Requesting Contributions

Requesting contributions when scoring.

Prerequisites

• h2o_mlops_scoring_client-*-py3-none-any.whl file or access to the Python Package Index (PyPI)
• Java

Setup

Install the h2o_mlops_scoring_client with pip.

Notes

• If running locally, the number of cores used (and thus parallel processes) can be overridden with:

num_cores = 10
h2o_mlops_scoring_client.spark_master = f"local[{num_cores}]"

Example Usage for Data Frames

import h2o_mlops_scoring_client
import pandas

Choose a feature type for contributions - either original or transformed.

CONTRIB_FEATURE_TYPE = h2o_mlops_scoring_client.FeatureType.ORIGINAL

Pass the feature type when scoring.

ID_COLUMN = "ID"
MLOPS_ENDPOINT_URL = "https://model.internal.dedicated.h2o.ai/65427177-dd10-44dd-abf8-76ab29f60799/model/score"
DATA_FRAME = pandas.read_csv("/Users/jgranados/datasets/creditcard.csv")

h2o_mlops_scoring_client.score_data_frame(
    mlops_endpoint_url=MLOPS_ENDPOINT_URL,
    data_frame=DATA_FRAME, 
    id_column=ID_COLUMN, 
    request_contributions=CONTRIB_FEATURE_TYPE
)

23/08/24 16:08:28 INFO h2o_mlops_scoring_client: Connecting to H2O.ai MLOps scorer at 'https://model.internal.dedicated.h2o.ai/65427177-dd10-44dd-abf8-76ab29f60799/model/score'
23/08/24 16:08:28 INFO h2o_mlops_scoring_client: Starting scoring data frame
23/08/24 16:09:04 INFO h2o_mlops_scoring_client: Scoring complete
23/08/24 16:09:04 INFO h2o_mlops_scoring_client: Total run time: 0:00:36
23/08/24 16:09:04 INFO h2o_mlops_scoring_client: Scoring run time: 0:00:36

	ID	default payment next month.0	default payment next month.1	contrib_LIMIT_BAL	contrib_MARRIAGE	contrib_AGE	contrib_PAt_0	contrib_PAY_2	contrib_PAY_3	contrib_PAY_4	...	contrib_BILL_AMT4	contrib_BILL_AMT5	contrib_BILL_AMT6	contrib_PAY_AMT1	contrib_PAY_AMT2	contrib_PAY_AMT3	contrib_PAY_AMT4	contrib_PAY_AMT5	contrib_PAY_AMT6	contrib_bias
0	1.0	0.548391	0.451609	0.318705	0.054260	0.024333	-0.143527	0.199437	0.123991	0.001567	...	-0.042873	0.004312	-0.044986	0.190008	0.116479	0.196877	0.107893	0.060028	0.063311	-1.516447
1	2.0	0.559799	0.440201	0.011380	-0.044559	-0.011563	0.017800	0.353959	0.258357	-0.009022	...	0.002973	-0.003932	-0.034921	0.184942	0.127812	-0.081676	-0.010390	0.062603	-0.015993	-1.516447
2	3.0	0.913488	0.086512	0.091553	-0.073675	-0.014241	-0.231872	-0.036224	-0.033809	-0.019079	...	-0.013316	-0.017721	-0.021681	0.091626	0.125422	-0.108036	-0.012150	-0.012263	-0.018186	-1.516447
3	4.0	0.796132	0.203868	0.267449	0.065499	0.019306	0.061837	-0.055391	-0.033248	-0.022298	...	-0.013313	-0.033297	0.000984	0.068004	-0.016718	-0.073336	-0.014324	-0.039674	-0.041868	-1.516447
4	5.0	0.601541	0.398459	0.170616	0.058417	-0.019828	0.940433	-0.060075	-0.303138	-0.018102	...	-0.004610	0.009174	-0.005628	0.064400	-0.391913	-0.065873	-0.021920	0.036659	0.023616	-1.516447
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
994	23995.0	0.866306	0.133694	0.478230	-0.091026	0.014865	-0.215355	-0.036565	-0.026022	-0.020773	...	-0.000289	-0.021694	0.008951	0.083203	-0.071253	-0.044852	-0.022899	-0.028537	0.018177	-1.516447
995	23996.0	0.608550	0.391450	0.087377	0.044217	-0.015629	0.014538	0.205578	0.199663	-0.011091	...	0.021188	-0.020175	0.056376	-0.027547	0.197609	-0.037477	-0.051646	-0.051575	0.029902	-1.516447
996	23997.0	0.814939	0.185061	0.536682	0.099411	0.017427	-0.208787	-0.027522	-0.022646	-0.023438	...	0.002923	-0.005027	-0.000314	0.081016	-0.031194	-0.071644	-0.015263	0.059949	-0.030882	-1.516447
997	23998.0	0.871589	0.128411	0.481494	-0.084286	-0.029002	-0.224106	-0.039429	-0.028018	-0.022154	...	0.005962	-0.013858	-0.035625	0.069608	-0.038320	-0.066714	0.018526	-0.049960	-0.024566	-1.516447
998	23999.0	0.868578	0.131422	0.461986	-0.083266	-0.035072	-0.229247	-0.036450	-0.013597	-0.024662	...	0.012886	0.010888	-0.066466	0.029243	0.109742	-0.081972	-0.028465	0.052664	0.036372	-1.516447

