-
Notifications
You must be signed in to change notification settings - Fork 86
/
tableau_rest_xml.py
executable file
·108 lines (100 loc) · 3.89 KB
/
tableau_rest_xml.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
# This is intended to be full of static helper methods
import xml.etree.ElementTree as ET
import random
from typing import Union, Optional, List, Dict, Tuple
import re
class TableauRestXml:
tableau_namespace = 'http://tableau.com/api'
ns_map = {'t': 'http://tableau.com/api'}
ns_prefix = '{' + ns_map['t'] + '}'
# ET.register_namespace('t', ns_map['t'])
# Generic method for XML lists for the "query" actions to name -> id dict
@staticmethod
def convert_xml_list_to_name_id_dict(xml_obj: ET.Element) -> Dict:
d = {}
for element in xml_obj:
e_id = element.get("id")
# If list is collection, have to run one deeper
if e_id is None:
for list_element in element:
e_id = list_element.get("id")
name = list_element.get("name")
d[name] = e_id
else:
name = element.get("name")
d[name] = e_id
return d
# Repeat of above method with shorter name
@staticmethod
def xml_list_to_dict(xml_obj: ET.Element) -> Dict:
d = {}
for element in xml_obj:
e_id = element.get("id")
# If list is collection, have to run one deeper
if e_id is None:
for list_element in element:
e_id = list_element.get("id")
name = list_element.get("name")
d[name] = e_id
else:
name = element.get("name")
d[name] = e_id
return d
@staticmethod
def luid_name_dict_from_xml(xml_obj: ET.Element) -> Dict:
d = {}
for element in xml_obj:
e_id = element.get("id")
# If list is collection, have to run one deeper
if e_id is None:
for list_element in element:
e_id = list_element.get("id")
name = list_element.get("name")
d[e_id] = name
else:
name = element.get("name")
d[e_id] = name
return d
@staticmethod
def luid_content_url_dict_from_xml(xml_obj: ET.Element) -> Dict:
d = {}
for element in xml_obj:
e_id = element.get("id")
# If list is collection, have to run one deeper
if e_id is None:
for list_element in element:
e_id = list_element.get("id")
name = list_element.get("contentUrl")
d[e_id] = name
else:
name = element.get("contentUrl")
d[e_id] = name
return d
# This corrects for the first element in any response by the plural collection tag, which leads to differences
# with the XPath search currently
@staticmethod
def make_xml_list_iterable(xml_obj: ET.Element) -> List[ET.Element]:
pass
# 32 hex characters with 4 dashes
@staticmethod
def is_luid(val: str) -> bool:
luid_pattern = r"[0-9a-fA-F]*-[0-9a-fA-F]*-[0-9a-fA-F]*-[0-9a-fA-F]*-[0-9a-fA-F]*"
if len(val) == 36:
if re.match(luid_pattern, val) is not None:
return True
else:
return False
else:
return False
# URI is different form actual URL you need to load a particular view in iframe
@staticmethod
def convert_view_content_url_to_embed_url(content_url: str) -> str:
split_url = content_url.split('/')
return 'views/{}/{}'.format(split_url[0], split_url[2])
# You must generate a boundary string that is used both in the headers and the generated request that you post.
# This builds a simple 30 hex digit string
@staticmethod
def generate_boundary_string() -> str:
random_digits = [random.SystemRandom().choice('0123456789abcdef') for n in range(30)]
s = "".join(random_digits)
return s