1
|
|
|
""" |
2
|
|
|
TableauHyperApiExtraLogic - a Hyper client library. |
3
|
|
|
|
4
|
|
|
This library allows packaging CSV content into HYPER format with data type checks |
5
|
|
|
""" |
6
|
|
|
# package regular expression |
7
|
|
|
import re |
8
|
|
|
# package to handle numerical structures |
9
|
|
|
import numpy |
10
|
|
|
# package to handle Data Frames (in this file) |
11
|
|
|
import pandas as pd |
12
|
|
|
# Custom classes from Tableau Hyper package |
13
|
|
|
from tableauhyperapi import HyperProcess, Telemetry, \ |
14
|
|
|
Connection, CreateMode, \ |
15
|
|
|
NOT_NULLABLE, NULLABLE, SqlType, TableDefinition, \ |
16
|
|
|
Inserter, \ |
17
|
|
|
TableName, \ |
18
|
|
|
HyperException |
19
|
|
|
|
20
|
|
|
|
21
|
|
|
class TableauHyperApiExtraLogic: |
22
|
|
|
|
23
|
|
|
def fn_build_hyper_columns_for_csv(self, logger, timmer, detected_csv_structure): |
24
|
|
|
timmer.start() |
25
|
|
|
list_to_return = [] |
26
|
|
|
for current_field_structure in detected_csv_structure: |
27
|
|
|
list_to_return.append(current_field_structure['order']) |
28
|
|
|
current_column_type = self.fn_convert_to_hyper_types(current_field_structure['type']) |
29
|
|
|
logger.debug('Column ' + str(current_field_structure['order']) + ' having name "' |
30
|
|
|
+ current_field_structure['name'] + '" and type "' |
31
|
|
|
+ current_field_structure['type'] + '" will become "' |
32
|
|
|
+ str(current_column_type) + '"') |
33
|
|
|
nullability_value = NULLABLE |
34
|
|
|
if current_field_structure['nulls'] == 0: |
35
|
|
|
nullability_value = NOT_NULLABLE |
36
|
|
|
list_to_return[current_field_structure['order']] = TableDefinition.Column( |
37
|
|
|
name=current_field_structure['name'], |
38
|
|
|
type=current_column_type, |
39
|
|
|
nullability=nullability_value |
40
|
|
|
) |
41
|
|
|
logger.info('Building Hyper columns completed') |
42
|
|
|
timmer.stop() |
43
|
|
|
return list_to_return |
44
|
|
|
|
45
|
|
|
@staticmethod |
46
|
|
|
def fn_convert_to_hyper_types(given_type): |
47
|
|
|
switcher = { |
48
|
|
|
'empty': SqlType.text(), |
49
|
|
|
'bool': SqlType.bool(), |
50
|
|
|
'int': SqlType.big_int(), |
51
|
|
|
'float-dot': SqlType.double(), |
52
|
|
|
'date-YMD': SqlType.date(), |
53
|
|
|
'date-MDY': SqlType.date(), |
54
|
|
|
'date-DMY': SqlType.date(), |
55
|
|
|
'time-24': SqlType.time(), |
56
|
|
|
'time-12': SqlType.time(), |
57
|
|
|
'datetime-24-YMD': SqlType.timestamp(), |
58
|
|
|
'datetime-12-MDY': SqlType.timestamp(), |
59
|
|
|
'datetime-24-DMY': SqlType.timestamp(), |
60
|
|
|
'str': SqlType.text() |
61
|
|
|
} |
62
|
|
|
identified_type = switcher.get(given_type) |
63
|
|
|
if identified_type is None: |
64
|
|
|
identified_type = SqlType.text() |
65
|
|
|
return identified_type |
66
|
|
|
|
67
|
|
|
def fn_create_hyper_file_from_csv(self, local_logger, timmer, input_csv_data_frame, |
68
|
|
|
in_data_type, given_parameters): |
69
|
|
|
hyper_cols = self.fn_build_hyper_columns_for_csv(local_logger, timmer, in_data_type) |
70
|
|
|
# The rows to insert into the <hyper_table> table. |
71
|
|
|
data_to_insert = self.fn_rebuild_csv_content_for_hyper(local_logger, timmer, |
72
|
|
|
input_csv_data_frame, |
73
|
|
|
in_data_type) |
74
|
|
|
# Starts the Hyper Process with telemetry enabled/disabled to send data to Tableau or not |
75
|
|
|
# To opt in, simply set telemetry=Telemetry.SEND_USAGE_DATA_TO_TABLEAU. |
76
|
|
|
# To opt out, simply set telemetry=Telemetry.DO_NOT_SEND_USAGE_DATA_TO_TABLEAU. |
77
|
|
|
with HyperProcess(telemetry=Telemetry.DO_NOT_SEND_USAGE_DATA_TO_TABLEAU) as hyper: |
78
|
|
|
# Creates new Hyper file <output_hyper_file> |
79
|
|
|
# Replaces file with CreateMode.CREATE_AND_REPLACE if it already exists. |
80
|
|
|
timmer.start() |
81
|
|
|
with Connection(endpoint=hyper.endpoint, |
82
|
|
|
database=given_parameters.output_file, |
83
|
|
|
create_mode=CreateMode.CREATE_AND_REPLACE) as hyper_connection: |
84
|
|
|
local_logger.info('Connection to the Hyper engine ' |
