[testing, CI] fix coverage statistics issue caused by test_common.py tracer patching

sklearnex/tests/test_common.py

@@ -15,12 +15,15 @@
 # ==============================================================================
 
 import importlib.util
+import io
 import os
 import pathlib
 import pkgutil
 import re
 import sys
 import trace
+from contextlib import redirect_stdout
+from multiprocessing import Pipe, Process, get_context
 
 import pytest
 from sklearn.utils import all_estimators
@@ -225,23 +228,120 @@ def _commonpath(inp):
 _TRACE_BLOCK_LIST = _whitelist_to_blacklist()
 
 
+def sklearnex_trace(estimator, method):
+    """Generate a trace of all function calls in calling estimator.method.
+
+    Parameters
+    ----------
+    estimator : str
+        name of estimator which is a key from PATCHED_MODELS or SPECIAL_INSTANCES
+
+    method : str
+        name of estimator method which is to be traced and stored
+
+    Returns
+    -------
+    text: str
+        Returns a string output (captured stdout of a python Trace call). It is a
+        modified version to be more informative, completed by a monkeypatching
+        of trace._modname.
+    """
+    # get estimator
+    try:
+        est = PATCHED_MODELS[estimator]()
+    except KeyError:
+        est = SPECIAL_INSTANCES[estimator]
+
+    # get dataset
+    X, y = gen_dataset(est)[0]
+    # fit dataset if method does not contain 'fit'
+    if "fit" not in method:
+        est.fit(X, y)
+
+    # monkeypatch new modname for clearer info
+    orig_modname = trace._modname
+    try:
+        # initialize tracer to have a more verbose module naming
+        # this impacts ignoremods, but it is not used.
+        trace._modname = _fullpath
+        tracer = trace.Trace(
+            count=0,
+            trace=1,
+            ignoredirs=_TRACE_BLOCK_LIST,
+        )
+        # call trace on method with dataset
+        f = io.StringIO()
+        with redirect_stdout(f):
+            tracer.runfunc(call_method, est, method, X, y)
+        return f.getvalue()
+    finally:
+        trace._modname = orig_modname
+
+
+def _trace_daemon(pipe):
+    """function interface for the other process. Information
+    exchanged using a multiprocess.Pipe"""
+    # a sent value with inherent conversion to False will break
+    # the while loop and complete the function
+    while key := pipe.recv():
+        text = ""
+        try:
+            estimator, method = key
+            text = sklearnex_trace(estimator, method)
+        finally:
+            pipe.send(text)
+
+
+@pytest.fixture(scope="module")
+def isolated_trace():
+    """Generates a separate python process for isolated sklearnex traces.
+
+    It is a module scope fixture due to the overhead of importing all the
+    various dependencies and is done once before all the various tests.
+    Each test will first check a cached value, if not existent it will have
+    the waiting child process generate the trace and return the text for
+    caching on its behalf. The isolated process is stopped at test teardown.
+
+    Yields
+    -------
+    pipe_parent: multiprocess.Connection
+        one end of a duplex pipe to be used by other pytest fixtures for
+        communicating with the special isolated tracing python instance
+        for sklearnex estimators.
+    """
+    try:
+        # force use of 'spawn' to guarantee a clean python environment
+        # from possible coverage arc tracing
+        ctx = get_context("spawn")
+        pipe_parent, pipe_child = ctx.Pipe()
+        p = ctx.Process(target=_trace_daemon, args=(pipe_child,), daemon=True)
+        p.start()
+        yield pipe_parent
+    finally:
+        # guarantee closing of the process via a try-catch-finally
+        # passing False terminates _trace_daemon's loop
+        pipe_parent.send(False)
+        pipe_parent.close()
+        pipe_child.close()
+        p.join()
+        p.close()
+
+
 @pytest.fixture
-def estimator_trace(estimator, method, cache, capsys, monkeypatch):
-    """Generate a trace of all function calls in calling estimator.method with cache.
+def estimator_trace(estimator, method, cache, isolated_trace):
+    """Create cache of all function calls in calling estimator.method.
 
     Parameters
     ----------
     estimator : str
-        name of estimator which is a key from PATCHED_MODELS or
+        name of estimator which is a key from PATCHED_MODELS or SPECIAL_INSTANCES
 
     method : str
         name of estimator method which is to be traced and stored
 
     cache: pytest.fixture (standard)
 
-    capsys: pytest.fixture (standard)
-
-    monkeypatch: pytest.fixture (standard)
+    isolated_trace: pytest.fixture (test_common.py)
 
     Returns
     -------
@@ -256,31 +356,12 @@ def estimator_trace(estimator, method, cache, capsys, monkeypatch):
     key = "-".join((str(estimator), method))
     flag = cache.get("key", "") != key
     if flag:
-        # get estimator
-        try:
-            est = PATCHED_MODELS[estimator]()
-        except KeyError:
-            est = SPECIAL_INSTANCES[estimator]
-
-        # get dataset
-        X, y = gen_dataset(est)[0]
-        # fit dataset if method does not contain 'fit'
-        if "fit" not in method:
-            est.fit(X, y)
 
-        # initialize tracer to have a more verbose module naming
-        # this impacts ignoremods, but it is not used.
-        monkeypatch.setattr(trace, "_modname", _fullpath)
-        tracer = trace.Trace(
-            count=0,
-            trace=1,
-            ignoredirs=_TRACE_BLOCK_LIST,
-        )
-        # call trace on method with dataset
-        tracer.runfunc(call_method, est, method, X, y)
+        isolated_trace.send((estimator, method))
+        text = isolated_trace.recv()
+        # provide a minimal error in the case that the tracing doesn't function
+        assert text, f"trace_daemon failure for {estimator}.{method}"
 
-        # collect trace for analysis
-        text = capsys.readouterr().out
         for modulename, file in _TRACE_ALLOW_DICT.items():
             text = text.replace(file, modulename)
         regex_func = (