Move utils.expand_module in its own file in lint

We never used it anywhere else than lint so it belong in lint.

3 files changed, 147 insertions, 3 deletions

diff --git a/pylint/lint/expand_modules.py b/pylint/lint/expand_modules.py
new file mode 100644
index 000000000..dc52811b5
--- /dev/null
+++ b/pylint/lint/expand_modules.py
@@ -0,0 +1,143 @@
+import os
+import sys
+
+from astroid import modutils
+
+
+def _modpath_from_file(filename, is_namespace, path=None):
+    def _is_package_cb(path, parts):
+        return modutils.check_modpath_has_init(path, parts) or is_namespace
+
+    return modutils.modpath_from_file_with_callback(
+        filename, path=path, is_package_cb=_is_package_cb
+    )
+
+
+def get_python_path(filepath: str) -> str:
+    dirname = os.path.realpath(os.path.expanduser(filepath))
+    if not os.path.isdir(dirname):
+        dirname = os.path.dirname(dirname)
+    while True:
+        if not os.path.exists(os.path.join(dirname, "__init__.py")):
+            return dirname
+        old_dirname = dirname
+        dirname = os.path.dirname(dirname)
+        if old_dirname == dirname:
+            return os.getcwd()
+    return None
+
+
+def _basename_in_ignore_list_re(base_name, ignore_list_re):
+    """Determines if the basename is matched in a regex ignorelist
+
+    :param str base_name: The basename of the file
+    :param list ignore_list_re: A collection of regex patterns to match against.
+        Successful matches are ignored.
+
+    :returns: `True` if the basename is ignored, `False` otherwise.
+    :rtype: bool
+    """
+    for file_pattern in ignore_list_re:
+        if file_pattern.match(base_name):
+            return True
+    return False
+
+
+def expand_modules(files_or_modules, ignore_list, ignore_list_re):
+    """take a list of files/modules/packages and return the list of tuple
+    (file, module name) which have to be actually checked
+    """
+    result = []
+    errors = []
+    path = sys.path.copy()
+
+    for something in files_or_modules:
+        if os.path.basename(something) in ignore_list:
+            continue
+        if _basename_in_ignore_list_re(os.path.basename(something), ignore_list_re):
+            continue
+
+        module_path = get_python_path(something)
+        additional_search_path = [".", module_path] + path
+        if os.path.exists(something):
+            # this is a file or a directory
+            try:
+                modname = ".".join(
+                    modutils.modpath_from_file(something, path=additional_search_path)
+                )
+            except ImportError:
+                modname = os.path.splitext(os.path.basename(something))[0]
+            if os.path.isdir(something):
+                filepath = os.path.join(something, "__init__.py")
+            else:
+                filepath = something
+        else:
+            # suppose it's a module or package
+            modname = something
+            try:
+                filepath = modutils.file_from_modpath(
+                    modname.split("."), path=additional_search_path
+                )
+                if filepath is None:
+                    continue
+            except (ImportError, SyntaxError) as ex:
+                # The SyntaxError is a Python bug and should be
+                # removed once we move away from imp.find_module: https://bugs.python.org/issue10588
+                errors.append({"key": "fatal", "mod": modname, "ex": ex})
+                continue
+
+        filepath = os.path.normpath(filepath)
+        modparts = (modname or something).split(".")
+
+        try:
+            spec = modutils.file_info_from_modpath(
+                modparts, path=additional_search_path
+            )
+        except ImportError:
+            # Might not be acceptable, don't crash.
+            is_namespace = False
+            is_directory = os.path.isdir(something)
+        else:
+            is_namespace = modutils.is_namespace(spec)
+            is_directory = modutils.is_directory(spec)
+
+        if not is_namespace:
+            result.append(
+                {
+                    "path": filepath,
+                    "name": modname,
+                    "isarg": True,
+                    "basepath": filepath,
+                    "basename": modname,
+                }
+            )
+
+        has_init = (
+            not (modname.endswith(".__init__") or modname == "__init__")
+            and os.path.basename(filepath) == "__init__.py"
+        )
+        if has_init or is_namespace or is_directory:
+            for subfilepath in modutils.get_module_files(
+                os.path.dirname(filepath), ignore_list, list_all=is_namespace
+            ):
+                if filepath == subfilepath:
+                    continue
+                if _basename_in_ignore_list_re(
+                    os.path.basename(subfilepath), ignore_list_re
+                ):
+                    continue
+
+                modpath = _modpath_from_file(
+                    subfilepath, is_namespace, path=additional_search_path
+                )
+                submodname = ".".join(modpath)
+                result.append(
+                    {
+                        "path": subfilepath,
+                        "name": submodname,
+                        "isarg": False,
+                        "basepath": filepath,
+                        "basename": modname,
+                    }
+                )
+    return result, errors
diff --git a/pylint/lint/pylinter.py b/pylint/lint/pylinter.py
index 5a662a079..e199b5437 100644
--- a/pylint/lint/pylinter.py
+++ b/pylint/lint/pylinter.py
@@ -16,6 +16,7 @@ import astroid
 
 from pylint import checkers, config, exceptions, interfaces, reporters
 from pylint.constants import MAIN_CHECKER_NAME, MSG_TYPES
+from pylint.lint.expand_modules import expand_modules
 from pylint.lint.parallel import check_parallel
 from pylint.lint.report_functions import (
     report_messages_by_module_stats,
@@ -970,7 +971,7 @@ class PyLinter(
 
     def _expand_files(self, modules):
         """get modules and errors from a list of modules and handle errors"""
-        result, errors = utils.expand_modules(
+        result, errors = expand_modules(
             modules, self.config.black_list, self.config.black_list_re
         )
         for error in errors:
diff --git a/pylint/lint/utils.py b/pylint/lint/utils.py
index 1f4d58bfe..3d17f3b38 100644
--- a/pylint/lint/utils.py
+++ b/pylint/lint/utils.py
@@ -4,7 +4,7 @@
 import contextlib
 import sys
 
-from pylint.utils import utils
+from pylint.lint.expand_modules import get_python_path
 
 
 class ArgumentPreprocessingError(Exception):
@@ -51,7 +51,7 @@ def _patch_sys_path(args):
     changes = []
     seen = set()
     for arg in args:
-        path = utils.get_python_path(arg)
+        path = get_python_path(arg)
         if path not in seen:
             changes.append(path)
             seen.add(path)