Avoid invalid escape sequences, use raw strings (#22936)

* Avoid invalid escape sequences, use raw strings

* Integrate PR feedback

utils/check_config_docstrings.py

@@ -31,7 +31,7 @@ CONFIG_MAPPING = transformers.models.auto.configuration_auto.CONFIG_MAPPING
 
 # Regex pattern used to find the checkpoint mentioned in the docstring of `config_class`.
 # For example, `[bert-base-uncased](https://huggingface.co/bert-base-uncased)`
-_re_checkpoint = re.compile("\[(.+?)\]\((https://huggingface\.co/.+?)\)")
+_re_checkpoint = re.compile(r"\[(.+?)\]\((https://huggingface\.co/.+?)\)")
 
 
 CONFIG_CLASSES_TO_IGNORE_FOR_DOCSTRING_CHECKPOINT_CHECK = {

utils/check_inits.py

@@ -35,9 +35,9 @@ _re_import_struct_add_one = re.compile(r'^\s*_import_structure\["\S*"\]\.append\
 # Catches a line _import_struct["bla"].extend(["foo", "bar"]) or _import_struct["bla"] = ["foo", "bar"]
 _re_import_struct_add_many = re.compile(r"^\s*_import_structure\[\S*\](?:\.extend\(|\s*=\s+)\[([^\]]*)\]")
 # Catches a line with an object between quotes and a comma:     "MyModel",
-_re_quote_object = re.compile('^\s+"([^"]+)",')
+_re_quote_object = re.compile(r'^\s+"([^"]+)",')
 # Catches a line with objects between brackets only:    ["foo", "bar"],
-_re_between_brackets = re.compile("^\s+\[([^\]]+)\]")
+_re_between_brackets = re.compile(r"^\s+\[([^\]]+)\]")
 # Catches a line with from foo import bar, bla, boo
 _re_import = re.compile(r"\s+from\s+\S*\s+import\s+([^\(\s].*)\n")
 # Catches a line with try:
@@ -78,7 +78,7 @@ def parse_init(init_file):
         # If we have everything on a single line, let's deal with it.
         if _re_one_line_import_struct.search(line):
             content = _re_one_line_import_struct.search(line).groups()[0]
-            imports = re.findall("\[([^\]]+)\]", content)
+            imports = re.findall(r"\[([^\]]+)\]", content)
             for imp in imports:
                 objects.extend([obj[1:-1] for obj in imp.split(", ")])
             line_index += 1

utils/check_repo.py

@@ -755,7 +755,7 @@ def find_all_documented_objects():
     for doc_file in Path(PATH_TO_DOC).glob("**/*.mdx"):
         with open(doc_file, "r", encoding="utf-8", newline="\n") as f:
             content = f.read()
-        raw_doc_objs = re.findall("\[\[autodoc\]\]\s+(\S+)\s+", content)
+        raw_doc_objs = re.findall(r"\[\[autodoc\]\]\s+(\S+)\s+", content)
         documented_obj += [obj.split(".")[-1] for obj in raw_doc_objs]
     return documented_obj
 

utils/notification_service_doc_tests.py

@@ -52,7 +52,7 @@ def extract_first_line_failure(failures_short_lines):
     file = None
     in_error = False
     for line in failures_short_lines.split("\n"):
-        if re.search("_ \[doctest\]", line):
+        if re.search(r"_ \[doctest\]", line):
             in_error = True
             file = line.split(" ")[2]
         elif in_error and not line.split(" ")[0].isdigit():

utils/sort_auto_mappings.py

@@ -23,7 +23,7 @@ PATH_TO_AUTO_MODULE = "src/transformers/models/auto"
 
 # re pattern that matches mapping introductions:
 #    SUPER_MODEL_MAPPING_NAMES = OrderedDict or SUPER_MODEL_MAPPING = OrderedDict
-_re_intro_mapping = re.compile("[A-Z_]+_MAPPING(\s+|_[A-Z_]+\s+)=\s+OrderedDict")
+_re_intro_mapping = re.compile(r"[A-Z_]+_MAPPING(\s+|_[A-Z_]+\s+)=\s+OrderedDict")
 # re pattern that matches identifiers in mappings
 _re_identifier = re.compile(r'\s*\(\s*"(\S[^"]+)"')