

def copy_previews_for_variant(variant, package_name, package_dir, previews_dir):
    if not 'previews' in variant:
        return

    for preview in variant['previews']:
        preview_src = os.path.join(package_dir, preview['path'])
        preview_dst = os.path.join(previews_dir, package_name + '_' + preview['path'])
        #print(preview_src, preview_dst, preview['path'])
        dir = os.path.dirname(preview_dst)
        if not os.path.isdir(dir):
            os.makedirs(dir)
        if os.path.exists(preview_src):
            shutil.copy2(preview_src, preview_dst)

def copy_previews_for_package(package, variants, package_name, package_dir, previews_dir):
    copy_previews_for_variant(package, package_name, package_dir, previews_dir)
    for v in variants:
        copy_previews_for_variant(v, package_name, package_dir, previews_dir)

def copy_thumbnail_for_variant(variant, package_name, package_dir, thumbnails_dir):
    if not 'thumbnail' in variant:
        return

    thumb_src = os.path.join(package_dir, variant['thumbnail'])
    thumb_dst = os.path.join(thumbnails_dir, package_name + '_' + variant['thumbnail'])

    dir = os.path.dirname(thumb_dst)
    if not os.path.isdir(dir):
        os.makedirs(dir)
    if os.path.exists(thumb_src):
        shutil.copy2(thumb_src, thumb_dst)

def copy_thumbnails_for_package(package, variants, package_name, package_dir, thumbnails_dir):
    copy_thumbnail_for_variant(package, package_name, package_dir, thumbnails_dir)

    # and now each variant in turn
    for v in variants:
        copy_thumbnail_for_variant(v, package_name, package_dir, thumbnails_dir)

def process_aircraft_dir(name, repo_path):
    global includes
    global download_base
    global output_dir
    global valid_zips
    global previews_dir
    global mirrors

    aircraft_dir = os.path.join(repo_path, name)
    if not os.path.isdir(aircraft_dir):
        return

    (package, variants) = catalog.scan_aircraft_dir(aircraft_dir, includes)
    if package == None:
        if not args.quiet:
            print("skipping: %s (no -set.xml files)" % name)
        return

    if not args.quiet:
        print("%s:" % name)

    package_node = make_aircraft_node(name, package, variants, download_base, mirrors)

    download_url = download_base + name + '.zip'
    if 'thumbnail' in package:
        # this is never even used, but breaks the script by assuming
        # all aircraft packages have thumbnails defined?
        thumbnail_url = download_base + 'thumbnails/' + name + '_' + package['thumbnail']

    # get cached md5sum if it exists
    md5sum = get_xml_text(md5sum_root.find(str('aircraft_' + name)))

    # now do the packaging and rev number stuff
    dir_mtime = scan_dir_for_change_date_mtime(aircraft_dir)
    if repo_type == 'svn':
        rev = last_change_date_svn(aircraft_dir)
    else:
        d = datetime.datetime.utcfromtimestamp(dir_mtime)
        rev = d.strftime("%Y%m%d")
    package_node.append( catalog.make_xml_leaf('revision', rev) )
    #print("rev: %s" % rev)
    #print("dir mtime: %s" % dir_mtime)
    zipfile = os.path.join( output_dir, name + '.zip' )
    valid_zips.append(name + '.zip')
    if not os.path.exists(zipfile) \
       or dir_mtime > os.path.getmtime(zipfile) \
       or args.clean:
        # rebuild zip file
        if not args.quiet:
            print("updating: %s" % zipfile)
        make_aircraft_zip(repo_path, name, zipfile, zip_excludes, verbose=not args.quiet)
        md5sum = get_md5sum(zipfile)
    else:
        if not args.quiet:
            print("(no change)")
        if md5sum == "":
            md5sum = get_md5sum(zipfile)
    filesize = os.path.getsize(zipfile)
    package_node.append( catalog.make_xml_leaf('md5', md5sum) )
    package_node.append( catalog.make_xml_leaf('file-size-bytes', filesize) )

    # handle md5sum cache
    node = md5sum_root.find('aircraft_' + name)
    if node != None:
        node.text = md5sum
    else:
        md5sum_root.append( catalog.make_xml_leaf('aircraft_' + name, md5sum) )

    # handle sharing
    if share_md5sum_root != None:
        sharedNode = share_md5sum_root.find(str('aircraft_' + name))
        if node != None:
            shared_md5 = get_xml_text(sharedNode)
            if shared_md5 == md5sum:
                if not args.quiet:
                    print("Sharing zip with share catalog for: %s" % name)
                os.remove(zipfile)
                os.symlink(os.path.join( share_output_dir, name + '.zip' ), zipfile)


    # handle thumbnails
    copy_thumbnails_for_package(package, variants, name, aircraft_dir, thumbnail_dir)

    catalog_node.append(package_node)

    # copy previews for the package and variants into the
    # output directory
    copy_previews_for_package(package, variants, name, aircraft_dir, previews_dir)

