python/.coveragerc

# Coverage scope for the pyspark drop-in's hand-written Python shim.
# The bulk of the client is the compiled Rust extension (_pyspark); this measures
# the Python glue only, excluding vendored third-party code (cloudpickle), the
# test-support utilities, pandas-on-spark, generated wrappers, and the compiled
# extension itself.
[run]
source = pyspark
branch = True
# The badge scopes to the core DataFrame/SQL drop-in shim exercised by the
# offline suite. Vendored third-party code (cloudpickle), generated wrappers,
# the compiled extension, and subsystems with their own (server-backed) tests
# rather than offline unit tests (streaming, resource, ml, the pickling
# serializers, datasource) are omitted so the number reflects the code this
# suite actually drives.
omit =
    # Vendored upstream code (byte-identical to Apache Spark, enforced by
    # scripts/check_vendored_upstream.sh; covered by Apache's own tests and by our dedicated
    # pandas/pipelines parity workflows, not this offline shim suite). Keep in sync with that
    # script's VENDORED_DIRS/VENDORED_FILES. What remains measured is the hand-written drop-in
    # glue (pyspark/sql/*.py shims and pyspark/sql/connect/* shims).
    */cloudpickle/*
    */pyspark/testing/*
    */pyspark/pandas/*
    */pyspark/ml/*
    */pyspark/mllib/*
    */pyspark/resource/*
    */pyspark/streaming/*
    */pyspark/pipelines/*
    */pyspark/logger/*
    */pyspark/errors/*
    */pyspark/sql/streaming/*
    */pyspark/sql/worker/*
    */pyspark/sql/plot/*
    */pyspark/sql/pandas/*
    */pyspark/sql/functions_generated.py
    */pyspark/sql/datasource.py
    */pyspark/sql/conversion.py
    */pyspark/sql/internal.py
    */pyspark/sql/variant_utils.py
    */pyspark/sql/datasource_internal.py
    */pyspark/sql/profiler.py
    */pyspark/serializers.py
    */pyspark/loose_version.py
    */pyspark/_globals.py
    # Classic-Spark / JVM / launcher utilities that a Connect-only client never uses and
    # cannot exercise offline (SparkContext helpers, gateway launch, local-install discovery).
    */pyspark/util.py
    */pyspark/sql/utils.py
    */pyspark/find_spark_home.py
    */pyspark/version.py
    */pyspark/accumulators.py
    */pyspark/profiler.py
    */pyspark/memory_profiler_ext.py
    */pyspark/worker_util.py
    */pyspark/instrumentation_utils.py
    */pyspark/_pyspark*
    */tests/*

[report]
exclude_lines =
    pragma: no cover
    raise NotImplementedError
    raise PySparkNotImplementedError
    if __name__ == .__main__.:
    # Never executed at runtime: type-checking-only imports and optional-dependency guards.
    if TYPE_CHECKING:
    except ImportError: