Skip to content

Commit d54edb4

Browse files
authored
change: Update data wrangler images (#3770)
1 parent 2bac010 commit d54edb4

File tree

4 files changed

+58
-10
lines changed

4 files changed

+58
-10
lines changed

src/sagemaker/image_uri_config/data-wrangler.json

Lines changed: 29 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -7,6 +7,35 @@
77
"ap-east-1": "707077482487",
88
"ap-northeast-1": "649008135260",
99
"ap-northeast-2": "131546521161",
10+
"ap-northeast-3": "913387583493",
11+
"ap-south-1": "089933028263",
12+
"ap-southeast-1": "119527597002",
13+
"ap-southeast-2": "422173101802",
14+
"ca-central-1": "557239378090",
15+
"eu-central-1": "024640144536",
16+
"eu-north-1": "054986407534",
17+
"eu-south-1": "488287956546",
18+
"eu-west-1": "245179582081",
19+
"eu-west-2": "894491911112",
20+
"eu-west-3": "807237891255",
21+
"me-south-1": "376037874950",
22+
"sa-east-1": "424196993095",
23+
"us-east-1": "663277389841",
24+
"us-east-2": "415577184552",
25+
"us-west-1": "926135532090",
26+
"us-west-2": "174368400705",
27+
"cn-north-1": "245909111842",
28+
"cn-northwest-1": "249157047649"
29+
},
30+
"repository": "sagemaker-data-wrangler-container"
31+
},
32+
"2.x": {
33+
"registries": {
34+
"af-south-1": "143210264188",
35+
"ap-east-1": "707077482487",
36+
"ap-northeast-1": "649008135260",
37+
"ap-northeast-2": "131546521161",
38+
"ap-northeast-3": "913387583493",
1039
"ap-south-1": "089933028263",
1140
"ap-southeast-1": "119527597002",
1241
"ap-southeast-2": "422173101802",

src/sagemaker/image_uris.py

Lines changed: 4 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -36,6 +36,7 @@
3636
SKLEARN_FRAMEWORK = "sklearn"
3737
TRAINIUM_ALLOWED_FRAMEWORKS = "pytorch"
3838
INFERENCE_GRAVITON = "inference_graviton"
39+
DATA_WRANGLER_FRAMEWORK = "data-wrangler"
3940

4041

4142
@override_pipeline_parameter_var
@@ -461,6 +462,9 @@ def _validate_version_and_set_if_needed(version, config, framework):
461462

462463
return available_versions[0]
463464

465+
if version is None and framework in [DATA_WRANGLER_FRAMEWORK]:
466+
version = _get_latest_versions(available_versions)
467+
464468
_validate_arg(version, available_versions + aliased_versions, "{} version".format(framework))
465469
return version
466470

tests/unit/sagemaker/image_uris/test_data_wrangler.py

Lines changed: 24 additions & 9 deletions
Original file line numberDiff line numberDiff line change
@@ -20,6 +20,7 @@
2020
"ap-east-1": "707077482487",
2121
"ap-northeast-1": "649008135260",
2222
"ap-northeast-2": "131546521161",
23+
"ap-northeast-3": "913387583493",
2324
"ap-south-1": "089933028263",
2425
"ap-southeast-1": "119527597002",
2526
"ap-southeast-2": "422173101802",
@@ -39,15 +40,29 @@
3940
"cn-north-1": "245909111842",
4041
"cn-northwest-1": "249157047649",
4142
}
43+
VERSIONS = ["1.x", "2.x"]
4244

4345

4446
def test_data_wrangler_ecr_uri():
45-
for region in DATA_WRANGLER_ACCOUNTS.keys():
46-
actual_uri = image_uris.retrieve("data-wrangler", region=region)
47-
expected_uri = expected_uris.algo_uri(
48-
"sagemaker-data-wrangler-container",
49-
DATA_WRANGLER_ACCOUNTS[region],
50-
region,
51-
version="1.x",
52-
)
53-
assert expected_uri == actual_uri
47+
for version in VERSIONS:
48+
for region in DATA_WRANGLER_ACCOUNTS.keys():
49+
actual_uri = image_uris.retrieve("data-wrangler", region=region, version="1.x")
50+
expected_uri = expected_uris.algo_uri(
51+
"sagemaker-data-wrangler-container",
52+
DATA_WRANGLER_ACCOUNTS[region],
53+
region,
54+
version="1.x",
55+
)
56+
assert expected_uri == actual_uri
57+
58+
59+
def test_data_wrangler_ecr_uri_none():
60+
region = "us-west-2"
61+
actual_uri = image_uris.retrieve("data-wrangler", region=region)
62+
expected_uri = expected_uris.algo_uri(
63+
"sagemaker-data-wrangler-container",
64+
DATA_WRANGLER_ACCOUNTS[region],
65+
region,
66+
version=VERSIONS[-1],
67+
)
68+
assert expected_uri == actual_uri

tests/unit/sagemaker/wrangler/test_processing.py

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -23,7 +23,7 @@
2323
REGION = "us-west-2"
2424
DATA_WRANGLER_RECIPE_SOURCE = "s3://data_wrangler_flows/flow-26-18-43-16-0b48ac2e.flow"
2525
DATA_WRANGLER_CONTAINER_URI = (
26-
"174368400705.dkr.ecr.us-west-2.amazonaws.com/sagemaker-data-wrangler-container:1.x"
26+
"174368400705.dkr.ecr.us-west-2.amazonaws.com/sagemaker-data-wrangler-container:2.x"
2727
)
2828
MOCK_S3_URI = "s3://mock_data/mock.csv"
2929

0 commit comments

Comments
 (0)