23999 rows × 25 columns

Example Usage for Source/Sink

import h2o_mlops_scoring_client

Choose a feature type for contributions - either original or transformed.

CONTRIB_FEATURE_TYPE = h2o_mlops_scoring_client.FeatureType.TRANSFORMED

Pass the feature type when scoring.

ID_COLUMN = "ID"
MLOPS_ENDPOINT_URL = "https://model.internal.dedicated.h2o.ai/65427177-dd10-44dd-abf8-76ab29f60799/model/score"
SOURCE_DATA = "file:///Users/jgranados/datasets/creditcard.csv"
SINK_LOCATION = "file:///Users/jgranados/datasets/output/"
SOURCE_FORMAT = h2o_mlops_scoring_client.Format.CSV
SINK_FORMAT = h2o_mlops_scoring_client.Format.CSV
SINK_WRITE_MODE = h2o_mlops_scoring_client.WriteMode.OVERWRITE

def preprocess(spark_df):
    return spark_df.repartition(30)

h2o_mlops_scoring_client.score_source_sink(
    mlops_endpoint_url=MLOPS_ENDPOINT_URL,
    id_column=ID_COLUMN,
    source_data=SOURCE_DATA,
    source_format=SOURCE_FORMAT,
    sink_location=SINK_LOCATION,
    sink_format=SINK_FORMAT,
    sink_write_mode=SINK_WRITE_MODE,
    preprocess_method=preprocess,
    request_contributions=CONTRIB_FEATURE_TYPE
) 

23/08/24 16:09:42 INFO h2o_mlops_scoring_client: Connecting to H2O.ai MLOps scorer at 'https://model.internal.dedicated.h2o.ai/65427177-dd10-44dd-abf8-76ab29f60799/model/score'
23/08/24 16:09:45 INFO h2o_mlops_scoring_client: Applying preprocess method
23/08/24 16:09:45 INFO h2o_mlops_scoring_client: Starting scoring from 'file:///Users/jgranados/datasets/creditcard-borked.csv' to 'file:///Users/jgranados/datasets/output/'
23/08/24 16:10:16 INFO h2o_mlops_scoring_client: Scoring complete               
23/08/24 16:10:16 INFO h2o_mlops_scoring_client: Total run time: 0:00:36
23/08/24 16:10:16 INFO h2o_mlops_scoring_client: Scoring run time: 0:00:31

---

Feedback

• Submit and view feedback for this page
• Send feedback about H2O MLOps to cloud-feedback@h2o.ai