errors

errors

Custom warnings and exceptions for ts-shape.

Follows the pandas/scikit-learn pattern: - warnings.warn() for user-facing feedback (visible by default). - logging for internal diagnostics (silent unless configured).

Users can filter specific categories, e.g.::

import warnings
from ts_shape.errors import PerformanceWarning
warnings.filterwarnings("ignore", category=PerformanceWarning)

TsShapeWarning

Bases: UserWarning

Base warning for all ts-shape warnings.

PerformanceWarning

Bases: TsShapeWarning

Warn when an operation may be slow due to data size or shape.

DataQualityWarning

Bases: TsShapeWarning

Warn about potential data quality issues (gaps, duplicates, NaNs).

LoaderConfigWarning

Bases: TsShapeWarning

Warn when a loader returns no data, likely due to misconfiguration.

ColumnNotFoundError

Bases: ValueError

Raised when a required column is missing from the DataFrame.

Subclasses ValueError so existing except ValueError handlers keep working, while callers that want to react specifically to a missing column can catch this narrower type.

LoaderError

Bases: Exception

Raised when a loader cannot read from its configured source.

Covers misconfiguration (missing path, absent credentials) and I/O that fails after the configured number of retries. Callers can catch this single type regardless of the underlying backend (local, S3, Azure, database).