dhiegopiroto commented 1 year ago

Hi, When I tried to run the call for the UDF for inference with new data, I got this error message:

code execution python

test_df = session.create_dataframe([[250000,250000,200000,450000],[500000,500000,500000,500000],[8500,9500,2000,500]],  schema=['SEARCH_ENGINE','SOCIAL_MEDIA','VIDEO','EMAIL'])
test_df.select(
    'SEARCH_ENGINE','SOCIAL_MEDIA','VIDEO','EMAIL', 
    call_udf("predict_roi", 
    array_construct(col("SEARCH_ENGINE"), col("SOCIAL_MEDIA"), col("VIDEO"),col("EMAIL"))).as_("PREDICTED_ROI")).show()

Error returned

AttributeError: Can't get attribute '_passthrough_scorer' on <module 'sklearn.metrics._scorer' from '/usr/lib/python_udf/5f20772ab347018f6391e4ceaf2d4f3e3c074ff214498b0fb4f554f0cc1a7099/lib/python3.9/site-packages/sklearn/metrics/_scorer.py'>
 in function BATCH_PREDICT_ROI with handler udf_py_715127722.compute

relevant libraries and versions: snowflake-snowpark-python 1.7.0 snowflake-ml-python 1.0.8 scikit-learn 1.2.2

iamontheinet commented 1 year ago

Hi @dhiegopiroto! Did you create a new environment and follow the steps in the QS or did you install the packages individually? Can you also confirm your Python version. Thanks!

dhiegopiroto commented 1 year ago

Hey, @iamontheinet . I've tried to use the quickstart script and tried to create the dependencies based on the environment.yml file. In both cases, I got the same error.

(snowpark-de-ml) dpiroto@ % python --version Python 3.9.17

Thanks for the reply :)

dhiegopiroto commented 1 year ago

let me just to add all the exception message:

`--------------------------------------------------------------------------- SnowparkSQLException Traceback (most recent call last) /Users/dpiroto/src/github/sfguide-getting-started-dataengineering-ml-snowpark-python/Snowpark_For_Python_ML.ipynb Cell 17 line 3 1 test_df = session.create_dataframe([[250000,250000,200000,450000],[500000,500000,500000,500000],[8500,9500,2000,500]], 2 schema=['SEARCH_ENGINE','SOCIAL_MEDIA','VIDEO','EMAIL']) ----> 3 test_df.select( 4 'SEARCH_ENGINE','SOCIAL_MEDIA','VIDEO','EMAIL', 5 call_udf("predict_roi", 6 array_construct(col("SEARCH_ENGINE"), col("SOCIALMEDIA"), col("VIDEO"), col("EMAIL"))).as("PREDICTED_ROI")).show()

File ~/opt/anaconda3/envs/snowpark-de-ml/lib/python3.9/site-packages/snowflake/snowpark/_internal/telemetry.py:139, in df_collect_api_telemetry..wrap(*args, kwargs) 136 @functools.wraps(func) 137 def wrap(*args, *kwargs): 138 with args[0]._session.query_history() as query_history: --> 139 result = func(args, kwargs) 140 plan = args[0]._select_statement or args[0]._plan 141 api_calls = [ 142 *plan.api_calls, 143 {TelemetryField.NAME.value: f"DataFrame.{func.name}"}, 144 ]

File ~/opt/anaconda3/envs/snowpark-de-ml/lib/python3.9/site-packages/snowflake/snowpark/dataframe.py:2855, in DataFrame.show(self, n, max_width, statement_params) 2836 @df_collect_api_telemetry 2837 def show( 2838 self, ... return _getattribute(sys.modules[module], name)[0] File "/usr/lib/python_udf/cc189aaa19e7dc666cf02176e2f22eea863747958e19ccd50dfa7f40bf5684dd/lib/python3.9/pickle.py", line 331, in _getattribute raise AttributeError("Can't get attribute {!r} on {!r}" AttributeError: Can't get attribute '_passthrough_scorer' on <module 'sklearn.metrics._scorer' from '/usr/lib/python_udf/cc189aaa19e7dc666cf02176e2f22eea863747958e19ccd50dfa7f40bf5684dd/lib/python3.9/site-packages/sklearn/metrics/_scorer.py'> in function PREDICT_ROI with handler compute`

and all conda libraries used:

Name Version Build Channel

aelkhoury79 commented 1 year ago

Hi, Also stuck at the same step. Error message below:

SnowparkSQLException Traceback (most recent call last) Cell In[13], line 1 ----> 1 test_df.select( 2 'SEARCH_ENGINE','SOCIAL_MEDIA','VIDEO','EMAIL', 3 call_udf("predict_roi", 4 array_construct(col("SEARCH_ENGINE"), col("SOCIALMEDIA"), col("VIDEO"), col("EMAIL"))).as("PREDICTED_ROI")).show()

