def predict(filePath):
project_id = "762814515048"
model_id = "ICN244084984296505344"
file_path = filePath
prediction_client = automl.PredictionServiceClient()
# Get the full path of the model.
model_full_id = automl.AutoMlClient.model_path(project_id, "us-central1",
model_id)
# Read the file.
with open(file_path, "rb") as content_file:
content = content_file.read()
image = automl.Image(image_bytes=content)
payload = automl.ExamplePayload(image=image)
# params is additional domain-specific parameters.
# score_threshold is used to filter the result
# https://cloud.google.com/automl/docs/reference/rpc/google.cloud.automl.v1#predictrequest
params = {"score_threshold": "0.7"}
request = automl.PredictRequest(name=model_full_id,
payload=payload,
params=params)
response = prediction_client.predict(request=request)
print("Prediction results:")
results = []
for result in response.payload:
results.append(result.display_name)
results.append(result.classification.score)
print("Predicted class name: {}".format(result.display_name))
print("Predicted class score: {}".format(result.classification.score))
return (results)
def get_results_from_automl(filepath):
project_id = "numeric-polygon-283403"
model_id = "ICN2007288218877165568"
prediction_client = automl.PredictionServiceClient()
# Get the full path of the model.
model_full_id = automl.AutoMlClient.model_path(project_id, "us-central1",
model_id)
# Read the file.
with open(filepath, "rb") as content_file:
content = content_file.read()
image = automl.Image(image_bytes=content)
payload = automl.ExamplePayload(image=image)
# params is additional domain-specific parameters, score_threshold is used to filter the result
# https://cloud.google.com/automl/docs/reference/rpc/google.cloud.automl.v1#predictrequest
params = {"score_threshold": "0.5"}
request = automl.PredictRequest(name=model_full_id,
payload=payload,
params=params)
response = prediction_client.predict(request=request)
results = []
for result in response.payload:
results.append([result.display_name, result.classification.score])
return results
def isWearingMask(file_path):
prediction_client = automl.PredictionServiceClient()
# Get the full path of the model.
model_full_id = automl.AutoMlClient.model_path(project_id, "us-central1",
model_id)
# Read the file.
with open(file_path, "rb") as content_file:
content = content_file.read()
image = automl.Image(image_bytes=content)
payload = automl.ExamplePayload(image=image)
# params is additional domain-specific parameters.
# score_threshold is used to filter the result
# https://cloud.google.com/automl/docs/reference/rpc/google.cloud.automl.v1#predictrequest
params = {"score_threshold": "0.1"}
request = automl.PredictRequest(name=model_full_id,
payload=payload,
params=params)
response = prediction_client.predict(request=request)
#print("Prediction results:")
print(len(response.payload))
for result in response.payload:
print("Predicted class name: {}".format(result.display_name))
print("Predicted class score: {}".format(result.classification.score))
return True if (result.display_name == "with_mask"
and result.classification.score > 0.9999) else False
return False
def gesture_detection(file_path):
project_id = 'gesture-detection-265519'
model_id = 'ICN7380709295957999616'
prediction_client = automl.PredictionServiceClient()
# Get the full path of the model.
model_full_id = prediction_client.model_path(project_id, 'us-central1',
model_id)
# Read the file.
with open(file_path, 'rb') as content_file:
content = content_file.read()
image = automl.types.Image(image_bytes=content)
payload = automl.types.ExamplePayload(image=image)
# params is additional domain-specific parameters.
# score_threshold is used to filter the result
params = {'score_threshold': '0.5'}
response = prediction_client.predict(model_full_id, payload, params)
display_dict = {
"ONE": 1,
"TWO": 2,
"THREE": 3,
"FOUR": 4,
"FIVE": 5,
"NONE": 0
}
for result in response.payload:
return display_dict[result.display_name]
return 0
def predict(project_id, model_id, file_path):
"""Predict."""
