-
Notifications
You must be signed in to change notification settings - Fork 1
/
prefect_tasks.py
352 lines (328 loc) · 15.2 KB
/
prefect_tasks.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
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
"""
Each task is nothing but a publisher. These tasks go on publishing the pipeline tasks one by
one in the given order and receive the data back from the workers.
Python specific operations - like updating the pipeline data, schema etc. are the responsibilities of prefect tasks.
Actual tasks can be found under tasks/scripts which can be implemented in any language of choice.
"""
import os
import random
import re
import pandas as pd
import pdfkit
from json2xml import json2xml
from pandas.io.json import build_table_schema
from prefect import task, flow
from task_utils import *
from io import StringIO
@task
def skip_column(context, pipeline, task_obj):
column = context['columns']
col = column
if not isinstance(column, list):
column = list()
column.append(col)
try:
pipeline.data = pipeline.data.drop(column, axis=1)
for col in column:
for sc in pipeline.schema:
if sc['key'] == col:
sc['key'] = ""
sc['format'] = ""
sc['description'] = ""
pipeline.logger.info(f"INFO: skip_column task is done")
set_task_model_values(task_obj, pipeline)
except Exception as e:
send_error_to_prefect_cloud(e)
pipeline.logger.error(f"ERROR: skip_column failed with an error - {str(e)}. Setting "
f"pipeline status to failed")
pipeline.model.err_msg = (f"ERROR: skip_column failed with an error - {str(e)}. Setting "
f"pipeline status to failed")
pipeline.model.save()
task_obj.status = "Failed"
task_obj.save()
@task
def merge_columns(context, pipeline, task_obj):
column1, column2, output_column = context['column1'], context['column2'], context['output_column']
retain_cols = False
separator = context['separator']
try:
retain_cols = context['retain_cols']
except:
pass
try:
pipeline.data[output_column] = pipeline.data[column1].astype(str) + separator + pipeline.data[column2] \
.astype(str)
if not retain_cols:
pipeline.data = pipeline.data.drop([column1, column2], axis=1)
""" setting up the schema after task"""
data_schema = pipeline.data.convert_dtypes(infer_objects=True, convert_string=True,
convert_integer=True, convert_boolean=True, convert_floating=True)
names_types_dict = data_schema.dtypes.astype(str).to_dict()
new_col_format = names_types_dict[output_column]
if not retain_cols:
for sc in pipeline.schema:
if sc['key'] == column1:
sc['key'] = ""
sc['format'] = ""
sc['description'] = ""
if sc['key'] == column2:
sc['key'] = ""
sc['format'] = ""
sc['description'] = ""
pipeline.schema.append({
"key": output_column, "format": new_col_format,"parent": "", "array_field":"", "path":"", "parent_path":"",
"description": "Result of merging columns " + column1 + " & " + column2 + " by pipeline - "
+ pipeline.model.pipeline_name
})
pipeline.logger.info(f"INFO: task - merge_columns is done.")
set_task_model_values(task_obj, pipeline)
except Exception as e:
pipeline.logger.error(f"ERROR: merge_columns failed with an error - {str(e)}. Setting "
f"pipeline status to failed")
pipeline.model.err_msg = (f"ERROR: merge_columns failed with an error - {str(e)}. Setting "
f"pipeline status to failed")
pipeline.model.save()
send_error_to_prefect_cloud(e)
task_obj.status = "Failed"
task_obj.save()
@task
def anonymize(context, pipeline, task_obj):
# TODO - decide on the context contents
option = context['option']
special_char = context['special_char']
col = context['column'].lower()
try:
df_cols = [each.lower() for each in pipeline.data.columns]
col_index = df_cols.index(col)
df_col_values = pipeline.data.iloc[:, col_index].values.tolist() #pipeline.data[col].values.tolist()
new_vals = []
for val in df_col_values:
val = str(val)
if option == "replace_all":
if special_char == "random":
replace_val = "".join(random.choices("!@#$%^&*()<>?{}[]~`", k=len(val)))
new_vals.append(replace_val)
else:
replace_val = special_char * len(val)
new_vals.append(replace_val)
elif option == "replace_nth":
n = context.get('n')
n = int(n) #- 1
if special_char == "random":
replacement = "".join(random.choices("!@#$%^&*()<>?{}[]~`", k=1))
for i in range((n-1), len(val)+(n-1), n):
val = val[ : i] + replacement + val[i + 1: ]
# for i in range(0, len(val), int(n)):
# val = val[ : i] + replacement + val[i + 1: ]
# replace_val = val[0:int(n)] + replacement + val[int(n) + 1:]
new_vals.append(val)
else:
for i in range((n-1), len(val)+(n-1), n):
val = val[ : i] + special_char + val[i + 1: ]
# for i in range(0, len(val), int(n)):
# val = val[ : i] + special_char + val[i + 1: ]
new_vals.append(val)
