def create_snapshots_on_cloud(bucket, project, runner):
import datetime, os
import apache_beam as beam
import hurricanes.goes_to_jpeg as g2j
input_file = 'maria/input/maria.csv'
g2j.copy_togcs('MARIA.csv', bucket, input_file)
OUTPUT_DIR = 'gs://{}/maria/'.format(bucket)
options = {
'staging_location': os.path.join(OUTPUT_DIR, 'tmp', 'staging'),
'temp_location': os.path.join(OUTPUT_DIR, 'tmp'),
'job_name':
'maria-' + datetime.datetime.now().strftime('%y%m%d-%H%M%S'),
'project': project,
'max_num_workers': 12,
'setup_file': './setup.py',
'teardown_policy': 'TEARDOWN_ALWAYS',
'no_save_main_session': True
}
opts = beam.pipeline.PipelineOptions(flags=[], **options)
p = beam.Pipeline(runner, options=opts)
(p
| 'lines' >> beam.io.ReadFromText('gs://{}/{}'.format(bucket, input_file))
| 'parse' >> beam.Map(lambda line: g2j.parse_line(line))
| 'to_jpg' >> beam.Map(lambda (dt, lat, lon): g2j.goes_to_jpeg(
g2j.get_objectId_at(dt), lat, lon, bucket,
'maria/images/ir_{}{:02d}{:02d}{:02d}{:02d}.jpg'.format(
dt.year, dt.month, dt.day, dt.hour, dt.second))))
job = p.run()
if runner == 'DirectRunner':
job.wait_until_finish()
def create_snapshots_on_cloud(bucket, project, runner, opts):
import datetime, os
import apache_beam as beam
import hurricanes.goes_to_jpeg as g2j
query = create_query(opts)
OUTPUT_DIR = 'gs://{}/hurricane/'.format(bucket)
options = {
'staging_location': os.path.join(OUTPUT_DIR, 'tmp', 'staging'),
'temp_location': os.path.join(OUTPUT_DIR, 'tmp'),
'job_name':
'maria-' + datetime.datetime.now().strftime('%y%m%d-%H%M%S'),
'project': project,
'max_num_workers': 12,
'setup_file': './setup.py',
'teardown_policy': 'TEARDOWN_ALWAYS',
'no_save_main_session': True
}
opts = beam.pipeline.PipelineOptions(flags=[], **options)
p = beam.Pipeline(runner, options=opts)
(p
| 'get_tracks' >> beam.io.Read(
beam.io.BigQuerySource(query=query, use_standard_sql=True))
| 'loc_at_time' >>
beam.Map(lambda rowdict:
(g2j.parse_timestamp(rowdict['iso_time']), rowdict['name'].lower(
), rowdict['latitude'], rowdict['longitude']))
| 'to_jpg' >> beam.Map(lambda (dt, name, lat, lon): g2j.goes_to_jpeg(
g2j.get_objectId_at(dt), lat, lon, bucket,
'hurricane/images/{}/ir_{}{:02d}{:02d}{:02d}{:02d}.jpg'.format(
name, dt.year, dt.month, dt.day, dt.hour, dt.second))))
job = p.run()
if runner == 'DirectRunner':
job.wait_until_finish()
def create_snapshots_one_by_one(outdir):
import shutil, os
import hurricanes.goes_to_jpeg as g2j
shutil.rmtree(outdir, ignore_errors=True)
os.mkdir(outdir)
with open('MARIA.csv', 'r') as ifp:
for line in ifp:
dt, lat, lon = g2j.parse_line(line)
objectId = g2j.get_objectId_at(dt)
outfilename = os.path.join(
outdir, 'ir_{}{:02d}{:02d}{:02d}{:02d}.jpg'.format(
dt.year, dt.month, dt.day, dt.hour, dt.second))
jpgfile = g2j.goes_to_jpeg(objectId, lat, lon, None, outfilename)
break # take out this to process all the timestamps ...
def create_local_snapshots(outdir, hurricane_file):
import shutil,os
import hurricanes.goes_to_jpeg as g2j
shutil.rmtree(outdir, ignore_errors=True)
os.mkdir(outdir)
with open(hurricane_file, 'r') as ifp:
for line in ifp:
dt, lat, lon = g2j.parse_line(line)
objectId = g2j.get_objectId_at(dt)
outfilename = os.path.join(
outdir,
'ir_{}{:02d}{:02d}{:02d}{:02d}.jpg'.format(
dt.year, dt.month, dt.day, dt.hour, dt.second))
jpgfile = g2j.goes_to_jpeg(objectId, lat, lon, None, outfilename)
break # take out this to process all the timestamps ...
def create_snapshots_on_cloud(bucket, project, runner, opts):
import datetime, os
import apache_beam as beam
import hurricanes.goes_to_jpeg as g2j
query = create_query(opts)
OUTPUT_DIR = 'gs://{}/hurricane/'.format(bucket)
options = {
'staging_location': os.path.join(OUTPUT_DIR, 'tmp', 'staging'),
'temp_location': os.path.join(OUTPUT_DIR, 'tmp'),
'job_name': 'maria-' + datetime.datetime.now().strftime('%y%m%d-%H%M%S'),
'project': project,
'max_num_workers': 12,
'setup_file': './setup.py',
'teardown_policy': 'TEARDOWN_ALWAYS',
'no_save_main_session': True
}
opts = beam.pipeline.PipelineOptions(flags=[], **options)
p = beam.Pipeline(runner, options=opts)
(p
| 'get_tracks' >> beam.io.Read(beam.io.BigQuerySource(query=query, use_standard_sql=True))
| 'loc_at_time' >> beam.Map(lambda rowdict: (
g2j.parse_timestamp(rowdict['iso_time']),
rowdict['name'].lower(),
rowdict['latitude'],
rowdict['longitude']))
| 'to_jpg' >> beam.Map(lambda (dt,name,lat,lon):
g2j.goes_to_jpeg(
g2j.get_objectId_at(dt),
lat, lon,
bucket,
'hurricane/images/{}/ir_{}{:02d}{:02d}{:02d}{:02d}.jpg'.format(
name, dt.year, dt.month, dt.day, dt.hour, dt.second)))
)
job = p.run()
if runner == 'DirectRunner':
job.wait_until_finish()