PropertyClassification/dags/train_pipeline.py at master · Sardhendu/PropertyClassification · GitHub

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
import os
dir_path  =  os.path.abspath(os.path.join(__file__ ,"../..")) # Moves one level up in the directory

import sys
sys.path.append(dir_path)

from datetime import datetime, timedelta
from jobs.data_load import dump_aerial, dump_aerial_cropped#, dump_overlaid
from jobs.batch_create import prepare_batches, remove_batches
from jobs.train_cv_test import train, cvalid, test
from airflow import DAG
from airflow.operators.python_operator import PythonOperator

from airflow.models import Variable

which_run = Variable.get('which_run')
image_type = Variable.get('image_type')
use_checkpoint_of_run = Variable.get('use_checkpoint_of_run')
filter_conditions = Variable.get('data_filter_conditions')
use_checkpoint = bool(Variable.get('train_using_previous_checkpoints'))
save_checkpoint = bool(Variable.get('save_new_checkpoints'))
write_tensorboard_summary = Variable.get('write_tensorboard_summary')
which_net = str(Variable.get('which_net'))
batch_size = int(Variable.get('batch_size'))
proportion_cv_data = float(Variable.get('proportion_cv_data'))
proportion_test_data = float(Variable.get('proportion_test_data'))


# Parse Variables
cond_dict = {}
filter_conditions = filter_conditions.split('\r\n')
for conds in filter_conditions:
    k, func, v = conds.split(':')
    if func.strip() == 'None':
        cond_dict[k.strip()] = None
    elif func.strip() == 'bool':
        cond_dict[k.strip()] = bool(v.strip())
    elif func.strip() == 'int':
        cond_dict[k.strip()] = int(v.strip())
    elif func.strip() == 'float':
        cond_dict[k.strip()] = float(v.strip())
    elif func.strip() == 'str':
        cond_dict[k.strip()] = str(v.strip())


default_args = {
    'owner': 'Newline Financial',
    'depends_on_past': False,
    'start_date': datetime(2018, 5, 1),
    'email': ['airflow@airflow.com'],
    'email_on_failure': False,
    'email_on_retry': False,
    'retries': 1,
    'retry_delay': timedelta(minutes=2)
}


dag = DAG('PropertyClassification_training_pipeline', default_args=default_args)

#
fetch_aerial = PythonOperator(dag=dag,
                              task_id='fetch_aerial_images',
                              provide_context=True,
                              python_callable=dump_aerial,
                              params=dict(
                                      which_run=which_run,img_type= image_type,
                                      cond_dict=cond_dict)
                              )

fetch_aerial_cropped = PythonOperator(dag=dag,
                                      task_id='fetch_aerial_cropped_images',
                                      provide_context=True,
                                      python_callable=dump_aerial_cropped,
                                      params=dict(
                                              which_run=which_run, img_type=image_type)
                                      )

# fetch_overlaid = PythonOperator(dag=dag,
#                            task_id='fetch_overlaid_images',
#                            provide_context=False,
#                            python_callable=dump_overlaid)

create_batches_train_cv_test = PythonOperator(dag=dag,
                                          task_id='create_aerial_cropped_batches_train_cv_test',
                                          provide_context=True,
                                          python_callable=prepare_batches,
                                          params = dict(
                                                  which_run=which_run,
                                                  img_type=image_type,
                                                  is_cvalid_test=True,
                                                  batch_size=batch_size,
                                                  proportion_cv_data=proportion_cv_data,
                                                  proportion_test_data=proportion_test_data
                                          )
                                          )


train_batches = PythonOperator(dag=dag,
                              task_id='train_images',
                              provide_context=True,
                              python_callable=train,
                              params = dict(
                                      which_run=which_run,
                                      img_type=image_type,
                                      use_checkpoint=use_checkpoint,
                                      save_checkpoint=save_checkpoint,
                                      write_tensorboard_summary=write_tensorboard_summary,
                                      which_net=which_net)
                              )

cross_validate_nw_batch = PythonOperator(dag=dag,
                              task_id='cross_validate_images',
                              provide_context=True,
                              python_callable=cvalid,
                              params = dict(
                                      which_run=which_run,
                                      img_type=image_type,
                                      which_net=which_net)
                              )

test_nw_batch = PythonOperator(dag=dag,
                              task_id='test_images',
                              provide_context=True,
                              python_callable=test,
                              params = dict(
                                      which_run=which_run,
                                      img_type=image_type,
                                      which_net=which_net)
                              )


remove_batches = PythonOperator(dag=dag,
                                task_id='remove_batches',
                                provide_context=True,
                                python_callable=remove_batches,
                                params = dict(
                                        which_run=which_run,
                                        img_type=image_type,
                                        which_net=which_net)
                                )


fetch_aerial >> fetch_aerial_cropped >> create_batches_train_cv_test >> train_batches >> cross_validate_nw_batch >> test_nw_batch >> remove_batches