nannyml: post-deployment data science in python
APACHE-2.0 License
Bot releases are visible (Hide)
Published by github-actions[bot] about 2 years ago
problem_type
parameter is required during calculator/estimator initialization, thistox
running speed by skipping some unnecessary package installations.Reader
instance would raise a WriteException
. Thanks for those eagle eyesPublished by github-actions[bot] about 2 years ago
Chunk
size to estimating the sampling error for an operation on a Chunk
.
Chunk
in the result data forPublished by github-actions[bot] about 2 years ago
Published by github-actions[bot] about 2 years ago
make clean
now also clears __pycache__
Published by github-actions[bot] over 2 years ago
Published by github-actions[bot] over 2 years ago
regression
use cases: create or extract RegressionMetadata
and use it for driftDefaultChunker
splits into 10 chunks of equal size.SizeBasedChunker
no longer drops incomplete last chunk by default, but this is now configurable behavior.Published by github-actions[bot] over 2 years ago
roc_auc
,f1
, precision
, recall
, specificity
and accuracy
.multiclass classification metadata
+ support in automated metadata extraction (by introducing amodel_type
parameter).CBPE
metrics.PerformanceCalculator
.identifier
property from the ModelMetadata
class. Joining analysis
data andanalysis target
values should be done upfront or index-based.exclude_columns
parameter to the extract_metadata
function. Use it to specify the columns that shouldfit
methods now return the fitted object. This allows chaining Calculator
/Estimator
instantiationPerformanceCalculator
. Only the predefined metrics remain supported.CBPE
step plotsPublished by github-actions[bot] over 2 years ago
Published by github-actions[bot] over 2 years ago
Published by github-actions[bot] over 2 years ago
ModelMetadata
.PerformanceCalculator
.TargetDistributionCalculator
nannyml.drift
package and subpackages. Breaking changes!FeatureType.UNKNOWN
.PeriodBasedChunker
, using theblack
pre-commit hook due to breaking changes in its click
dependency.calculator
/ estimator
/ metric
,scipy
dependency to be >=1.7.3, <1.8.0
. Planned to be relaxed ASAP.NaN
values when concatenating.ModelMetadata
printout