# [START automl_vision_classification_predict]
from google.cloud import automl
# TODO(developer): Uncomment and set the following variables
# project_id = "YOUR_PROJECT_ID"
# model_id = "YOUR_MODEL_ID"
# file_path = "path_to_local_file.jpg"
prediction_client = automl.PredictionServiceClient()
# Get the full path of the model.
model_full_id = prediction_client.model_path(project_id, "us-central1",
model_id)
# Read the file.
with open(file_path, "rb") as content_file:
content = content_file.read()
image = automl.types.Image(image_bytes=content)
payload = automl.types.ExamplePayload(image=image)
# params is additional domain-specific parameters.
# score_threshold is used to filter the result
params = {"score_threshold": "0.8"}
response = prediction_client.predict(model_full_id, payload, params)
print("Prediction results:")
for result in response.payload:
print("Predicted class name: {}".format(result.display_name))
print("Predicted class score: {}".format(result.classification.score))
def predict(project_id, model_id, file_path):
"""Predict."""
# [START automl_translate_predict]
from google.cloud import automl
# TODO(developer): Uncomment and set the following variables
# project_id = "YOUR_PROJECT_ID"
# model_id = "YOUR_MODEL_ID"
# file_path = "path_to_local_file.txt"
prediction_client = automl.PredictionServiceClient()
# Get the full path of the model.
model_full_id = prediction_client.model_path(project_id, "us-central1",
model_id)
# Read the file content for translation.
with open(file_path, "rb") as content_file:
content = content_file.read()
content.decode("utf-8")
text_snippet = automl.types.TextSnippet(content=content)
payload = automl.types.ExamplePayload(text_snippet=text_snippet)
response = prediction_client.predict(model_full_id, payload)
translated_content = response.payload[0].translation.translated_content
print(u"Translated content: {}".format(translated_content.content))
def get_Google_Prediction():
prediction_client = automl.PredictionServiceClient()
# Get the full path of the model.
model_full_id = prediction_client.model_path(project_id, "us-central1",
model_id)
# Read the file.
with open(file_path, "rb") as content_file:
content = content_file.read()
image = automl.types.Image(image_bytes=content)
payload = automl.types.ExamplePayload(image=image)
# params is additional domain-specific parameters.
# score_threshold is used to filter the result
# https://cloud.google.com/automl/docs/reference/rpc/google.cloud.automl.v1#predictrequest
params = {"score_threshold": "0.1"}
response = prediction_client.predict(model_full_id, payload, params)
# no analyze the results
google_result = response.payload
#google_response = getResponse(google_result.display_name, google_result.classification.score)
return google_result
def batch_predict(project_id, model_id, input_uri, output_uri):
"""Batch predict"""
# [START automl_batch_predict]
from google.cloud import automl
# TODO(developer): Uncomment and set the following variables
# project_id = "YOUR_PROJECT_ID"
# model_id = "YOUR_MODEL_ID"
# input_uri = "gs://YOUR_BUCKET_ID/path/to/your/input/csv_or_jsonl"
# output_uri = "gs://YOUR_BUCKET_ID/path/to/save/results/"
prediction_client = automl.PredictionServiceClient()
# Get the full path of the model.
model_full_id = f"projects/{project_id}/locations/us-central1/models/{model_id}"
gcs_source = automl.GcsSource(input_uris=[input_uri])
input_config = automl.BatchPredictInputConfig(gcs_source=gcs_source)
gcs_destination = automl.GcsDestination(output_uri_prefix=output_uri)
output_config = automl.BatchPredictOutputConfig(
gcs_destination=gcs_destination)
response = prediction_client.batch_predict(name=model_full_id,
input_config=input_config,
output_config=output_config)
print("Waiting for operation to complete...")
print(
f"Batch Prediction results saved to Cloud Storage bucket. {response.result()}"
)
def __init__(self, MODEL_CONFIG):
# The projectID in Google Cloud
self.project_id = MODEL_CONFIG['project_id']
# The modelID of the model in the natural language section of GCP
self.model_id = MODEL_CONFIG['model_id']
# Creating a client to receive predictions
self.client = automl.PredictionServiceClient()
def predict(project_id, model_id, content):
"""Predict."""
