Fix false positive unspecified-encoding when using kwargs (#8728)

False positives were being generated when passing arguments as
kwargs to open() and other IO calls. This has been fixed by
using inference whenever the argument was not found through
previously existing methods (position and keyword) and
kwargs are present.

The confidence levels for both methods with/without inference
have also been updated.

Closes #8719

Author: crazybolillo

doc/whatsnew/fragments/8719.false_positive

@@ -0,0 +1,3 @@
+Fix false positives generated when supplying arguments as ``**kwargs`` to IO calls like open().
+
+Closes #8719

pylint/checkers/stdlib.py

@@ -16,7 +16,7 @@
 
 from pylint import interfaces
 from pylint.checkers import BaseChecker, DeprecatedMixin, utils
-from pylint.interfaces import INFERENCE
+from pylint.interfaces import HIGH, INFERENCE
 from pylint.typing import MessageDefinitionTuple
 
 if TYPE_CHECKING:
@@ -699,6 +699,7 @@ def _check_open_call(
     ) -> None:
         """Various checks for an open call."""
         mode_arg = None
+        confidence = HIGH
         try:
             if open_module == "_io":
                 mode_arg = utils.get_argument_from_call(
@@ -709,11 +710,12 @@ def _check_open_call(
                     node, position=0, keyword="mode"
                 )
         except utils.NoSuchArgumentError:
-            pass
+            mode_arg = utils.infer_kwarg_from_call(node, keyword="mode")
+            if mode_arg:
+                confidence = INFERENCE
 
         if mode_arg:
             mode_arg = utils.safe_infer(mode_arg)
-
             if (
                 func_name in OPEN_FILES_MODE
                 and isinstance(mode_arg, nodes.Const)
@@ -723,14 +725,15 @@ def _check_open_call(
                     "bad-open-mode",
                     node=node,
                     args=mode_arg.value or str(mode_arg.value),
+                    confidence=confidence,
                 )
 
         if (
             not mode_arg
             or isinstance(mode_arg, nodes.Const)
             and (not mode_arg.value or "b" not in str(mode_arg.value))
         ):
-            encoding_arg = None
+            confidence = HIGH
             try:
                 if open_module == "pathlib":
                     if node.func.attrname == "read_text":
@@ -750,13 +753,21 @@ def _check_open_call(
                         node, position=3, keyword="encoding"
                     )
             except utils.NoSuchArgumentError:
-                self.add_message("unspecified-encoding", node=node)
+                encoding_arg = utils.infer_kwarg_from_call(node, keyword="encoding")
+                if encoding_arg:
+                    confidence = INFERENCE
+                else:
+                    self.add_message(
+                        "unspecified-encoding", node=node, confidence=confidence
+                    )
 
             if encoding_arg:
                 encoding_arg = utils.safe_infer(encoding_arg)
 
                 if isinstance(encoding_arg, nodes.Const) and encoding_arg.value is None:
-                    self.add_message("unspecified-encoding", node=node)
+                    self.add_message(
+                        "unspecified-encoding", node=node, confidence=confidence
+                    )
 
     def _check_env_function(self, node: nodes.Call, infer: nodes.FunctionDef) -> None:
         env_name_kwarg = "key"

pylint/checkers/utils.py

@@ -736,6 +736,25 @@ def get_argument_from_call(
     raise NoSuchArgumentError
 
 
+def infer_kwarg_from_call(call_node: nodes.Call, keyword: str) -> nodes.Name | None:
+    """Returns the specified argument from a function's kwargs.
+
+    :param nodes.Call call_node: Node representing a function call to check.
+    :param str keyword: Name of the argument to be extracted.
+
+    :returns: The node representing the argument, None if the argument is not found.
+    :rtype: nodes.Name
+    """
+    for arg in call_node.kwargs:
+        inferred = safe_infer(arg.value)
+        if isinstance(inferred, nodes.Dict):
+            for item in inferred.items:
+                if item[0].value == keyword:
+                    return item[1]
+
+    return None
+
+
 def inherit_from_std_ex(node: nodes.NodeNG | astroid.Instance) -> bool:
     """Return whether the given class node is subclass of
     exceptions.Exception.

tests/functional/b/bad_open_mode.txt

@@ -1,6 +1,6 @@
-bad-open-mode:12:0:12:35::"""rwx"" is not a valid mode for open.":UNDEFINED
-bad-open-mode:13:0:13:34::"""rr"" is not a valid mode for open.":UNDEFINED
-bad-open-mode:14:0:14:33::"""+"" is not a valid mode for open.":UNDEFINED
-bad-open-mode:15:0:15:34::"""xw"" is not a valid mode for open.":UNDEFINED
-bad-open-mode:21:0:21:34::"""Ua"" is not a valid mode for open.":UNDEFINED
-bad-open-mode:22:0:22:36::"""Ur++"" is not a valid mode for open.":UNDEFINED
+bad-open-mode:12:0:12:35::"""rwx"" is not a valid mode for open.":HIGH
+bad-open-mode:13:0:13:34::"""rr"" is not a valid mode for open.":HIGH
+bad-open-mode:14:0:14:33::"""+"" is not a valid mode for open.":HIGH
+bad-open-mode:15:0:15:34::"""xw"" is not a valid mode for open.":HIGH
+bad-open-mode:21:0:21:34::"""Ua"" is not a valid mode for open.":HIGH
+bad-open-mode:22:0:22:36::"""Ur++"" is not a valid mode for open.":HIGH

tests/functional/u/unspecified_encoding_py38.py

@@ -162,3 +162,35 @@ class IOArgs:
 