elif option == "retain_first_n":
n = context.get('n')
if special_char == "random":
replacement = "".join(random.choices("!@#$%^&*()<>?{}[]~`", k=(len(val) - int(n))))
replace_val = val[:int(n)] + replacement
new_vals.append(replace_val)
else:
replace_val = val[:int(n)] + (special_char * (len(val) - int(n)))
new_vals.append(replace_val)
#pipeline.data[col] = new_vals
pipeline.data.iloc[:, col_index] = new_vals
data_schema = pipeline.data.convert_dtypes(infer_objects=True, convert_string=True,
convert_integer=True, convert_boolean=True, convert_floating=True)
names_types_dict = data_schema.dtypes.astype(str).to_dict()
for sc in pipeline.schema:
if sc['key'] == col:
sc['format'] = names_types_dict[col]
pipeline.logger.info(f"INFO: task - anonymize task is done")
set_task_model_values(task_obj, pipeline)
except Exception as e:
pipeline.logger.error(f"ERROR: task - anonymize failed with an error - {str(e)}. Setting "
f"pipeline status to failed")
pipeline.model.err_msg = (f"ERROR: task - anonymize failed with an error - {str(e)}. Setting "
f"pipeline status to failed")
pipeline.model.save()
send_error_to_prefect_cloud(e)
task_obj.status = "Failed"
task_obj.save()
@task
def change_format_to_pdf(context, pipeline, task_obj):
file_format = context['format']
result_file_name = pipeline.model.pipeline_name
dir = "format_changed_files/"
if file_format == "xml" or file_format =="XML":
try:
data_string = pipeline.data.to_json(orient='records')
json_data = json.loads(data_string)
xml_data = json2xml.Json2xml(json_data).to_xml()
with open(dir+result_file_name + '.xml', 'w') as f:
f.write(xml_data)
pipeline.logger.info(f"INFO: Resource format changed to xml")
set_task_model_values(task_obj, pipeline)
except Exception as e:
pipeline.logger.error(f"ERROR: task - change_format failed with an error - {str(e)}. Setting "
f"pipeline status to failed")
pipeline.model.err_msg = (f"ERROR: task - change_format failed with an error - {str(e)}. Setting "
f"pipeline status to failed")
pipeline.model.save()
send_error_to_prefect_cloud(e)
task_obj.status = "Failed"
task_obj.save()
elif file_format == "pdf" or file_format == "PDF":
try:
pipeline.data.to_html("data.html", index=False)
pdfkit.from_file("data.html", dir+result_file_name + ".pdf")
os.remove('data.html')
pipeline.logger.info(f"INFO: Resource format changed to pdf")
set_task_model_values(task_obj, pipeline)
except Exception as e:
pipeline.logger.error(f"ERROR: task - change_format failed with an error - {str(e)}. Setting "
f"pipeline status to failed")
pipeline.model.err_msg = (f"ERROR: task - change_format failed with an error - {str(e)}. Setting "
f"pipeline status to failed")
pipeline.model.save()
send_error_to_prefect_cloud(e)
task_obj.status = "Failed"
task_obj.save()
elif file_format == "json" or file_format == "JSON":
try:
data_string = pipeline.data.to_json(orient='records')
with open(dir + result_file_name + ".json", "w") as f:
f.write(data_string)
pipeline.logger.info(f"INFO: Resource format changed to json")
set_task_model_values(task_obj, pipeline)
except Exception as e:
pipeline.logger.error(f"ERROR: task - change_format failed with an error - {str(e)}. Setting "
f"pipeline status to failed")
pipeline.model.err_msg = (f"ERROR: task - change_format failed with an error - {str(e)}. Setting "
f"pipeline status to failed")
pipeline.model.save()
send_error_to_prefect_cloud(e)
task_obj.status = "Failed"
task_obj.save()
@task
def aggregate(context, pipeline, task_obj):
print("inside aggregate")
index = context['index']
columns = context['columns']
values = context['values']
index = index.split(",")
columns = columns.split(",")
values = values.split(",")
none_list = [None]
for col in columns:
none_list.append(col)
try:
pipeline.data = pd.pivot_table(pipeline.data, index=index, columns=columns, values=values, aggfunc='count')
pipeline.data = pipeline.data.rename_axis(none_list, axis=1)
pipeline.data = pipeline.data.reset_index()
inferred_schema = build_table_schema(pipeline.data)
fields = inferred_schema['fields']
new_schema = []
for field in fields:
key = field['name']
description = ""
format = field['type']
for sc in pipeline.schema:
if sc['key'] == key or sc['key'] == key[0]:
description = sc['description']
if key == "index" or key == "":
continue
if isinstance(key, tuple):
key = "-".join(map(str, key))
new_schema.append({"key": key, "format": format, "description": description,
"parent": "", "array_field":"", "path":"", "parent_path":""})
pipeline.schema = new_schema
pipeline.logger.info(f"INFO: task - aggregate is done.")
