def thumbnail(event, context):
print("Event: ", event)
print("Event Data: ", event['data'])
bucket = os.path.dirname(event['data']['Key'])
_, file_extension = os.path.splitext(event['data']['Key'])
filename = os.path.basename(event['data']['Key'])
print file_extension.upper()
if file_extension.upper() not in [".JPEG", ".JPG", ".PNG", ".GIF"]:
return "Not a picture"
if event['data'][
'EventName'] == "s3:ObjectCreated:Put" and bucket == 'uploads':
tf = tempfile.NamedTemporaryFile(delete=False)
tf_thumb = tempfile.NamedTemporaryFile(delete=False)
try:
client.fget_object(bucket, filename, tf.name)
except ResponseError as err:
print("Can't get object from the bucket: ", bucket, filename,
tf.name)
print err
size = (120, 120)
img = Image.open(tf.name)
img.thumbnail(size)
img.save(tf_thumb.name, "JPEG")
# puts the thumbnail in a thumbnail bucket
thumb_name = filename + '.thumb'
try:
client.fput_object('thumbs', thumb_name, tf_thumb.name)
except ResponseError as err:
print err
else:
print "Minio file deletion event"
return "Thumbnail creation triggered"
def handler(context):
if context['EventType'] == "s3:ObjectCreated:Put":
tf = tempfile.NamedTemporaryFile(delete=False)
bucket = context['Key'].split('/')[0]
filename = context['Key'].split('/')[1]
# Fetching source file from Minio
try:
print('Fetching file')
client.fget_object(bucket, filename, tf.name)
except ResponseError as err:
print('Error fetching file')
print err
# OCR text extraction performed by Tika
print 'Sending file to Tika'
parsed = parser.from_file(tf.name, 'http://tika-tika-server:80/tika')
ocrdata = json.dumps(parsed, ensure_ascii=True)
# MongoDB document insertion
db = mongo['ocr']
result = db.processed.insert_one(parsed)
print 'Document Saved!'
print('Document proccessed: {0}'.format(result.inserted_id))
# move OCRd file to done bucket
try:
# Copy from input bucket to done bucket
fullpath = 'input/' + filename
client.copy_object('done', filename, fullpath)
# Remove from input bucket
client.remove_object('input', filename)
except ResponseError as err:
print err
else:
print "Minio file deletion event"
return "OCR Finished"
def ocr(context):
if context['EventType'] == "s3:ObjectCreated:Put" and context['Key'].split(
'/')[0] == 'foobar':
tf = tempfile.NamedTemporaryFile(delete=False)
bucket = context['Key'].split('/')[0]
filename = context['Key'].split('/')[1]
try:
client.fget_object(bucket, filename, tf.name)
except ResponseError as err:
print err
# puts the same file in a OCR bucket
ocr_name = filename + '.ocr'
try:
client.fput_object('ocr', ocr_name, tf.name)
except ResponseError as err:
print err
else:
print "Minio file deletion event"
return "OCR called"
def thumbnail(context):
bucket = os.path.dirname(context['Key'])
_, file_extension = os.path.splitext(context['Key'])
filename = os.path.basename(context['Key'])
print file_extension.upper()
if file_extension.upper() != ".JPEG":
return "Not a picture"
if context['EventType'] == "s3:ObjectCreated:Put" and bucket == 'foobar':
tf = tempfile.NamedTemporaryFile(delete=False)
tf_thumb = tempfile.NamedTemporaryFile(delete=False)
try:
client.fget_object(bucket, filename, tf.name)
except ResponseError as err:
print err
size = (120, 120)
img = Image.open(tf.name)
img.thumbnail(size)
img.save(tf_thumb.name, "JPEG")
# puts the thumbnail in a thumbnail bucket
thumb_name = filename + '.thumb'
try:
client.fput_object('thumb', thumb_name, tf_thumb.name)
except ResponseError as err:
print err
else:
print "Minio file deletion event"
return "Thumbnail creation triggered"