Organize a list of string in a structured way (dictionary, list structure)

string_structurize.py

"""
Organize a list of string in a structured way (very slow)
:Author: chaonan99
:Date: 2016/10/21
"""
import re
import numpy as np

class StructureDir(object):
	""" StructureDir
		:type pattern: `_sre.SRE_Pattern` (returned by re.compile)
		:param pattern: a regular expression object to match each string in the list
		:type format_list: `list` of `string`
		:value format_list: "list" or "dict"
		:param format_list: indict a group should be organized as dictionary or list
	"""
	def __init__(self, pattern, format_list):
		super(StructureDir, self).__init__()
		self.pattern = pattern
		self.format_list = format_list
		assert pattern.groups == len(format_list), "length of pattern groups and format list do not match!"

	def structured(self, list_of_dir):
		def f(x):
			return self.pattern.findall(x)[0]

		matches = tuple(map(f, list_of_dir))
		return self.__merge_one_layer(dict(zip(matches, list_of_dir)))

	def __merge_one_layer(self, current_merge):
		matches = list(current_merge.keys())
		files = list(current_merge.values())
		n = len(matches[0]) - 1

		if n == -1:
			return current_merge
		if n == 0:
			if self.format_list[n] == 'list':
				return files
			elif self.format_list[n] == 'dict':
				return dict(zip(matches, files))
			else:
				raise ValueError("Unexpected format indicator: {}".format, self.format_list[n])
		else:
			if self.format_list[n] == 'list':
				next_merge = {x:[files[ind] for ind, y in enumerate(np.array(matches)) if np.all(y[0:n]==np.array(x))] \
				 for x in {tuple(row) for row in np.array(matches)[:, 0:n]}}
			elif self.format_list[n] == 'dict':
				next_merge = {x:{y[n]:files[ind] for ind, y in enumerate(np.array(matches)) if np.all(y[0:n]==np.array(x))} \
				 for x in {tuple(row) for row in np.array(matches)[:, 0:n]}}
			else:
				raise ValueError("Unexpected format indicator: {}".format, self.format_list[n])

		return self.__merge_one_layer(next_merge)


if __name__ == '__main__':
	"""example"""
	pattern = re.compile(r'/data/datasets/MARS/bbox_test/\d+/(\d+)C(\d+)T(\d+)F(\d+)')
	files = [
		"/data/datasets/MARS/bbox_test/0001/0001C2T001F0001",
		"/data/datasets/MARS/bbox_test/0001/0001C2T001F0002",
		"/data/datasets/MARS/bbox_test/0001/0001C2T001F0003",
		"/data/datasets/MARS/bbox_test/0001/0001C2T002F0001",
		"/data/datasets/MARS/bbox_test/0001/0001C2T002F0002",
		"/data/datasets/MARS/bbox_test/0001/0001C2T002F0003",
		"/data/datasets/MARS/bbox_test/0001/0001C4T003F0001",
		"/data/datasets/MARS/bbox_test/0001/0001C4T003F0002",
		"/data/datasets/MARS/bbox_test/0001/0002C1T001F0001",
		"/data/datasets/MARS/bbox_test/0001/0002C1T001F0002",
		"/data/datasets/MARS/bbox_test/0001/0002C1T002F0001",
		"/data/datasets/MARS/bbox_test/0001/0002C2T003F0001",
		"/data/datasets/MARS/bbox_test/0001/0002C2T003F0002",
		]
	stc = StructureDir(pattern, ["dict", "dict", "list", "list"])
	res = stc.structured(files)
	from IPython import embed; embed()