# [START automl_language_text_classification_predict]
from google.cloud import automl
# TODO(developer): Uncomment and set the following variables
# project_id = "YOUR_PROJECT_ID"
# model_id = "YOUR_MODEL_ID"
# content = "text to predict"
prediction_client = automl.PredictionServiceClient()
# Get the full path of the model.
model_full_id = prediction_client.model_path(
project_id, "us-central1", model_id
)
# Supported mime_types: 'text/plain', 'text/html'
# https://cloud.google.com/automl/docs/reference/rpc/google.cloud.automl.v1#textsnippet
text_snippet = automl.types.TextSnippet(
content=content, mime_type="text/plain"
)
payload = automl.types.ExamplePayload(text_snippet=text_snippet)
response = prediction_client.predict(model_full_id, payload)
for annotation_payload in response.payload:
print(
u"Predicted class name: {}".format(annotation_payload.display_name)
)
print(
u"Predicted class score: {}".format(
annotation_payload.classification.score
)
)
def get_prediction(content, project_id, model_id):
prediction_client = automl.PredictionServiceClient()
name = 'projects/{}/locations/us-central1/models/{}'.format(
project_id, model_id)
payload = {'image': {'image_bytes': content}}
params = {}
request = prediction_client.predict(name, payload, params)
return request # waits till request is returned
def test_default_credentials_prediction_client(self):
env = EnvironmentVarGuard()
env.set('KAGGLE_USER_SECRETS_TOKEN', 'foobar')
env.set('KAGGLE_KERNEL_INTEGRATIONS', 'CLOUDAI')
with env:
prediction_client = automl.PredictionServiceClient()
self.assertIsNotNone(prediction_client.credentials)
self.assertIsInstance(prediction_client.credentials, KaggleKernelCredentials)
self.assertTrue(prediction_client._connection.user_agent.startswith("kaggle-gcp-client/1.0"))
def automl_prediction_service_batch_predict(
model_path,
gcs_input_uris: str = None,
gcs_output_uri_prefix: str = None,
bq_input_uri: str = None,
bq_output_uri: str = None,
params=None,
retry=None, #google.api_core.gapic_v1.method.DEFAULT,
timeout=None, #google.api_core.gapic_v1.method.DEFAULT,
metadata: dict = None,
) -> NamedTuple('Outputs', [('gcs_output_directory', str),
('bigquery_output_dataset', str)]):
import sys
import subprocess
subprocess.run([
sys.executable, '-m', 'pip', 'install', 'google-cloud-automl==0.4.0',
'--quiet', '--no-warn-script-location'
],
env={'PIP_DISABLE_PIP_VERSION_CHECK': '1'},
check=True)
input_config = {}
if gcs_input_uris:
input_config['gcs_source'] = {'input_uris': gcs_input_uris}
if bq_input_uri:
input_config['bigquery_source'] = {'input_uri': bq_input_uri}
output_config = {}
if gcs_output_uri_prefix:
output_config['gcs_destination'] = {
'output_uri_prefix': gcs_output_uri_prefix
}
if bq_output_uri:
output_config['bigquery_destination'] = {'output_uri': bq_output_uri}
from google.cloud import automl
client = automl.PredictionServiceClient()
response = client.batch_predict(
model_path,
input_config,
output_config,
params,
retry,
timeout,
metadata,
)
print('Operation started:')
print(response.operation)
result = response.result()
metadata = response.metadata
print('Operation finished:')
print(metadata)
output_info = metadata.batch_predict_details.output_info
# Workaround for Argo issue - it fails when output is empty: https://github.com/argoproj/argo/pull/1277/files#r326028422
return (output_info.gcs_output_directory
or '-', output_info.bigquery_output_dataset or '-')
def __init__(self, model_name=None, prediction_client=None):
self.config = None
# The model name
# e.g. a value like value like
# "projects/976279526634/locations/us-central1/models/TCN654213816573231104'"
self.model_name = model_name
if not prediction_client:
prediction_client = automl.PredictionServiceClient()
self._prediction_client = prediction_client
def predict(content):
options = ClientOptions(api_endpoint="automl.googleapis.com")
prediction_client = automl.PredictionServiceClient(client_options=options)
payload = {
"text_snippet": {
"content": content,
"mime_type": "text/plain"
}
}
request = prediction_client.predict(name=model_name, payload=payload)
data = {}
for item in request.payload:
data[item.display_name] = item.classification.score
return data
def func():
prediction_client = automl.PredictionServiceClient()
# Get the full path of the model.
model_full_id = prediction_client.model_path(project_id, "us-central1",
model_id)
# Read the file.
with open(file_path, "rb") as content_file:
content = content_file.read()
image = automl.types.Image(image_bytes=content)
payload = automl.types.ExamplePayload(image=image)
return get_prediction(content, project_id, model_id)
def predict(project_id, model_id, file_path):
"""Predict."""