 # Test for crash reported in https://github.com/pylint-dev/pylint/issues/6414
 open('foo', mode=2)  # [bad-open-mode, unspecified-encoding]
+
+# Infer kwargs
+KWARGS = {"mode": "rb"}
+open(FILENAME, **KWARGS)
+
+KWARGS = {"mode": "w", "encoding": "utf-8"}
+open(FILENAME, **KWARGS)
+io.open(FILENAME, **KWARGS)
+Path(FILENAME).open(**KWARGS)
+
+KWARGS = {"mode": 5}
+open(FILENAME, **KWARGS)  # [bad-open-mode, unspecified-encoding]
+io.open(FILENAME, **KWARGS)  # [bad-open-mode, unspecified-encoding]
+
+KWARGS = {"mode": "wt", "encoding": None}
+with open(FILENAME, **KWARGS) as fd:  # [unspecified-encoding]
+    pass
+
+Path(FILENAME).open(**KWARGS)  # [unspecified-encoding]
+
+KWARGS = {"encoding": None}
+Path(FILENAME).write_text("hello", **KWARGS)  # [unspecified-encoding]
+
+KWARGS = {"encoding": "utf-8"}
+Path(FILENAME).write_text("goodbye", **KWARGS)
+
+# No one does it this way, but it is a possibility
+KWARGS = {"encoding": None}
+Path(FILENAME).read_text(**KWARGS)  # [unspecified-encoding]
+
+KWARGS = {"encoding": "utf-8"}
+Path(FILENAME).read_text(**KWARGS)

