# Licensed to the Apache Software Foundation (ASF) under one
# or more contributor license agreements. See the NOTICE file
# distributed with this work for additional information
# regarding copyright ownership. The ASF licenses this file
# to you under the Apache License, Version 2.0 (the
# "License"); you may not use this file except in compliance
# with the License. You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing,
# software distributed under the License is distributed on an
# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
# KIND, either express or implied. See the License for the
# specific language governing permissions and limitations
# under the License.
from __future__ import annotations
import os
from datetime import datetime
from airflow import DAG, settings
from airflow.decorators import task, task_group
from airflow.models import Connection
from airflow.models.baseoperator import chain
from airflow.providers.amazon.aws.hooks.comprehend import ComprehendHook
from airflow.providers.amazon.aws.operators.comprehend import (
ComprehendCreateDocumentClassifierOperator,
)
from airflow.providers.amazon.aws.operators.s3 import (
S3CreateBucketOperator,
S3CreateObjectOperator,
S3DeleteBucketOperator,
)
from airflow.providers.amazon.aws.sensors.comprehend import (
ComprehendCreateDocumentClassifierCompletedSensor,
)
from airflow.providers.amazon.aws.transfers.http_to_s3 import HttpToS3Operator
from airflow.utils.trigger_rule import TriggerRule
from providers.tests.system.amazon.aws.utils import SystemTestContextBuilder
[docs]ROLE_ARN_KEY = "ROLE_ARN"
[docs]sys_test_context_task = SystemTestContextBuilder().add_variable(ROLE_ARN_KEY).build()
[docs]DAG_ID = "example_comprehend_document_classifier"
[docs]ANNOTATION_BUCKET_KEY = "training-labels/label.csv"
[docs]TRAINING_DATA_PREFIX = "training-docs"
# To create a custom document classifier, we need a minimum of 10 documents for each label.
# for testing purpose, we will generate 10 copies of each document referenced below.
[docs]PUBLIC_DATA_SOURCES = [
{
"fileName": "discharge-summary.pdf",
"endpoint": "aws-samples/amazon-comprehend-examples/blob/master/building-custom-classifier/sample-docs/discharge-summary.pdf?raw=true",
},
{
"fileName": "doctors-notes.pdf",
"endpoint": "aws-samples/amazon-comprehend-examples/blob/master/building-custom-classifier/sample-docs/doctors-notes.pdf?raw=true",
},
]
# Annotations file won't allow headers
# label,document name,page number
[docs]ANNOTATIONS = """DISCHARGE_SUMMARY,discharge-summary-0.pdf,1
DISCHARGE_SUMMARY,discharge-summary-1.pdf,1
DISCHARGE_SUMMARY,discharge-summary-2.pdf,1
DISCHARGE_SUMMARY,discharge-summary-3.pdf,1
DISCHARGE_SUMMARY,discharge-summary-4.pdf,1
DISCHARGE_SUMMARY,discharge-summary-5.pdf,1
DISCHARGE_SUMMARY,discharge-summary-6.pdf,1
DISCHARGE_SUMMARY,discharge-summary-7.pdf,1
DISCHARGE_SUMMARY,discharge-summary-8.pdf,1
DISCHARGE_SUMMARY,discharge-summary-9.pdf,1
DOCTOR_NOTES,doctors-notes-0.pdf,1
DOCTOR_NOTES,doctors-notes-1.pdf,1
DOCTOR_NOTES,doctors-notes-2.pdf,1
DOCTOR_NOTES,doctors-notes-3.pdf,1
DOCTOR_NOTES,doctors-notes-4.pdf,1
DOCTOR_NOTES,doctors-notes-5.pdf,1
DOCTOR_NOTES,doctors-notes-6.pdf,1
DOCTOR_NOTES,doctors-notes-7.pdf,1
DOCTOR_NOTES,doctors-notes-8.pdf,1
DOCTOR_NOTES,doctors-notes-9.pdf,1"""
@task_group
[docs]def document_classifier_workflow():
# [START howto_operator_create_document_classifier]
create_document_classifier = ComprehendCreateDocumentClassifierOperator(
task_id="create_document_classifier",
document_classifier_name=classifier_name,
input_data_config=input_data_configurations,
output_data_config=output_data_configurations,
mode="MULTI_CLASS",
data_access_role_arn=test_context[ROLE_ARN_KEY],
language_code="en",
document_classifier_kwargs=document_classifier_kwargs,
)
# [END howto_operator_create_document_classifier]
create_document_classifier.wait_for_completion = False
# [START howto_sensor_create_document_classifier]
await_create_document_classifier = ComprehendCreateDocumentClassifierCompletedSensor(
task_id="await_create_document_classifier", document_classifier_arn=create_document_classifier.output
)
# [END howto_sensor_create_document_classifier]
@task(trigger_rule=TriggerRule.ALL_DONE)
def delete_classifier(document_classifier_arn: str):
ComprehendHook().conn.delete_document_classifier(DocumentClassifierArn=document_classifier_arn)
chain(
create_document_classifier,
await_create_document_classifier,
delete_classifier(create_document_classifier.output),
)
@task_group
[docs]def copy_data_to_s3(bucket: str, sources: list[dict], prefix: str, number_of_copies=1):
"""
Copy some sample data to S3 using HttpToS3Operator.