# [START automl_vision_object_detection_predict]
from google.cloud import automl
# TODO(developer): Uncomment and set the following variables
# project_id = "YOUR_PROJECT_ID"
# model_id = "YOUR_MODEL_ID"
# file_path = "path_to_local_file.jpg"
prediction_client = automl.PredictionServiceClient()
# Get the full path of the model.
model_full_id = prediction_client.model_path(
project_id, "us-central1", model_id
)
# Read the file.
with open(file_path, "rb") as content_file:
content = content_file.read()
image = automl.types.Image(image_bytes=content)
payload = automl.types.ExamplePayload(image=image)
# params is additional domain-specific parameters.
# score_threshold is used to filter the result
# https://cloud.google.com/automl/docs/reference/rpc/google.cloud.automl.v1#predictrequest
params = {"score_threshold": "0.8"}
response = prediction_client.predict(model_full_id, payload, params)
print("Prediction results:")
for result in response.payload:
print("Predicted class name: {}".format(result.display_name))
print(
"Predicted class score: {}".format(
result.image_object_detection.score
)
)
bounding_box = result.image_object_detection.bounding_box
print("Normalized Vertices:")
for vertex in bounding_box.normalized_vertices:
print("\tX: {}, Y: {}".format(vertex.x, vertex.y))
def predict(content):
prediction_client = automl.PredictionServiceClient()
# Get the full path of the model.
model_full_id = automl.AutoMlClient.model_path(project_id, "us-central1",
model_id)
# Supported mime_types: 'text/plain', 'text/html'
# https://cloud.google.com/automl/docs/reference/rpc/google.cloud.automl.v1#textsnippet
text_snippet = automl.TextSnippet(content=content, mime_type="text/plain")
payload = automl.ExamplePayload(text_snippet=text_snippet)
response = prediction_client.predict(name=model_full_id, payload=payload)
for annotation_payload in response.payload:
print("Predicted class name: {}".format(
annotation_payload.display_name))
print("Predicted sentiment score: {}".format(
annotation_payload.text_sentiment.sentiment))
return annotation_payload.text_sentiment.sentiment
def predict(self, text):
"""
Receives a text input
Outputs a payload with the prediction of the text sentiment
"""
try:
# All below is given code from GCP
prediction_client = automl.PredictionServiceClient()
model_full_id = prediction_client.model_path(
self.project_id, 'us-central1', self.model_id)
text_snippet = automl.types.TextSnippet(
content=text,
mime_type='text/plain') # Types: 'text/plain', 'text/html'
payload = automl.types.ExamplePayload(text_snippet=text_snippet)
response = prediction_client.predict(model_full_id, payload)
return response.payload
except Exception as error:
print(error)
def predict(content):
from google.cloud import automl
# You must first create a dataset, using the `eu` endpoint, before you can
# call other operations such as: list, get, import, delete, etc.
client_options = {'api_endpoint': 'eu-automl.googleapis.com:443'}
project_id = '685330484131'
# (model AI Crowd) model_id = 'TCN3300918624537018368'
model_id = 'TCN4629621252099670016'
prediction_client = automl.PredictionServiceClient(
client_options=client_options)
# Get the full path of the model.
model_full_id = automl.AutoMlClient.model_path(project_id, "eu", model_id)
text_snippet = automl.TextSnippet(content=content, mime_type="text/plain")
payload = automl.ExamplePayload(text_snippet=text_snippet)
response = prediction_client.predict(name=model_full_id, payload=payload)
return response
def get_prediction(file_path):
project_id = "msds-434-final"
model_id = "ICN7344581542892535808"
# file_path = "uploads/city.png" ## for local image testing
credentials = service_account.Credentials.from_service_account_file("/home/jesse_lybianto/msds-434-final/msds-434-final-f32a0ccc78d5.json")
prediction_client = automl.PredictionServiceClient(credentials=credentials)
# Get the full path of the model.
model_full_id = automl.AutoMlClient.model_path(
project_id, "us-central1", model_id
)
# Read the file.
with open(file_path, "rb") as content_file:
content = content_file.read()
image = automl.Image(image_bytes=content)
payload = automl.ExamplePayload(image=image)