85
|
|
|
+ f'file "{given_parameters.output_file}" has been created.') |
86
|
|
|
timmer.stop() |
87
|
|
|
timmer.start() |
88
|
|
|
hyper_connection.catalog.create_schema("Extract") |
89
|
|
|
local_logger.info('Hyper schema "Extract" has been created.') |
90
|
|
|
hyper_table = TableDefinition( |
91
|
|
|
TableName("Extract", "Extract"), |
92
|
|
|
columns=hyper_cols |
93
|
|
|
) |
94
|
|
|
hyper_connection.catalog.create_table(table_definition=hyper_table) |
95
|
|
|
local_logger.info('Hyper table "Extract" has been created.') |
96
|
|
|
timmer.stop() |
97
|
|
|
timmer.start() |
98
|
|
|
# Execute the actual insert |
99
|
|
|
with Inserter(hyper_connection, hyper_table) as hyper_insert: |
100
|
|
|
hyper_insert.add_rows(rows=data_to_insert) |
101
|
|
|
hyper_insert.execute() |
102
|
|
|
local_logger.info('Data has been inserted into Hyper table') |
103
|
|
|
timmer.stop() |
104
|
|
|
timmer.start() |
105
|
|
|
# Number of rows in the <hyper_table> table. |
106
|
|
|
# `execute_scalar_query` is for executing a query |
107
|
|
|
# that returns exactly one row with one column. |
108
|
|
|
query_to_run = f'SELECT COUNT(*) FROM {hyper_table.table_name}' |
109
|
|
|
row_count = hyper_connection.execute_scalar_query(query=query_to_run) |
110
|
|
|
local_logger.info(f'Table {hyper_table.table_name} has {row_count} rows') |
111
|
|
|
timmer.stop() |
112
|
|
|
local_logger.info('Connection to the Hyper engine file has been closed') |
113
|
|
|
local_logger.info('Hyper engine process has been shut down') |
114
|
|
|
|
115
|
|
|
def fn_rebuild_csv_content_for_hyper(self, logger, timmer, input_df, detected_fields_type): |
116
|
|
|
timmer.start() |
117
|
|
|
input_df.replace(to_replace=[numpy.nan], value=[None], inplace=True) |
118
|
|
|
# Cycle through all found columns |
119
|
|
|
for current_field in detected_fields_type: |
120
|
|
|
fld_nm = current_field['name'] |
121
|
|
|
logger.debug(f'Column {fld_nm} has panda_type = ' + str(current_field['panda_type']) |
122
|
|
|
+ ' and python type = ' + str(current_field['type'])) |
123
|
|
|
input_df[fld_nm] = self.reevaluate_single_column(input_df, fld_nm, current_field) |
124
|
|
|
logger.info('Re-building CSV content for maximum Hyper compatibility has been completed') |
125
|
|
|
timmer.stop() |
126
|
|
|
return input_df.values |
127
|
|
|
|
128
|
|
|
def reevaluate_single_column(self, given_df, given_field_name, current_field_details): |
129
|
|
|
if current_field_details['type'] == 'str': |
130
|
|
|
given_df[given_field_name] = given_df[given_field_name].astype(str) |
131
|
|
|
elif current_field_details['panda_type'] == 'float64' \ |
132
|
|
|
and current_field_details['type'] == 'int': |
133
|
|
|
given_df[given_field_name] = given_df[given_field_name].fillna(0).astype('int64') |
134
|
|
|
elif current_field_details['type'][0:5] in ('date-', 'datet', 'time-'): |
135
|
|
|
given_df[given_field_name] = self.fn_string_to_date(given_field_name, given_df) |
136
|
|
|
return given_df[given_field_name] |
137
|
|
|
|
138
|
|
|
def fn_run_hyper_creation(self, local_logger, timmer, input_data_frame, input_data_type, |
139
|
|
|
given_parameters): |
140
|
|
|
try: |
141
|
|
|
self.fn_create_hyper_file_from_csv(local_logger, timmer, input_data_frame, |
142
|
|
|
input_data_type, given_parameters) |
143
|
|
|
except HyperException as ex: |
144
|
|
|
local_logger.error(str(ex).replace(chr(10), ' ')) |
145
|
|
|
exit(1) |
146
|
|
|
|
147
|
|
|
@staticmethod |
148
|
|
|
def fn_string_to_date(in_col_name, in_data_frame): |
149
|
|
|
if re.match('-YMD', in_col_name): |
150
|
|
|
in_data_frame[in_col_name] = pd.to_datetime(in_data_frame[in_col_name], yearfirst=True) |
151
|
|
|
elif re.match('-DMY', in_col_name): |
152
|
|
|
in_data_frame[in_col_name] = pd.to_datetime(in_data_frame[in_col_name], dayfirst=True) |
153
|
|
|
else: |
154
|
|
|
in_data_frame[in_col_name] = pd.to_datetime(in_data_frame[in_col_name]) |
155
|
|
|
return in_data_frame[in_col_name] |
156
|
|
|
|