:param bucket: Name of the Amazon S3 bucket to send the data.
:param prefix: Folder to store the files
:param number_of_copies: Number of files to create for a document from the sources
:param sources: Public available data locations
"""
"""
EX: If number_of_copies is 2, sources has file name 'file.pdf', and prefix is 'training-docs'.
Will generate two copies and upload to s3:
- training-docs/file-0.pdf
- training-docs/file-1.pdf
"""
http_to_s3_configs = [
{
"endpoint": source["endpoint"],
"s3_key": f"{prefix}/{os.path.splitext(os.path.basename(source['fileName']))[0]}-{counter}{os.path.splitext(os.path.basename(source['fileName']))[1]}",
}
for counter in range(number_of_copies)
for source in sources
]
@task
def create_connection(conn_id):
conn = Connection(
conn_id=conn_id,
conn_type="http",
host="https://github.com/",
)
session = settings.Session()
session.add(conn)
session.commit()
@task(trigger_rule=TriggerRule.ALL_DONE)
def delete_connection(conn_id):
session = settings.Session()
conn_to_details = session.query(Connection).filter(Connection.conn_id == conn_id).first()
session.delete(conn_to_details)
session.commit()
http_to_s3_task = HttpToS3Operator.partial(
task_id="http_to_s3_task",
http_conn_id=http_conn_id,
s3_bucket=bucket,
).expand_kwargs(http_to_s3_configs)
chain(create_connection(http_conn_id), http_to_s3_task, delete_connection(http_conn_id))
with DAG(
dag_id=DAG_ID,
schedule="@once",
start_date=datetime(2021, 1, 1),
tags=["example"],
catchup=False,
) as dag:
[docs] test_context = sys_test_context_task()
env_id = test_context["ENV_ID"]
classifier_name = f"{env_id}-custom-document-classifier"
bucket_name = f"{env_id}-comprehend-document-classifier"
http_conn_id = f"{env_id}-git"
input_data_configurations = {
"S3Uri": f"s3://{bucket_name}/{ANNOTATION_BUCKET_KEY}",
"DataFormat": "COMPREHEND_CSV",
"DocumentType": "SEMI_STRUCTURED_DOCUMENT",
"Documents": {"S3Uri": f"s3://{bucket_name}/{TRAINING_DATA_PREFIX}/"},
"DocumentReaderConfig": {
"DocumentReadAction": "TEXTRACT_DETECT_DOCUMENT_TEXT",
"DocumentReadMode": "SERVICE_DEFAULT",
},
}
output_data_configurations = {"S3Uri": f"s3://{bucket_name}/output/"}
document_classifier_kwargs = {"VersionName": "v1"}
create_bucket = S3CreateBucketOperator(
task_id="create_bucket",
bucket_name=bucket_name,
)
upload_annotation_file = S3CreateObjectOperator(
task_id="upload_annotation_file",
s3_bucket=bucket_name,
s3_key=ANNOTATION_BUCKET_KEY,
data=ANNOTATIONS.encode("utf-8"),
)
delete_bucket = S3DeleteBucketOperator(
task_id="delete_bucket",
trigger_rule=TriggerRule.ALL_DONE,
bucket_name=bucket_name,
force_delete=True,
)
chain(
test_context,
create_bucket,
upload_annotation_file,
copy_data_to_s3(
bucket=bucket_name, sources=PUBLIC_DATA_SOURCES, prefix=TRAINING_DATA_PREFIX, number_of_copies=10
),
# TEST BODY
document_classifier_workflow(),
# TEST TEARDOWN
delete_bucket,
)
from dev.tests_common.test_utils.watcher import watcher
# This test needs watcher in order to properly mark success/failure
# when "tearDown" task with trigger rule is part of the DAG
list(dag.tasks) >> watcher()
from dev.tests_common.test_utils.system_tests import get_test_run # noqa: E402
# Needed to run the example DAG with pytest (see: tests/system/README.md#run_via_pytest)
[docs]test_run = get_test_run(dag)