# params is additional domain-specific parameters.
# score_threshold is used to filter the result
# https://cloud.google.com/automl/docs/reference/rpc/google.cloud.automl.v1#predictrequest
params = {"score_threshold": "0.0"}
request = automl.PredictRequest(
name=model_full_id,
payload=payload,
params=params
)
response = prediction_client.predict(request=request)
print(response.payload)
return response.payload
# For non-modular console output
# print("Prediction results:")
# for result in response.payload:
# print("Predicted class name: {}".format(result.display_name))
# print("Predicted class score: {}".format(result.classification.score))
def get_disaster(bucket, filename):
project_id = "swamphacksvi-266915"
model_id = "ICN6150171066523189248"
prediction_client = automl.PredictionServiceClient()
model_full_id = prediction_client.model_path(project_id, "us-central1",
model_id)
bucket = storage_client.bucket(bucket)
blob = bucket.blob(filename)
image = automl.types.Image(image_bytes=blob.download_as_string())
payload = automl.types.ExamplePayload(image=image)
params = {"score_threshold": "0.8"}
response = prediction_client.predict(model_full_id, payload, params)
disasters = [result.display_name for result in response.payload]
print("Prediction Results: ", disasters)
doc_id = filename.split('.')[0]
doc = disaster_docs.document(doc_id)
doc.update({'disaster': disasters})
print(f'After update: {doc.get().to_dict()}')
def home_page():
# from flask import request
if request.method == 'GET':
context = {"done": False}
return render_template('site.html', **context)
else:
# get image content from POST request inputs
file = request.files["myFile"]
image_content = b''
for data in file.stream: # read image as a stream
image_content += data
# google api variables
project_id = "alpine-infinity-290102"
prediction_client = automl.PredictionServiceClient()
# Get the full path of the model.
model_full_id = automl.AutoMlClient.model_path(project_id,
"us-central1", model_id)
image = automl.Image(image_bytes=image_content)
payload = automl.ExamplePayload(image=image)
req = automl.PredictRequest(
name=model_full_id,
payload=payload,
)
response = prediction_client.predict(request=req)
class_ = response.payload[0].display_name
measure = response.payload[0].classification.score
context = {
"done": True,
"guess_mal": class_ == "Malignant",
"class_": class_,
"measure": '{:.2%}'.format(measure)
}
return render_template('site.html', **context)
def produce_automl_results(dir_img_path: str = "",
dir_destination_path: str = ""):
from google.cloud import automl
project_id = "hidden"
model_id = "hidden"
imgs = os.listdir(dir_img_path)
relevant_images = [
os.path.join(dir_img_path, i) for i in imgs if "png" in i
]
prediction_client = automl.PredictionServiceClient()
project_id = "hidden"
model_id = "hidden"
# Get the full path of the model.
model_full_id = automl.AutoMlClient.model_path(project_id, "us-central1",
model_id)
file_path = "../../data/005.png"
with open(file_path, "rb") as content_file:
content = content_file.read()
image = automl.Image(image_bytes=content)
payload = automl.ExamplePayload(image=image)
params = {}
request = automl.PredictRequest(name=model_full_id,
payload=payload,
params=params)
response = prediction_client.predict(request=request)
print("Prediction results:")
for result in response.payload:
print("Predicted class name: {}".format(result.display_name))
print("Predicted class score: {}".format(result.classification.score))
def __init__(self, prediction_threshold: float = 0.8):
"""
Wrapper class for asynchronous, batch-oriented image classification with Google Vision AutoML.
:param prediction_threshold: Score threshold for predictions (default: 0.8)
:raises ValueError for invalid prediction_threshold values
"""
if not (0 <= prediction_threshold <= 1):
raise ValueError(
"Prediction threshold must be within the range [0,1].")
