fix: shorten base job name to avoid collision (#1170)

* fix: shorten job base name to avoid collision

* use unique_name_from_base

* use unique_name_from_base for integ tests

Author: chuyang-deng

src/sagemaker/automl/automl.py

@@ -19,7 +19,7 @@
 from sagemaker.automl.candidate_estimator import CandidateEstimator
 from sagemaker.job import _Job
 from sagemaker.session import Session
-from sagemaker.utils import name_from_base
+from sagemaker.utils import unique_name_from_base
 
 
 class AutoML(object):
@@ -379,9 +379,9 @@ def _prepare_for_auto_ml_job(self, job_name=None):
             if self.base_job_name:
                 base_name = self.base_job_name
             else:
-                base_name = "sagemaker-auto-ml"
+                base_name = "automl"
             # CreateAutoMLJob API validates that member length less than or equal to 32
-            self.current_job_name = name_from_base(base_name, max_length=32)
+            self.current_job_name = unique_name_from_base(base_name, max_length=32)
 
         if self.output_path is None:
             self.output_path = "s3://{}/".format(self.sagemaker_session.default_bucket())

tests/integ/test_auto_ml.py

@@ -58,9 +58,10 @@ def test_auto_ml_fit(sagemaker_session):
         max_candidates=3,
     )
 
+    job_name = unique_name_from_base("auto-ml", max_length=32)
     inputs = sagemaker_session.upload_data(path=TRAINING_DATA, key_prefix=PREFIX + "/input")
     with timeout(minutes=AUTO_ML_DEFAULT_TIMEMOUT_MINUTES):
-        auto_ml.fit(inputs)
+        auto_ml.fit(inputs, job_name=job_name)
 
 
 @pytest.mark.skipif(
@@ -76,8 +77,9 @@ def test_auto_ml_fit_local_input(sagemaker_session):
     )
 
     inputs = TRAINING_DATA
+    job_name = unique_name_from_base("auto-ml", max_length=32)
     with timeout(minutes=AUTO_ML_DEFAULT_TIMEMOUT_MINUTES):
-        auto_ml.fit(inputs)
+        auto_ml.fit(inputs, job_name=job_name)
 
 
 @pytest.mark.skipif(
@@ -91,10 +93,11 @@ def test_auto_ml_input_object_fit(sagemaker_session):
         sagemaker_session=sagemaker_session,
         max_candidates=1,
     )
+    job_name = unique_name_from_base("auto-ml", max_length=32)
     s3_input = sagemaker_session.upload_data(path=TRAINING_DATA, key_prefix=PREFIX + "/input")
     inputs = AutoMLInput(inputs=s3_input, target_attribute_name=TARGET_ATTRIBUTE_NAME)
     with timeout(minutes=AUTO_ML_DEFAULT_TIMEMOUT_MINUTES):
-        auto_ml.fit(inputs)
+        auto_ml.fit(inputs, job_name=job_name)
 
 
 @pytest.mark.skipif(
@@ -134,11 +137,12 @@ def test_auto_ml_invalid_target_attribute(sagemaker_session):
     auto_ml = AutoML(
         role=ROLE, target_attribute_name="y", sagemaker_session=sagemaker_session, max_candidates=1
     )
+    job_name = unique_name_from_base("auto-ml", max_length=32)
     inputs = sagemaker_session.upload_data(path=TRAINING_DATA, key_prefix=PREFIX + "/input")
     with pytest.raises(
         UnexpectedStatusException, match="Could not complete the data builder processing job."
     ):
-        auto_ml.fit(inputs)
+        auto_ml.fit(inputs, job_name=job_name)
 
 
 @pytest.mark.skipif(

tests/unit/sagemaker/automl/test_auto_ml.py

@@ -13,14 +13,13 @@
 from __future__ import absolute_import
 
 import pytest
-from mock import Mock, patch
+from mock import Mock
 from sagemaker import AutoML, AutoMLJob, AutoMLInput, CandidateEstimator
 
 MODEL_DATA = "s3://bucket/model.tar.gz"
 MODEL_IMAGE = "mi"
 ENTRY_POINT = "blah.py"
 
-TIMESTAMP = "2017-11-06-14:14:15.671"
 BUCKET_NAME = "mybucket"
 INSTANCE_COUNT = 1
 INSTANCE_TYPE = "ml.c5.2xlarge"
@@ -32,7 +31,6 @@
 DEFAULT_OUTPUT_PATH = "s3://{}/".format(BUCKET_NAME)
 LOCAL_DATA_PATH = "file://data"
 DEFAULT_MAX_CANDIDATES = 500
-DEFAULT_JOB_NAME = "sagemake-{}".format(TIMESTAMP)
 
 JOB_NAME = "default-job-name"
 JOB_NAME_2 = "banana-auto-ml-job"
@@ -283,38 +281,34 @@ def test_auto_ml_additional_optional_params(sagemaker_session):
     }
 
 
-@patch("time.strftime", return_value=TIMESTAMP)
-def test_auto_ml_default_fit(strftime, sagemaker_session):
+def test_auto_ml_default_fit(sagemaker_session):
     auto_ml = AutoML(
         role=ROLE, target_attribute_name=TARGET_ATTRIBUTE_NAME, sagemaker_session=sagemaker_session
     )
     inputs = DEFAULT_S3_INPUT_DATA
     auto_ml.fit(inputs)
     sagemaker_session.auto_ml.assert_called_once()
     _, args = sagemaker_session.auto_ml.call_args
-    assert args == {
-        "input_config": [
-            {
-                "DataSource": {
-                    "S3DataSource": {"S3DataType": "S3Prefix", "S3Uri": DEFAULT_S3_INPUT_DATA}
-                },
-                "TargetAttributeName": TARGET_ATTRIBUTE_NAME,
-            }
-        ],
-        "output_config": {"S3OutputPath": DEFAULT_OUTPUT_PATH},
-        "auto_ml_job_config": {
-            "CompletionCriteria": {"MaxCandidates": DEFAULT_MAX_CANDIDATES},
-            "SecurityConfig": {
-                "EnableInterContainerTrafficEncryption": ENCRYPT_INTER_CONTAINER_TRAFFIC
+    assert args["input_config"] == [
+        {
+            "DataSource": {
+                "S3DataSource": {"S3DataType": "S3Prefix", "S3Uri": DEFAULT_S3_INPUT_DATA}
             },
+            "TargetAttributeName": TARGET_ATTRIBUTE_NAME,
+        }
+    ]
+    assert args["output_config"] == {"S3OutputPath": DEFAULT_OUTPUT_PATH}
+    assert args["auto_ml_job_config"] == {
+        "CompletionCriteria": {"MaxCandidates": DEFAULT_MAX_CANDIDATES},
+        "SecurityConfig": {
+            "EnableInterContainerTrafficEncryption": ENCRYPT_INTER_CONTAINER_TRAFFIC
         },
-        "role": ROLE,
-        "job_name": DEFAULT_JOB_NAME,
-        "problem_type": None,
-        "job_objective": None,
-        "generate_candidate_definitions_only": GENERATE_CANDIDATE_DEFINITIONS_ONLY,
-        "tags": None,
     }
+    assert args["role"] == ROLE
+    assert args["problem_type"] is None
+    assert args["job_objective"] is None
+    assert args["generate_candidate_definitions_only"] == GENERATE_CANDIDATE_DEFINITIONS_ONLY
+    assert args["tags"] is None
 
 
 def test_auto_ml_local_input(sagemaker_session):