File ~\AppData\Local\miniconda3\envs\snowpark-de-ml\lib\site-packages\snowflake\snowpark_internal\telemetry.py:139, in df_collect_api_telemetry..wrap(*args, kwargs) 136 @functools.wraps(func) 137 def wrap(*args, *kwargs): 138 with args[0]._session.query_history() as query_history: --> 139 result = func(args, kwargs) 140 plan = args[0]._select_statement or args[0]._plan 141 api_calls = [ 142 *plan.api_calls, 143 {TelemetryField.NAME.value: f"DataFrame.{func.name}"}, 144 ]

File ~\AppData\Local\miniconda3\envs\snowpark-de-ml\lib\site-packages\snowflake\snowpark\dataframe.py:2811, in DataFrame.show(self, n, max_width, statement_params) 2792 @df_collect_api_telemetry 2793 def show( 2794 self, (...) 2798 statement_params: Optional[Dict[str, str]] = None, 2799 ) -> None: 2800 """Evaluates this DataFrame and prints out the first n rows with the 2801 specified maximum number of characters per column. 2802 (...) 2808 statement_params: Dictionary of statement level parameters to be set while executing this action. 2809 """ 2810 print( -> 2811 self._show_string( 2812 n, 2813 max_width, 2814 _statement_params=create_or_update_statement_params_with_query_tag( 2815 statement_params or self._statement_params, 2816 self._session.query_tag, 2817 SKIP_LEVELS_TWO, 2818 ), 2819 ) 2820 )

File ~\AppData\Local\miniconda3\envs\snowpark-de-ml\lib\site-packages\snowflake\snowpark\dataframe.py:2929, in DataFrame._show_string(self, n, max_width, kwargs) 2926 query = self._plan.queries[-1].sql.strip().lower() 2928 if is_sql_select_statement(query): -> 2929 result, meta = self._session._conn.get_result_and_metadata( 2930 self.limit(n)._plan, kwargs 2931 ) 2932 else: 2933 res, meta = self._session._conn.get_result_and_metadata( 2934 self._plan, **kwargs 2935 )

File ~\AppData\Local\miniconda3\envs\snowpark-de-ml\lib\site-packages\snowflake\snowpark_internal\server_connection.py:586, in ServerConnection.get_result_and_metadata(self, plan, kwargs) 583 def get_result_and_metadata( 584 self, plan: SnowflakePlan, kwargs 585 ) -> Tuple[List[Row], List[Attribute]]: --> 586 result_set, result_meta = self.get_result_set(plan, **kwargs) 587 result = result_set_to_rows(result_set["data"]) 588 meta = convert_result_meta_to_attribute(result_meta)

File ~\AppData\Local\miniconda3\envs\snowpark-de-ml\lib\site-packages\snowflake\snowpark_internal\analyzer\snowflake_plan.py:181, in SnowflakePlan.Decorator.wrap_exception..wrap(*args, **kwargs) 177 else: 178 ne = SnowparkClientExceptionMessages.SQL_EXCEPTION_FROM_PROGRAMMING_ERROR( 179 e 180 ) --> 181 raise ne.with_traceback(tb) from None

File ~\AppData\Local\miniconda3\envs\snowpark-de-ml\lib\site-packages\snowflake\snowpark_internal\analyzer\snowflake_plan.py:111, in SnowflakePlan.Decorator.wrap_exception..wrap(*args, kwargs) 109 def wrap(*args, *kwargs): 110 try: --> 111 return func(args, kwargs) 112 except snowflake.connector.errors.ProgrammingError as e: 113 query = None

File ~\AppData\Local\miniconda3\envs\snowpark-de-ml\lib\site-packages\snowflake\snowpark_internal\server_connection.py:546, in ServerConnection.get_result_set(self, plan, to_pandas, to_iter, block, data_type, log_on_exception, casesensitive, **kwargs) 544 for holder, id in placeholders.items(): 545 final_query = finalquery.replace(holder, id) --> 546 result = self.run_query( 547 final_query, 548 to_pandas, 549 to_iter and (i == len(plan.queries) - 1), 550 is_ddl_on_temp_object=query.is_ddl_on_temp_object, 551 block=not is_last, 552 data_type=data_type, 553 async_job_plan=plan, 554 log_on_exception=log_on_exception, 555 case_sensitive=case_sensitive, 556 params=query.params, 557 **kwargs, 558 ) 559 placeholders[query.query_id_place_holder] = ( 560 result["sfqid"] if not is_last else result.query_id 561 ) 562 result_meta = self._cursor.description

