Splitting an objects data_path into hierachal list

Question

Drivers, and in particular driver variable targets have data_paths that will break in the UI using layout.prop(target.id, target.data_path) an example is a mesh target with shape_keys.key_blocks["Key 1"].value as a datapath.

def split_path(data_path):
    '''
    Split a data_path into parts
    '''
    if not len(data_path):
        return []
     # remove all collection names   
    match = re.findall(r'\[\"(.*?)\"\]\.', data_path)

    namedic = {}
    for i, m in enumerate(match):
        key = "Collection___NAME________%d" % i # surely not lol.
        data_path = data_path.replace(m, key, 1)
        namedic[key] = m

    parts = data_path.rpartition(".")
    props = []
    while parts[0] != parts[1] != '':
        dp = parts[0]
        props.append(parts[2])
        parts = dp.rpartition(".")

    props.append(parts[2])

    # replace the names
    propstring = ",".join(props)
    for key, name in namedic.items():
        propstring = propstring.replace(key, name)
    props = propstring.split(",")

    # reverse list
    props.reverse()
    return props

will return ['shape_keys', 'key_blocks["Key 1"]', 'value'] for aforementioned data_path

However it's possible for users to name their objects with quotes " and square brackets et al which will break the re.

Any suggestions on how to parse more effectively for names?

Answer 1

Problems with parsing data_path can occur when the user gives entities names containing apostrophes(names containing quotes will not work in driver data_paths). Such will still work in drivers, i.e:

"""pose.bones["Bone'bla'"].location.x"""

So we can assume every name has to be enclosed with [""] in such datastring that drivers will accept:

def split_path(data_path):
    '''
    Split a data_path into parts
    '''
    if not data_path:
        return []

    # extract names from data_path
    names = data_path.split('"')[1::2]
    data_path_no_names = ''.join(data_path.split('"')[0::2])

    # segment into chunks
    # ID props will be segmented by replacing ][ with ].[
    data_chunks = data_path_no_names.replace('][', '].[').split('.')
    # probably regex should be used here and things put into dictionary
    # so it's clear what chunk is what
    # depends of use case, the main idea is to extract names, segment, then put back

    # put names back into chunks where [] are
    for id, chunk in enumerate(data_chunks):
        if chunk.find('[]') > 0:
            data_chunks[id] = chunk.replace('[]', '["' + names.pop(0) + '"]')

    return data_chunks

The part # segment into chunks is just basic, some more advanced segmenting can be implemented (for example one that will understand location.x is one thing or that will extract correct target.id and target.data_path)