From b3e750be9c2768f16bcfdbf51eea374442f5bf0b Mon Sep 17 00:00:00 2001 From: tierno Date: Wed, 5 Sep 2018 11:25:23 +0200 Subject: [PATCH] new utility deep_update following rfc7396 plus array edition Change-Id: I32afac9dccc1343a9664633d0a11d59c39fd9e69 Signed-off-by: tierno --- osm_common/__init__.py | 4 +- osm_common/dbbase.py | 200 ++++++++++++++++++++++++++++++++ osm_common/tests/test_dbbase.py | 105 ++++++++++++++++- 3 files changed, 305 insertions(+), 4 deletions(-) diff --git a/osm_common/__init__.py b/osm_common/__init__.py index 601b140..882df08 100644 --- a/osm_common/__init__.py +++ b/osm_common/__init__.py @@ -1,2 +1,2 @@ -version = '0.1.6' -date_version = '2018-08-31' +version = '0.1.7' +date_version = '2018-09-07' diff --git a/osm_common/dbbase.py b/osm_common/dbbase.py index 2dada76..3700b3a 100644 --- a/osm_common/dbbase.py +++ b/osm_common/dbbase.py @@ -1,4 +1,6 @@ +import yaml from http import HTTPStatus +from copy import deepcopy __author__ = "Alfonso Tierno " @@ -36,3 +38,201 @@ class DbBase(object): def del_one(self, table, filter={}, fail_on_empty=True): raise DbException("Method 'del_one' not implemented") + + +def deep_update(dict_to_change, dict_reference, key_list=None): + """ + Modifies one dictionary with the information of the other following https://tools.ietf.org/html/rfc7396 + Basically is a recursive python 'dict_to_change.update(dict_reference)', but a value of None is used to delete. + It implements an extra feature that allows modifying an array. RFC7396 only allows replacing the entire array. + For that, dict_reference should contains a dict with keys starting by "$" with the following meaning: + $[index] is an integer for targeting a concrete index from dict_to_change array. If the value is None + the element of the array is deleted, otherwise it is edited. + $+[index] The value is inserted at this . A value of None has not sense and an exception is raised. + $+ The value is appended at the end. A value of None has not sense and an exception is raised. + $val It looks for all the items in the array dict_to_change equal to . is evaluated as yaml, + that is, numbers are taken as type int, true/false as boolean, etc. Use quotes to force string. + Nothing happens if no match is found. If the value is None the matched elements are deleted. + $key: val In case a dictionary is passed in yaml format, if looks for all items in the array dict_to_change + that are dictionaries and contains this equal to . Several keys can be used by yaml + format '{key: val, key: val, ...}'; and all of them mast match. Nothing happens if no match is + found. If value is None the matched items are deleted, otherwise they are edited. + $+val If no match if found (see '$val'), the value is appended to the array. If any match is found nothing + is changed. A value of None has not sense. + $+key: val If no match if found (see '$key: val'), the value is appended to the array. If any match is found + nothing is changed. A value of None has not sense. + If there are several editions, insertions and deletions; editions and deletions are done first in reverse index + order; then insertions also in reverse index order; and finally appends in any order. So indexes used at + insertions must take into account the deleted items. + :param dict_to_change: Target dictionary to be changed. + :param dict_reference: Dictionary that contains changes to be applied. + :param key_list: This is used internally for recursive calls. Do not fill this parameter. + :return: none or raises and exception only at array modification when there is a bad format or conflict. + """ + def _deep_update_array(array_to_change, _dict_reference, _key_list): + to_append = {} + to_insert_at_index = {} + values_to_edit_delete = {} + indexes_to_edit_delete = [] + array_edition = None + _key_list.append("") + for k in _dict_reference: + _key_list[-1] = str(k) + if not isinstance(k, str) or not k.startswith("$"): + if array_edition is True: + raise DbException("Found array edition (keys starting with '$') and pure dictionary edition in the" + " same dict at '{}'".format(":".join(_key_list[:-1]))) + array_edition = False + continue + else: + if array_edition is False: + raise DbException("Found array edition (keys starting with '$') and pure dictionary edition in the" + " same dict at '{}'".format(":".join(_key_list[:-1]))) + array_edition = True + insert = False + indexes = [] # indexes to edit or insert + kitem = k[1:] + if kitem.startswith('+'): + insert = True + kitem = kitem[1:] + if _dict_reference[k] is None: + raise DbException("A value of None has not sense for insertions at '{}'".format( + ":".join(_key_list))) + + if kitem.startswith('[') and kitem.endswith(']'): + try: + index = int(kitem[1:-1]) + if index < 0: + index += len(array_to_change) + if index < 0: + index = 0 # skip outside index edition + indexes.append(index) + except Exception: + raise DbException("Wrong format at '{}'. Expecting integer index inside quotes".format( + ":".join(_key_list))) + elif kitem: + # match_found_skip = False + try: + filter_in = yaml.safe_load(kitem) + except Exception: + raise DbException("Wrong format at '{}'. Expecting '$'".format(":".join(_key_list))) + if isinstance(filter_in, dict): + for index, item in enumerate(array_to_change): + for filter_k, filter_v in filter_in.items(): + if not isinstance(item, dict) or filter_k not in item or item[filter_k] != filter_v: + break + else: # match found + if insert: + # match_found_skip = True + insert = False + break + else: + indexes.append(index) + else: + index = 0 + try: + while True: # if not match a ValueError exception will be raise + index = array_to_change.index(filter_in, index) + if insert: + # match_found_skip = True + insert = False + break + indexes.append(index) + index += 1 + except ValueError: + pass + + # if match_found_skip: + # continue + elif not insert: + raise DbException("Wrong format at '{}'. Expecting '$+', '$[]'".format( + ":".join(_key_list))) + for index in indexes: + if insert: + if index in to_insert_at_index and to_insert_at_index[index] != _dict_reference[k]: + # Several different insertions on the same item of the array + raise DbException("Conflict at '{}'. Several insertions on same array index {}".format( + ":".join(_key_list), index)) + to_insert_at_index[index] = _dict_reference[k] + else: + if index in indexes_to_edit_delete and values_to_edit_delete[index] != _dict_reference[k]: + # Several different editions on the same item of the array + raise DbException("Conflict at '{}'. Several editions on array index {}".format( + ":".join(_key_list), index)) + indexes_to_edit_delete.append(index) + values_to_edit_delete[index] = _dict_reference[k] + if not indexes: + if insert: + to_append[k] = _dict_reference[k] + # elif _dict_reference[k] is not None: + # raise DbException("Not found any match to edit in the array, or wrong format at '{}'".format( + # ":".join(_key_list))) + + # edition/deletion is done before insertion + indexes_to_edit_delete.sort(reverse=True) + for index in indexes_to_edit_delete: + _key_list[-1] = str(index) + try: + if values_to_edit_delete[index] is None: # None->Anything + try: + del (array_to_change[index]) + except IndexError: + pass # it is not consider an error if this index does not exist + elif not isinstance(values_to_edit_delete[index], dict): # NotDict->Anything + array_to_change[index] = deepcopy(values_to_edit_delete[index]) + elif