import os
import os.path
import filecmp
import shutil

class Image:
    def __init__(self, path):
        assert os.path.isfile(path)
        self.path = path
    
    def last_dir_name(self):
        dirname = os.path.dirname(self.path)
        return os.path.basename(dirname)
    
    def filename(self):
        return os.path.basename(self.path)
    
    def size(self):
        return os.path.getsize(self.path)
    
    def __eq__(self, other):
        return (self.filename() == other.filename() and
            self.size() == other.size() and
            filecmp.cmp(self.path, other.path, 0))
    
    def __hash__(self):
        return hash(self.filename()) ^ hash(self.size())


def collect_all_files(path, existing = None):
    if existing is None:
        existing = set()
    
    results = set()
    
    dupes = []
    
    def callback(arg, dirname, fnames):
        print "Current: %20s\r" % os.path.basename(dirname)[:20],
        fnames.sort()
        for filename in fnames:
            f = os.path.join(dirname, filename)
            if not os.path.isfile(f):
                continue
            
            i = Image(f)
            if i in results or i in existing:
                dupes.append(i)
            else:
                results.add(i)
    
    os.path.walk(path, callback, None)
    
    return results, dupes

def copy_files_to(path, images):
    for image in images:
        this_dir = os.path.join(path, image.last_dir_name())
        if not os.path.isdir(this_dir):
            os.mkdir(this_dir)
        
        dest_path = os.path.join(this_dir, image.filename())
        
        if os.path.exists(dest_path) and image.filename() in ['Thumbs.db', '.picasa.ini']:
            continue
        
        while os.path.exists(dest_path):
            print "File " + dest_path + " already exists, changing name."
            root, ext = os.path.splitext(dest_path)
            root += "x"
            dest_path = root + ext
        
        shutil.copy2(image.path, dest_path)

if __name__ == '__main__':
    import sys
    
    if len(sys.argv) == 2:
        assert os.path.isdir(sys.argv[1])
        source = sys.argv[1]
        dest = os.path.dirname(os.path.abspath(__file__))
        print "Copying non-dupes to " + dest
    elif len(sys.argv) == 3 and sys.argv[1] == '-n':
        assert os.path.isdir(sys.argv[2])
        print "Checking only"
        source = sys.argv[2]
        dest = None
    elif len(sys.argv) == 3:
        assert os.path.isdir(sys.argv[1])
        assert os.path.isdir(sys.argv[2])
        source = sys.argv[1]
        dest = sys.argv[2]
        print "Copying non-dupes to " + dest
    else:
        print "Usage: " + sys.argv[0] + " [-n] source [dest]"
        raise SystemExit
    
    oldset = None
    if dest:
        oldset, dupes = collect_all_files(dest)
        if dupes:
            print "Note: there are", len(dupes), "duplicates in destination path."
    
    results, dupes = collect_all_files(source, oldset)
    print "Found", len(results), "unique files and", len(dupes), "duplicates."
    
    if dest:
        print "Copying..."
        copy_files_to(dest, results)
        print "Done."