# Let Google Cloud libraries pick up credentials from environment variables
config, _ = read_config()
os.environ["PROJECT_ID"] = config["project_id"]
os.environ["GOOGLE_APPLICATION_CREDENTIALS"] = config["credentials"]
# Google Cloud Vision AutoML
self.__gvision_client = automl.PredictionServiceClient()
self.__full_model_id = self.__gvision_client.model_path(
project=config["project_id"],
location=config["location"],
model=config["model_id"])
self.__prediction_threshold = prediction_threshold
def shape():
project_id = "quixotic-sol-281406"
model_id = "ICN8423774195787235328"
file_path = "images/result.jpg"
prediction_client = automl.PredictionServiceClient()
# Get the full path of the model.
model_full_id = prediction_client.model_path(project_id, "us-central1",
model_id)
# Read the file.
with open(file_path, "rb") as content_file:
content = content_file.read()
image = automl.types.Image(image_bytes=content)
payload = automl.types.ExamplePayload(image=image)
# params is additional domain-specific parameters.
# score_threshold is used to filter the result
# https://cloud.google.com/automl/docs/reference/rpc/google.cloud.automl.v1#predictrequest
params = {"score_threshold": "0.8"}
response = prediction_client.predict(model_full_id, payload, params)
# print("Prediction results:")
for result in response.payload:
# print("Predicted class name: {}".format(result.display_name))
# print(
#"Predicted class score: {}".format(
# result.image_object_detection.score
#)
#)
bounding_box = result.image_object_detection.bounding_box
# print("Normalized Vertices:")
#for vertex in bounding_box.normalized_vertices:
# print("\tX: {}, Y: {}".format(vertex.x, vertex.y))
return result.display_name
def label(img):
project_id = "salary-dost"
model_id = "ICN4622096194519171072"
prediction_client = automl.PredictionServiceClient()
# Get the full path of the model.
model_full_id = prediction_client.model_path(project_id, "us-central1",
model_id)
content = img
imgByte = io.BytesIO()
content.save(imgByte, format='jpeg')
imgByte = imgByte.getvalue()
image = automl.types.Image(image_bytes=imgByte)
payload = automl.types.ExamplePayload(image=image)
# params is additional domain-specific parameters.
# score_threshold is used to filter the result
# https://cloud.google.com/automl/docs/reference/rpc/google.cloud.automl.v1#predictrequest
params = {"score_threshold": "0.8"}
response = prediction_client.predict(model_full_id,
payload,
params,
timeout=30.0)
#print("Prediction results:")
for result in response.payload:
#print("Predicted class name: {}".format(result.display_name))
#print("Predicted class score: {}".format(result.classification.score))
res = result.display_name
return res
print("Training operation name: {}".format(response.operation.name))
print("Training started...")
model_id = response.result().name.split('/')[-1]
start_deploying_time = time.time()
model_full_id = client.model_path(project_id, "us-central1", model_id)
response = client.deploy_model(name=model_full_id)
response.result()
start_prediction_time = time.time()
prediction_client = automl.PredictionServiceClient()
preds = []
for img in tqdm.tqdm(data['test_images']):
score = None
while True:
try:
score = get_prediction(img)
break
except:
time.sleep(5)
preds.append(score)
preds = np.array(preds)
metrics = get_metrics(data['test_labels'], preds)
def automl_vision(file_path):
"""Predict."""
# [START automl_vision_classification_predict]
#print(file_path)
# autoML JSon 키 설정
#credential_path = r"키.json"
os.environ['GOOGLE_APPLICATION_CREDENTIALS'] = credential_path
project_id = "수정"
model_id = "수정"
# 원본 이미지
img = cv2.imread(file_path)
# 원본 이미지 크기
img_height = img.shape[0]
img_width = img.shape[1]
# RGB 색상으로 변환
orig = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
prediction_client = automl.PredictionServiceClient()
# Get the full path of the model.
model_full_id = automl.AutoMlClient.model_path(project_id, "us-central1",
model_id)
# 파일 읽어들이기
with open(file_path, "rb") as content_file:
content = content_file.read()
image = automl.Image(image_bytes=content)
payload = automl.ExamplePayload(image=image)