isinstance(array_to_change[index], dict): # Dict->Dict + deep_update(array_to_change[index], values_to_edit_delete[index], _key_list) + else: # Dict->NotDict + if isinstance(array_to_change[index], list): # Dict->List. Check extra array edition + if _deep_update_array(array_to_change[index], values_to_edit_delete[index], _key_list): + continue + array_to_change[index] = deepcopy(values_to_edit_delete[index]) + # calling deep_update to delete the None values + deep_update(array_to_change[index], values_to_edit_delete[index], _key_list) + except IndexError: + raise DbException("Array edition index out of range at '{}'".format(":".join(_key_list))) + + # insertion with indexes + to_insert_indexes = list(to_insert_at_index.keys()) + to_insert_indexes.sort(reverse=True) + for index in to_insert_indexes: + array_to_change.insert(index, to_insert_at_index[index]) + + # append + for k, insert_value in to_append.items(): + _key_list[-1] = str(k) + insert_value_copy = deepcopy(insert_value) + if isinstance(insert_value_copy, dict): + # calling deep_update to delete the None values + deep_update(insert_value_copy, insert_value, _key_list) + array_to_change.append(insert_value_copy) + + _key_list.pop() + if array_edition: + return True + return False + + if key_list is None: + key_list = [] + key_list.append("") + for k in dict_reference: + key_list[-1] = str(k) + if dict_reference[k] is None: # None->Anything + if k in dict_to_change: + del dict_to_change[k] + elif not isinstance(dict_reference[k], dict): # NotDict->Anything + dict_to_change[k] = deepcopy(dict_reference[k]) + elif k not in dict_to_change: # Dict->Empty + dict_to_change[k] = deepcopy(dict_reference[k]) + # calling deep_update to delete the None values + deep_update(dict_to_change[k], dict_reference[k], key_list) + elif isinstance(dict_to_change[k], dict): # Dict->Dict + deep_update(dict_to_change[k], dict_reference[k], key_list) + else: # Dict->NotDict + if isinstance(dict_to_change[k], list): # Dict->List. Check extra array edition + if _deep_update_array(dict_to_change[k], dict_reference[k], key_list): + continue + dict_to_change[k] = deepcopy(dict_reference[k]) + # calling deep_update to delete the None values + deep_update(dict_to_change[k], dict_reference[k], key_list) + key_list.pop() diff --git a/osm_common/tests/test_dbbase.py b/osm_common/tests/test_dbbase.py index c2af52f..2da4506 100644 --- a/osm_common/tests/test_dbbase.py +++ b/osm_common/tests/test_dbbase.py @@ -1,7 +1,7 @@ import http import pytest - -from osm_common.dbbase import DbBase, DbException +import unittest +from osm_common.dbbase import DbBase, DbException, deep_update def exception_message(message): @@ -60,3 +60,104 @@ def test_del_one(db_base): db_base.del_one(None, None, None) assert str(excinfo.value).startswith(exception_message("Method 'del_one' not implemented")) assert excinfo.value.http_code == http.HTTPStatus.NOT_FOUND + + +class TestDeepUpdate(unittest.TestCase): + def test_update_dict(self): + # Original, patch, expected result + TEST = ( + ({"a": "b"}, {"a": "c"}, {"a": "c"}), + ({"a": "b"}, {"b": "c"}, {"a": "b", "b": "c"}), + ({"a": "b"}, {"a": None}, {}), + ({"a": "b", "b": "c"}, {"a": None}, {"b": "c"}), + ({"a": ["b"]}, {"a": "c"}, {"a": "c"}), + ({"a": "c"}, {"a": ["b"]}, {"a": ["b"]}), + ({"a": {"b": "c"}}, {"a": {"b": "d", "c": None}}, {"a": {"b": "d"}}), + ({"a": [{"b": "c"}]}, {"a": [1]}, {"a": [1]}), + ({1: ["a", "b"]}, {1: ["c", "d"]}, {1: ["c", "d"]}), + ({1: {"a": "b"}}, {1: ["c"]}, {1: ["c"]}), + ({1: {"a": "foo"}}, {1: None}, {}), + ({1: {"a": "foo"}}, {1: "bar"}, {1: "bar"}), + ({"e": None}, {"a": 1}, {"e": None, "a": 1}), + ({1: [1, 2]}, {1: {"a": "b", "c": None}}, {1: {"a": "b"}}), + ({}, {"a": {"bb": {"ccc": None}}}, {"a": {"bb": {}}}), + ) + for t in TEST: + deep_update(t[0], t[1]) + self.assertEqual(t[0], t[2]) + # test deepcopy is done. So that original dictionary does not reference the pach + test_original = {1: {"a": "b"}} + test_patch = {1: {"c": {"d": "e"}}} + test_result = {1: {"a": "b", "c": {"d": "e"}}} + deep_update(test_original, test_patch) + self.assertEqual(test_original, test_result) + test_patch[1]["c"]["f"] = "edition of patch, must not modify original" + self.assertEqual(test_original, test_result) + + def test_update_array(self): + # This TEST contains a list with the the Original, patch, and expected result + TEST = ( + # delete all instances of "a"/"d" + ({"A": ["a", "b", "a"]}, {"A": {"$a": None}}, {"A": ["b"]}), + ({"A": ["a", "b", "a"]}, {"A": {"$d": None}}, {"A": ["a", "b", "a"]}), + # delete and insert at 0 + ({"A": ["a", "b", "c"]}, {"A": {"$b": None, "$+[0]": "b"}}, {"A": ["b", "a", "c"]}), + # delete and edit + ({"A": ["a", "b", "a"]}, {"A": {"$a": None, "$[1]": {"c": "d"}}}, {"A": [{"c": "d"}]}), + # insert if not exist + ({"A": ["a", "b", "c"]}, {"A": {"$+b": "b"}}, {"A": ["a", "b", "c"]}), + ({"A": ["a", "b", "c"]}, {"A": {"$+d": "f"}}, {"A": ["a", "b", "c", "f"]}), + # edit by filter + ({"A": ["a", "b", "a"]}, {"A": {"$b": {"c": "d"}}}, {"A": ["a", {"c": "d"}, "a"]}), + ({"A": ["a", "b", "a"]}, {"A": {"$b": None, "$+[0]": "b", "$+": "c"}}, {"A": ["b", "a", "a", "c"]}), + ({"A": ["a", "b", "a"]}, {"A": {"$c": None}}, {"A": ["a", "b", "a"]}), + # index deletion out of range + ({"A": ["a", "b", "a"]}, {"A": {"$[5]": None}}, {"A": ["a", "b", "a"]}), + # nested array->dict + ({"A": ["a", "b", {"id": "1", "c": {"d": 2}}]}, {"A": {"$id: '1'": {"h": None, "c": {"d": "e", "f": "g"}}}}, + {"A": ["a", "b", {"id": "1", "c": {"d": "e", "f": "g"}}]}), + ({"A": [{"id": 1, "c": {"d": 2}}, {"id": 1, "c": {"f": []}}]}, + {"A": {"$id: 1": {"h": None, "c": {"d": "e", "f": "g"}}}}, + {"A": [{"id": 1, "c": {"d": "e", "f": "g"}}, {"id": 1, "c": {"d": "e", "f": "g"}}]}), + # nested array->array + ({"A": ["a", "b", ["a", "b"]]}, {"A": {"$b": None, "$[2]": {"$b": {}, "$+": "c"}}}, + {"A": ["a", ["a", {}, "c"]]}), + # types str and int different, so not found + ({"A": ["a", {"id": "1", "c": "d"}]}, {"A": {"$id: 1": {"c": "e"}}}, {"A": ["a", {"id": "1", "c": "d"}]}), + + ) + for t in TEST: + print(t) + deep_update(t[0], t[1]) + self.assertEqual(t[0], t[2]) + + def test_update_badformat(self): + # This TEST contains original, incorrect patch and #TODO text that must be present + TEST = ( + # conflict, index 0 is edited twice + ({"A": ["a", "b", "a"]}, {"A": {"$a": None, "$[0]": {"c": "d"}}}), + # conflict, two insertions at same index + ({"A": ["a", "b", "a"]}, {"A": {"$[1]": "c", "$[-2]": "d"}}), + ({"A": ["a", "b", "a"]}, {"A": {"$[1]": "c", "$[+1]": "d"}}), + # bad format keys with and without $ + ({"A": ["a", "b", "a"]}, {"A": {"$b": {"c": "d"}, "c": 3}}), + # bad format empty $ and yaml incorrect + ({"A": ["a", "b", "a"]}, {"A": {"$": 3}}), + ({"A": ["a", "b", "a"]}, {"A": {"$a: b: c": 3}}), + ({"A": ["a", "b", "a"]}, {"A": {"$a: b, c: d": 3}}), + # insertion of None + ({"A": ["a", "b", "a"]}, {"A": {"$+": None}}), + # Not found, insertion of None + ({"A": ["a", "b", "a"]}, {"A": {"$+c": None}}), + # index edition out of range + ({"A": ["a", "b", "a"]}, {"A": {"$[5]": 6}}), + # conflict, two editions on index 2 + ({"A": ["a", {"id": "1", "c": "d"}]}, {"A": {"$id: '1'": {"c": "e"}, "$c: d": {"c": "f"}}}), + ) + for t in TEST: + print(t) + self.assertRaises(DbException, deep_update, t[0], t[1]) + try: + deep_update(t[0], t[1]) + except DbException as e: + print(e) -- 2.17.1