File ~\AppData\Local\miniconda3\envs\snowpark-de-ml\lib\site-packages\snowflake\snowpark_internal\server_connection.py:102, in ServerConnection._Decorator.wrap_exception..wrap(*args, **kwargs) 98 raise SnowparkClientExceptionMessages.SERVER_SESSION_EXPIRED( 99 ex.cause 100 ) 101 except Exception as ex: --> 102 raise ex

File ~\AppData\Local\miniconda3\envs\snowpark-de-ml\lib\site-packages\snowflake\snowpark_internal\server_connection.py:96, in ServerConnection._Decorator.wrap_exception..wrap(*args, *kwargs) 94 raise SnowparkClientExceptionMessages.SERVER_SESSION_HAS_BEEN_CLOSED() 95 try: ---> 96 return func(args, **kwargs) 97 except ReauthenticationRequest as ex: 98 raise SnowparkClientExceptionMessages.SERVER_SESSION_EXPIRED( 99 ex.cause 100 )

File ~\AppData\Local\miniconda3\envs\snowpark-de-ml\lib\site-packages\snowflake\snowpark_internal\server_connection.py:366, in ServerConnection.run_query(self, query, to_pandas, to_iter, is_ddl_on_temp_object, block, data_type, async_job_plan, log_on_exception, case_sensitive, params, num_statements, **kwargs) 364 query_id_log = f" [queryID: {ex.sfqid}]" if hasattr(ex, "sfqid") else "" 365 logger.error(f"Failed to execute query{query_id_log} {query}\n{ex}") --> 366 raise ex 368 # fetch_pandas_all/batches() only works for SELECT statements 369 # We call fetchall() if fetch_pandas_all/batches() fails, 370 # because when the query plan has multiple queries, it will 371 # have non-select statements, and it shouldn't fail if the user 372 # calls to_pandas() to execute the query. 373 if block:

File ~\AppData\Local\miniconda3\envs\snowpark-de-ml\lib\site-packages\snowflake\snowpark_internal\server_connection.py:347, in ServerConnection.run_query(self, query, to_pandas, to_iter, is_ddl_on_temp_object, block, data_type, async_job_plan, log_on_exception, case_sensitive, params, num_statements, kwargs) 345 kwargs["_statement_params"]["SNOWPARK_SKIP_TXN_COMMIT_IN_DDL"] = True 346 if block: --> 347 results_cursor = self._cursor.execute(query, params=params, kwargs) 348 self.notify_query_listeners( 349 QueryRecord(results_cursor.sfqid, results_cursor.query) 350 ) 351 logger.debug(f"Execute query [queryID: {results_cursor.sfqid}] {query}")

File ~\AppData\Local\miniconda3\envs\snowpark-de-ml\lib\site-packages\snowflake\connector\cursor.py:908, in SnowflakeCursor.execute(self, command, params, _bind_stage, timeout, _exec_async, _no_retry, _do_reset, _put_callback, _put_azure_callback, _put_callback_output_stream, _get_callback, _get_azure_callback, _get_callback_output_stream, _show_progress_bar, _statement_params, _is_internal, _describe_only, _no_results, _is_put_get, _raise_put_get_error, _force_put_overwrite, _skip_upload_on_content_match, file_stream, num_statements) 904 is_integrity_error = ( 905 code == "100072" 906 ) # NULL result in a non-nullable column 907 error_class = IntegrityError if is_integrity_error else ProgrammingError --> 908 Error.errorhandler_wrapper(self.connection, self, error_class, errvalue) 909 return self

File ~\AppData\Local\miniconda3\envs\snowpark-de-ml\lib\site-packages\snowflake\connector\errors.py:290, in Error.errorhandler_wrapper(connection, cursor, error_class, error_value) 267 @staticmethod 268 def errorhandler_wrapper( 269 connection: SnowflakeConnection | None, (...) 272 error_value: dict[str, Any], 273 ) -> None: 274 """Error handler wrapper that calls the errorhandler method. 275 276 Args: (...) 287 exception to the first handler in that order. 288 """ --> 290 handed_over = Error.hand_to_other_handler( 291 connection, 292 cursor, 293 error_class, 294 error_value, 295 ) 296 if not handed_over: 297 raise Error.errorhandler_make_exception( 298 error_class, 299 error_value, 300 )

File ~\AppData\Local\miniconda3\envs\snowpark-de-ml\lib\site-packages\snowflake\connector\errors.py:345, in Error.hand_to_other_handler(connection, cursor, error_class, error_value) 343 if cursor is not None: 344 cursor.messages.append((error_class, error_value)) --> 345 cursor.errorhandler(connection, cursor, error_class, error_value) 346 return True 347 elif connection is not None:

