Skipping Uploader for google drive

uploader.py

__author__ = 'pgcd'
import sys
import os
import subprocess
import hashlib


def hashfile(fname, blocksize=65536):
    hasher = hashlib.md5()
    with open(fname, 'rb') as afile:
        buf = afile.read(blocksize)
        while len(buf) > 0:
            hasher.update(buf)
            buf = afile.read(blocksize)
    return hasher.hexdigest()


def splitData(d):
    return {
        'id': d[:28],
        'title': d[31:74].strip(),
        'size': d[74:85].strip(),
        'date': d[85:],
    }


def getId(folder, parent):
    args = ['drive', 'list', '-n', '-q',
            'title="%s" and \'%s\' in parents' % (os.path.basename(folder).replace('"', '\\"'),
                                                  parent)]
    print " ".join(args)
    result = subprocess.Popen(args, stdout=subprocess.PIPE).stdout.readlines()
    try:
        return splitData(result[0])['id']
    except IndexError:
        return ''


def createFolder(folder, parent=None):
    args = ['drive', 'folder', '-t', os.path.basename(folder), ]
    if parent:
        args += ['-p', parent]
    print " ".join(args)
    result = subprocess.Popen(args, stdout=subprocess.PIPE).stdout.readlines()
    t, sep, _id = result[0].partition(': ')
    if t != 'Id':
        return ''
    return _id.strip()


def getMD5(_id):
    args = ['drive', 'info', '-i', _id]
    result = subprocess.Popen(args, stdout=subprocess.PIPE)
    try:
        t, sep, md5sum = result.stdout.readlines()[6].partition(': ')
    except IndexError:
        return ''
    if t != 'Md5sum':
        return ''
    return md5sum.strip()


def down(folder_name, parent_id):
    folder_ids = {
        os.path.dirname(folder_name): parent_id
    }
    current_id = False
    try:
        with open('processed.log') as file_io:
            previous_process = file_io.readlines()
    except IOError:
        previous_process = []
    with open('processed.log', 'ab') as processed_log:
        for dirName, subdirList, fileList in os.walk(folder_name):
            parent_id = folder_ids.get(os.path.dirname(dirName))
            if parent_id:
                current_id = getId(dirName, parent_id)
            if not current_id:
                current_id = createFolder(dirName, parent_id)
                if not current_id:
                    raise Exception("Unexplicable error creating folder %s under %s" % (dirName, parent_id))
            print('Found local directory: %s, parentId: %s, currentId: %s' % (dirName, parent_id, current_id))
            folder_ids[dirName] = current_id
            for fname in fileList:
                fullname = os.path.join(dirName, fname)
                if fullname + '\n' in previous_process:
                    print "File %s (line %s) marked as processed, skipping" % (
                        fname, previous_process.index(fullname + '\n'))
                    continue
                fileid = getId(fname, current_id)
                if fileid:
                    remote_hash = getMD5(fileid)
                    local_hash = hashfile(fullname)
                    print "Local file %s; id %s; remote md5 %s; local md5 %s" % (fname, fileid, remote_hash,
                                                                                 local_hash)
                    if remote_hash == local_hash:  # Ideally we should resume...
                        print "Same MD5, skipping"
                        processed_log.write(fullname + '\n')
                        continue
                    elif remote_hash:
                        args = ['drive', 'delete', '-i', fileid]
                        print "Deleting non-matching file"
                        subprocess.call(args)
                # In this case we should upload
                args = ['drive', 'upload', '-f', fullname, '-p', current_id]
                print "Uploading local %s" % fullname
                subprocess.call(args)
                processed_log.write(fullname + '\n')


if __name__ == '__main__':
    foldername = sys.argv[1].rstrip('/')
    if len(sys.argv) == 3:  # Upload to parent
        down(foldername, sys.argv[2])
    else:
        down(foldername, 'root')

I'm using this with @prasmussen's gdrive (https://github.com/prasmussen/gdrive) to skip re-uploading the same stuff over and over again when my connection breaks up. It's extremely rough and untested outside what I'm doing right now so YMMV.