#def get_file_stats(file):
#    f = open(file, 'r')
#    md5 = hashlib.md5(f.read()).hexdigest()
#    file_size = os.path.getsize(file)
#    return (md5, file_size)

if not os.path.isdir(args.dir):
    print("A valid catalog directory must be provided")
    exit(0)

parser = ET.XMLParser(remove_blank_text=True)
config_node = parse_config_file(parser=parser, file_name=os.path.join(args.dir, 'catalog.config.xml'))
template_node = parse_template_file(parser=parser, file_name=os.path.join(args.dir, 'template.xml'))

md5sum_file = os.path.join(args.dir, 'md5sum.xml')
if os.path.exists(md5sum_file):
    md5sum_tree = ET.parse(md5sum_file, parser)
    md5sum_root = md5sum_tree.getroot()
else:
    md5sum_root = ET.Element('PropertyList')
    md5sum_tree = ET.ElementTree(md5sum_root)

# share .zip files with other output dirs
share_output_dir = get_xml_text(config_node.find('share-output'))
share_md5_file = get_xml_text(config_node.find('share-md5-sums'))
if share_output_dir != '' and share_md5_file != '':
    print("Output shared with: %s" % share_output_dir)
    share_md5sum_tree = ET.parse(share_md5_file, parser)
    share_md5sum_root = share_md5sum_tree.getroot()
else:
    share_md5sum_root = None

# SCM providers
scm_list = config_node.findall('scm')
upload_node = config_node.find('upload')

download_base = None
for i in config_node.findall("download-url"):
    url = get_xml_text(i)
    if not url.endswith('/'):
        url += '/'

    if download_base == None:
        # download_base is the first entry
        download_base = url
    else:
        mirrors.append(url)

output_dir = get_xml_text(config_node.find('local-output'))
if output_dir == '':
    output_dir = os.path.join(args.dir, 'output')
if not os.path.isdir(output_dir):
    os.mkdir(output_dir)

thumbnail_dir = os.path.join(output_dir, 'thumbnails')
if not os.path.isdir(thumbnail_dir):
    os.mkdir(thumbnail_dir)

previews_dir = os.path.join(output_dir, 'previews')
if not os.path.isdir(previews_dir):
    os.mkdir(previews_dir)

tmp = os.path.join(args.dir, 'zip-excludes.lst')
zip_excludes = os.path.realpath(tmp)

for i in config_node.findall("include-dir"):
    path = get_xml_text(i)
    if not os.path.exists(path):
        print("Skipping missing include path: %s" % path)
        continue
    includes.append(path)

# freshen repositories
if args.no_update:
    print('Skipping repository updates.')
else:
    cwd = os.getcwd()
    for scm in scm_list:
        # XML mandated skip, with command line override.
        if not args.update:
            skip = get_xml_text(scm.find('update'))
            if skip == 'false':
                continue

        repo_type = get_xml_text(scm.find('type'))
        repo_path = get_xml_text(scm.find('path'))
        includes.append(repo_path)

        if repo_type == 'svn':
            print("SVN update: %s" % repo_path)
            subprocess.call(['svn', 'update', repo_path])
        elif repo_type == 'git':
            print("GIT pull: %s" % repo_path)
            os.chdir(repo_path)
            subprocess.call(['git','pull'])
        elif repo_type == 'no-scm':
            print("No update of unmannaged files: %s" % repo_path)
        else:
            print("Unknown scm type: %s %s" % (scm, repo_path))
    os.chdir(cwd)

# names of zip files we want (so we can identify/remove orphans)
valid_zips = []

# create the catalog tree
catalog_node = ET.Element('PropertyList')
catalog_root = ET.ElementTree(catalog_node)

# include the template configuration
for child in template_node:
    catalog_node.append(child)

# scan repositories for catalog information
for scm in scm_list:
    repo_type = get_xml_text(scm.find('type'))
    repo_path = get_xml_text(scm.find('path'))
    skip_nodes = scm.findall('skip')
    skip_list = []
    for s in skip_nodes:
        skip_list.append(get_xml_text(s))

    # Selective list of craft to include, overriding the skip list.
    include_nodes = scm.findall('include')
    include_list = []
    for node in include_nodes:
        include_list.append(get_xml_text(node))
    if len(include_list):
        skip_list = []

    print("Skip list: %s" % skip_list)
    print("Include list: %s" % include_list)

    names = os.listdir(repo_path)
    for name in sorted(names, key=lambda s: s.lower()):
        if name in skip_list or (len(include_list) and name not in include_list):
            if not args.quiet:
                print("Skipping: %s" % name)
            continue

        # process each aircraft in turn
        # print("%s %s" % (name, repo_path))
        process_aircraft_dir(name, repo_path)

# write out the master catalog file
cat_file = os.path.join(output_dir, 'catalog.xml')
catalog_root.write(cat_file, encoding='utf-8', xml_declaration=True, pretty_print=True)

# write out the md5sum cache file
print(md5sum_file)
md5sum_tree.write(md5sum_file, encoding='utf-8', xml_declaration=True, pretty_print=True)

# look for orphaned zip files
files = os.listdir(output_dir)
for file in files:
    if file.endswith('.zip')and not file in valid_zips:
        print("orphaned zip: %s" % file)