# params is additional domain-specific parameters.
# score_threshold is used to filter the result
# https://cloud.google.com/automl/docs/reference/rpc/google.cloud.automl.v1#predictrequest
params = {"score_threshold": "0.8"}
request = automl.PredictRequest(name=model_full_id,
payload=payload,
params=params)
response = prediction_client.predict(request=request)
tags = []
for result in response.payload:
# 태그로 잡힌 박스 몇 개인지
bounding_box = result.image_object_detection.bounding_box
# normalized vertices로 되어있는 좌표를 실사이즈로 변환 및 직사각형의 네 꼭짓점 좌표를 구함
lu = bounding_box.normalized_vertices[0] # left/up x, y 좌표
rd = bounding_box.normalized_vertices[1] # right/down x, y 좌표
w = rd.x - lu.x
h = rd.y - lu.y
# cropping
cropped_img = orig[int(lu.y * img_height):int((lu.y + h) * img_height),
int(lu.x * img_width):int((lu.x + w) * img_width)]
texts = my_detect_text_mat(
cropped_img) #잘린 이미지들 속 글자 추출하기 위해 my_detect_text_mat() 호출
tags.append(texts[0].description)
return tags
def predict(model_id, project_id, csvpath_test, resultpath_cloud):
# project_id = "YOUR_PROJECT_ID"
# model_id = "YOUR_MODEL_ID"
input_uri = csvpath_test
output_uri = resultpath_cloud
prediction_client = automl.PredictionServiceClient()
# Get the full path of the model.
model_full_id = prediction_client.model_path(
project_id, "us-central1", model_id
)
gcs_source = automl.types.GcsSource(input_uris=[input_uri])
input_config = automl.types.BatchPredictInputConfig(gcs_source=gcs_source)
gcs_destination = automl.types.GcsDestination(output_uri_prefix=output_uri)
output_config = automl.types.BatchPredictOutputConfig(
gcs_destination=gcs_destination
)
# Print model stats
# client = automl.AutoMlClient()
# print("List of model evaluations:")
# for evaluation in client.list_model_evaluations(model_full_id, ""):
# print("Model evaluation name: {}".format(evaluation.name))
# print(
# "Model annotation spec id: {}".format(
# evaluation.annotation_spec_id
# )
# )
# print("Create Time:")
# print("\tseconds: {}".format(evaluation.create_time.seconds))
# print("\tnanos: {}".format(evaluation.create_time.nanos / 1e9))
# print(
# "Evaluation example count: {}".format(
# evaluation.evaluated_example_count
# )
# )
# print(
# "Classification model evaluation metrics: {}".format(
# evaluation.classification_evaluation_metrics
# )
# )
print('batch prediction beginning...')
while True:
try:
response = prediction_client.batch_predict(model_full_id, input_config, output_config, params={'score_threshold': '0'})
break
except google.api_core.exceptions.ResourceExhausted:
print('Concurrent batch prediction quota exhausted. This is a common error. Waiting 4 sec...')
sleep(4)
except google.api_core.exceptions.DeadlineExceeded:
print('Deadline exceeded, whatever, keep trying')
sleep(4)
except google.api_core.exceptions.ServiceUnavailable:
print('Service unavailable error... what??')
sleep(4)
except google.api_core.exceptions.NotFound:
print(f'Model {model_full_id} not found. probably training not finished. waiting 30 sec')
sleep(30)
except Exception as e:
raise # print(f'some other random error: {e}')
print(f"Batch prediction operation id {response.operation.name.split('/')[-1]} has started \n\n\n")
print(f"Waiting for batch prediction operation id {response.operation.name.split('/')[-1]} to complete...")
# def callback(operation_future):
# # Handle result.
# print("Batch Prediction results saved to Cloud Storage bucket. {}".format(operation_future.result()))
#
# response.add_done_callback(callback)
while True:
if response.done():
break
else:
sleep(300)
print('Model still predicting ...')
print("Batch Prediction results saved to Cloud Storage bucket. {}".format(response.result()))