tests/functional/u/unspecified_encoding_py38.txt

@@ -1,33 +1,41 @@
-unspecified-encoding:13:0:13:14::Using open without explicitly specifying an encoding:UNDEFINED
-unspecified-encoding:14:0:14:20::Using open without explicitly specifying an encoding:UNDEFINED
-unspecified-encoding:15:0:15:20::Using open without explicitly specifying an encoding:UNDEFINED
-unspecified-encoding:16:0:16:34::Using open without explicitly specifying an encoding:UNDEFINED
-unspecified-encoding:17:0:17:19::Using open without explicitly specifying an encoding:UNDEFINED
-unspecified-encoding:26:5:26:19::Using open without explicitly specifying an encoding:UNDEFINED
-unspecified-encoding:29:5:29:34::Using open without explicitly specifying an encoding:UNDEFINED
-unspecified-encoding:33:5:33:45::Using open without explicitly specifying an encoding:UNDEFINED
-unspecified-encoding:38:0:38:17::Using open without explicitly specifying an encoding:UNDEFINED
-unspecified-encoding:39:0:39:23::Using open without explicitly specifying an encoding:UNDEFINED
-unspecified-encoding:40:0:40:23::Using open without explicitly specifying an encoding:UNDEFINED
-unspecified-encoding:41:0:41:37::Using open without explicitly specifying an encoding:UNDEFINED
-unspecified-encoding:50:5:50:22::Using open without explicitly specifying an encoding:UNDEFINED
-unspecified-encoding:53:5:53:37::Using open without explicitly specifying an encoding:UNDEFINED
-unspecified-encoding:57:5:57:48::Using open without explicitly specifying an encoding:UNDEFINED
-unspecified-encoding:66:0:66:26::Using open without explicitly specifying an encoding:UNDEFINED
-unspecified-encoding:67:0:67:39::Using open without explicitly specifying an encoding:UNDEFINED
-unspecified-encoding:68:0:68:50::Using open without explicitly specifying an encoding:UNDEFINED
-unspecified-encoding:75:0:75:35::Using open without explicitly specifying an encoding:UNDEFINED
-unspecified-encoding:76:0:76:50::Using open without explicitly specifying an encoding:UNDEFINED
-unspecified-encoding:77:0:77:61::Using open without explicitly specifying an encoding:UNDEFINED
-unspecified-encoding:81:0:81:21::Using open without explicitly specifying an encoding:UNDEFINED
-unspecified-encoding:82:0:82:25::Using open without explicitly specifying an encoding:UNDEFINED
-unspecified-encoding:83:0:83:25::Using open without explicitly specifying an encoding:UNDEFINED
-unspecified-encoding:84:0:84:39::Using open without explicitly specifying an encoding:UNDEFINED
-unspecified-encoding:149:0:149:23::Using open without explicitly specifying an encoding:UNDEFINED
-unspecified-encoding:152:0:152:28::Using open without explicitly specifying an encoding:UNDEFINED
-unspecified-encoding:155:0:155:26::Using open without explicitly specifying an encoding:UNDEFINED
-unspecified-encoding:158:0:158:35::Using open without explicitly specifying an encoding:UNDEFINED
-bad-open-mode:161:0:161:25::"""None"" is not a valid mode for open.":UNDEFINED
-unspecified-encoding:161:0:161:25::Using open without explicitly specifying an encoding:UNDEFINED
-bad-open-mode:164:0:164:19::"""2"" is not a valid mode for open.":UNDEFINED
-unspecified-encoding:164:0:164:19::Using open without explicitly specifying an encoding:UNDEFINED
+unspecified-encoding:13:0:13:14::Using open without explicitly specifying an encoding:HIGH
+unspecified-encoding:14:0:14:20::Using open without explicitly specifying an encoding:HIGH
+unspecified-encoding:15:0:15:20::Using open without explicitly specifying an encoding:HIGH
+unspecified-encoding:16:0:16:34::Using open without explicitly specifying an encoding:HIGH
+unspecified-encoding:17:0:17:19::Using open without explicitly specifying an encoding:HIGH
+unspecified-encoding:26:5:26:19::Using open without explicitly specifying an encoding:HIGH
+unspecified-encoding:29:5:29:34::Using open without explicitly specifying an encoding:HIGH
+unspecified-encoding:33:5:33:45::Using open without explicitly specifying an encoding:HIGH
+unspecified-encoding:38:0:38:17::Using open without explicitly specifying an encoding:HIGH
+unspecified-encoding:39:0:39:23::Using open without explicitly specifying an encoding:HIGH
+unspecified-encoding:40:0:40:23::Using open without explicitly specifying an encoding:HIGH
+unspecified-encoding:41:0:41:37::Using open without explicitly specifying an encoding:HIGH
+unspecified-encoding:50:5:50:22::Using open without explicitly specifying an encoding:HIGH
+unspecified-encoding:53:5:53:37::Using open without explicitly specifying an encoding:HIGH
+unspecified-encoding:57:5:57:48::Using open without explicitly specifying an encoding:HIGH
+unspecified-encoding:66:0:66:26::Using open without explicitly specifying an encoding:HIGH
+unspecified-encoding:67:0:67:39::Using open without explicitly specifying an encoding:HIGH
+unspecified-encoding:68:0:68:50::Using open without explicitly specifying an encoding:HIGH
+unspecified-encoding:75:0:75:35::Using open without explicitly specifying an encoding:HIGH
+unspecified-encoding:76:0:76:50::Using open without explicitly specifying an encoding:HIGH
+unspecified-encoding:77:0:77:61::Using open without explicitly specifying an encoding:HIGH
+unspecified-encoding:81:0:81:21::Using open without explicitly specifying an encoding:HIGH
+unspecified-encoding:82:0:82:25::Using open without explicitly specifying an encoding:HIGH
+unspecified-encoding:83:0:83:25::Using open without explicitly specifying an encoding:HIGH
+unspecified-encoding:84:0:84:39::Using open without explicitly specifying an encoding:HIGH
+unspecified-encoding:149:0:149:23::Using open without explicitly specifying an encoding:HIGH
+unspecified-encoding:152:0:152:28::Using open without explicitly specifying an encoding:HIGH
+unspecified-encoding:155:0:155:26::Using open without explicitly specifying an encoding:HIGH
+unspecified-encoding:158:0:158:35::Using open without explicitly specifying an encoding:HIGH
+bad-open-mode:161:0:161:25::"""None"" is not a valid mode for open.":HIGH
+unspecified-encoding:161:0:161:25::Using open without explicitly specifying an encoding:HIGH
+bad-open-mode:164:0:164:19::"""2"" is not a valid mode for open.":HIGH
+unspecified-encoding:164:0:164:19::Using open without explicitly specifying an encoding:HIGH
+bad-open-mode:176:0:176:24::"""5"" is not a valid mode for open.":INFERENCE
+unspecified-encoding:176:0:176:24::Using open without explicitly specifying an encoding:HIGH
+bad-open-mode:177:0:177:27::"""5"" is not a valid mode for open.":INFERENCE
+unspecified-encoding:177:0:177:27::Using open without explicitly specifying an encoding:HIGH
+unspecified-encoding:180:5:180:29::Using open without explicitly specifying an encoding:INFERENCE
+unspecified-encoding:183:0:183:29::Using open without explicitly specifying an encoding:INFERENCE
+unspecified-encoding:186:0:186:44::Using open without explicitly specifying an encoding:INFERENCE
+unspecified-encoding:193:0:193:34::Using open without explicitly specifying an encoding:INFERENCE