Splunk Enterprise - 9.3.2
Splunk_ML_Toolkit: 5.4.2
PSC - 4.2.1
After deploying the custom Python package containing TestAlgo.py, the algorithm fails to initialize. I have confirmed installation of all required Python modules (pandas, scikit-learn, numpy, scipy) and verified directory ownership and the rebuilt PYTHONPATH. I suspect a persistent issue with Python path injection or a deeper structural configuration conflict.
Traceback (most recent call last):
File "/datastore/splunk/etc/apps/Splunk_ML_Toolkit/bin/chunked_controller.py", line 127, in initialize_processor
processor = processor_class(process_options, searchinfo)
File "/datastore/splunk/etc/apps/Splunk_ML_Toolkit/bin/processors/FitBatchProcessor.py", line 51, in __init__
self.algo = self.initialize_algo(self.algo_options, self.searchinfo)
File "/datastore/splunk/etc/apps/Splunk_ML_Toolkit/bin/processors/FitBatchProcessor.py", line 72, in initialize_algo
raise RuntimeError(
RuntimeError: Error while initializing algorithm "TestAlgo": TheTestAlgo algorithm cannot be initialized.
Could you help me why the algorithm is not initialized properly?
test_all_ml/local/algos.conf
[TestAlgo]
package=test_algos
disabled=false
test_all_ml/bin/test_algos/TestAlgo.py
from base import BaseAlgo
class TestAlgo(BaseAlgo):
def __init__(self, options, **kwargs):
super(TestAlgo, self).__init__(options, **kwargs)
self.model = None
def fit(self, df, args):
raise NotImplementedError("FIT SUCCESSFUL - LOGIC TEST")
def apply(self, df, args):
raise NotImplementedError("APPLY SUCCESSFUL - LOGIC TEST")