set_task_model_values(task_obj, pipeline)
except Exception as e:
pipeline.logger.error(f"ERROR: task - aggregate failed with an error - {str(e)}. Setting "
f"pipeline status to failed")
pipeline.model.err_msg = (f"ERROR: task - aggregate failed with an error - {str(e)}. Setting "
f"pipeline status to failed")
pipeline.model.save()
send_error_to_prefect_cloud(e)
task_obj.status = "Failed"
task_obj.save()
@task
def query_data_resource(context, pipeline, task_obj):
columns = context['columns']
num_rows = context["rows"]
columns = columns.split(",")
if len(columns) == 1 and len(columns[0]) == 0:
column_selected_df = pipeline.data
else:
column_selected_df = pipeline.data.loc[:, pipeline.data.columns.isin(columns)]
for sc in pipeline.schema:
if sc["key"] not in columns:
sc["key"] = ""
sc["format"] = ""
sc["description"] = ""
# if row length is not specified return all rows
if num_rows == "" or int(num_rows) > len(column_selected_df):
final_df = column_selected_df
else:
num_rows_int = int(num_rows)
final_df = column_selected_df.iloc[:num_rows_int]
pipeline.data = final_df
set_task_model_values(task_obj, pipeline)
@task
def fill_missing_fields(context, pipeline, task_obj):
data, exception_flag = publish_task_and_process_result(task_obj, context, pipeline.data)
if not exception_flag:
print(data, "???")
else:
pipeline.logger.error(f"""ERROR: {data} at fill_missing_fields""")
@task
def sample_scraper(context, pipeline, task_obj):
data, exception_flag = publish_task_and_process_result(task_obj, context, pipeline.data)
if not exception_flag:
print("HERE IS S3 LINK------", data)
else:
pipeline.logger.error(f"""ERROR: {data} at sample_scraper""")
@task
def db_loader(context, pipeline, task_obj):
# data = pipeline.data.to_json()
print(pipeline.data)
data, exception_flag = publish_task_and_process_result(task_obj, context, pipeline.data)
if not exception_flag:
print("data loaded in db")
else:
pipeline.logger.error(f"""ERROR: {data} at db_loader""")
@flow
def pipeline_executor(pipeline):
print("setting ", pipeline.model.pipeline_name, " status to In Progress")
pipeline.model.status = "In Progress"
pipeline.logger.info(f"INFO: set pipeline status to - 'In progress'")
print(pipeline.model.status)
pipeline.model.save()
tasks_objects = pipeline._commands
func_names = get_task_names(tasks_objects)
contexts = get_task_contexts(tasks_objects)
try:
for i in range(len(func_names)):
globals()[func_names[i]](contexts[i], pipeline, tasks_objects[i])
except Exception as e:
pipeline.model.err_msg = str(e)
pipeline.model.save()
raise e
# if any of the tasks is failed, set pipeline status as - Failed
for each_task in tasks_objects:
if each_task.status == "Failed":
pipeline.model.status = "Failed"
pipeline.model.save()
pipeline.logger.info(f"INFO: There was a failed task hence, setting pipeline status to failed")
break
# if none of the tasks failed, set pipeline status as - Done
if pipeline.model.status != "Failed":
pipeline.model.status = "Done"
pipeline.model.save()
pipeline.logger.info(f"INFO: All tasks were successful, setting pipeline status to Done")
pipeline.model.output_id = str(pipeline.model.pipeline_id) + "_" + pipeline.model.status
print("Data after pipeline execution\n", pipeline.data)
return