File ~\AppData\Local\miniconda3\envs\snowpark-de-ml\lib\site-packages\snowflake\connector\errors.py:221, in Error.default_errorhandler(connection, cursor, error_class, error_value) 219 errno = error_value.get("errno") 220 done_format_msg = error_value.get("done_format_msg") --> 221 raise error_class( 222 msg=error_value.get("msg"), 223 errno=None if errno is None else int(errno), 224 sqlstate=error_value.get("sqlstate"), 225 sfqid=error_value.get("sfqid"), 226 query=error_value.get("query"), 227 done_format_msg=( 228 None if done_format_msg is None else bool(done_format_msg) 229 ), 230 connection=connection, 231 cursor=cursor, 232 )

SnowparkSQLException: (1304): 01af1869-0001-c222-0000-003a7b0759c1: 100357 (P0000): Python Interpreter Error: Traceback (most recent call last): File "_udf_code.py", line 51, in compute File "_udf_code.py", line 40, in wrapper File "C:\Users\aekhoury\AppData\Local\Temp\ipykernel_24052\1935627540.py", line 12, in predict_roi File "/usr/lib/python_udf/cc189aaa19e7dc666cf02176e2f22eea863747958e19ccd50dfa7f40bf5684dd/lib/python3.9/site-packages/joblib/numpy_pickle.py", line 658, in load obj = _unpickle(fobj, filename, mmap_mode) File "/usr/lib/python_udf/cc189aaa19e7dc666cf02176e2f22eea863747958e19ccd50dfa7f40bf5684dd/lib/python3.9/site-packages/joblib/numpy_pickle.py", line 577, in _unpickle obj = unpickler.load() File "/usr/lib/python_udf/cc189aaa19e7dc666cf02176e2f22eea863747958e19ccd50dfa7f40bf5684dd/lib/python3.9/pickle.py", line 1212, in load dispatchkey[0] File "/usr/lib/python_udf/cc189aaa19e7dc666cf02176e2f22eea863747958e19ccd50dfa7f40bf5684dd/lib/python3.9/pickle.py", line 1537, in load_stack_global self.append(self.find_class(module, name)) File "/usr/lib/python_udf/cc189aaa19e7dc666cf02176e2f22eea863747958e19ccd50dfa7f40bf5684dd/lib/python3.9/pickle.py", line 1581, in find_class return _getattribute(sys.modules[module], name)[0] File "/usr/lib/python_udf/cc189aaa19e7dc666cf02176e2f22eea863747958e19ccd50dfa7f40bf5684dd/lib/python3.9/pickle.py", line 331, in _getattribute raise AttributeError("Can't get attribute {!r} on {!r}" AttributeError: Can't get attribute '_passthrough_scorer' on <module 'sklearn.metrics._scorer' from '/usr/lib/python_udf/cc189aaa19e7dc666cf02176e2f22eea863747958e19ccd50dfa7f40bf5684dd/lib/python3.9/site-packages/sklearn/metrics/_scorer.py'> in function PREDICT_ROI with handler compute

iamontheinet commented 1 year ago

Hi @dhiegopiroto and @aelkhoury79!

I have pushed a fix. Can you please try it from scratch and see if it works for you both?

Thank you! Dash

sfc-gh-dpiroto commented 1 year ago

did not work for me, @iamontheinet

What I did (suggestion of Haowei) is modify both inference scripts to use sklearn version dynamically, since the module is saved using a new version and the guide is trying to use an old one.

Scalar UDF: session.add_packages('pandas','joblib',f'scikit-learn=={sklearn.__version__}')

Vectorized UDF session.add_packages('pandas','joblib','cachetools',f'scikit-learn=={sklearn.__version__}')

and then add into the imports the library import sklearn

Worked for me :)

aelkhoury79 commented 1 year ago

Hi @iamontheinet your fix seems to have worked for me after I deleted my conda env and starting over. Thanks a lot!

iamontheinet commented 1 year ago

Thanks for confirming @aelkhoury79!

@sfc-gh-dpiroto Did you pull the latest before retrying? See my commit from yesterday https://github.com/Snowflake-Labs/sfguide-getting-started-dataengineering-ml-snowpark-python/commit/57331d8ba33468705bc9117c30875f0ce8e6dd48

Closing this issue. Let me know if you run into anything else. Thanks!

Snowflake-Labs / sfguide-getting-started-dataengineering-ml-snowpark-python

Error Executing Inference #11

Name Version Build Channel