Refactored the code to use a generator

Author: MomIsBestFriend

ci/code_checks.sh

@@ -217,7 +217,7 @@ if [[ -z "$CHECK" || "$CHECK" == "patterns" ]]; then
  RET=$(($RET + $?)) ; echo $MSG "DONE"
 
  MSG='Check for use of not concatenated strings' ; echo $MSG
- python $BASE_DIR/scripts/validate_string_concatenation.py pandas
+ $BASE_DIR/scripts/validate_string_concatenation.py pandas
  RET=$(($RET + $?)) ; echo $MSG "DONE"
 
  MSG='Check that no file in the repo contains trailing whitespaces' ; echo $MSG

scripts/validate_string_concatenation.py

@@ -12,13 +12,14 @@
 ... "baz"
 ... )
 
+
 into this:
 
 >>> foo = ("bar " "baz")
 
 Black is not considering this as an
-issue (see https://github.com/psf/black/issues/1051), so we are checking
-it here.
+issue (see issue https://github.com/psf/black/issues/1051),
+so we are checking it here.
 """
 
 import os
@@ -30,38 +31,31 @@
 FILE_EXTENSIONS_TO_CHECK = frozenset((".pxd", ".py", ".pyx", ".pyx.ini"))
 
 
-def is_concatenated(file_path):
+def strings_to_concatenate(file_path):
  """
- Checking if the file containing strings that needs to be concatenated.
+ Yielding the strings that needs to be concatenated in a given file.
 
  Parameters
  ----------
  file_path : str
  File path pointing to a single file.
 
- Returns
- -------
- int
- Status code representing if the file needs a fix.
- 0 - All good.
- 1 - Needs to be fixed.
+ Yields
+ ------
+ str
+ Message containing info about the string that needs to be concatenated.
  """
- need_fix = False
  with open(file_path, "r") as file_name:
  tokens = list(tokenize.generate_tokens(file_name.readline))
- for current_token, next_token in zip(tokens, tokens[1:]):
- if current_token[0] == next_token[0] == token.STRING:
- need_fix = True
- print(
- "{file_path}:{line_number}:\t{start} and {end}".format(
- file_path=file_path,
- line_number=current_token[2][0],
- start=current_token[1],
- end=next_token[1],
- )
- )
 
- return int(need_fix)
+ for current_token, next_token in zip(tokens, tokens[1:]):
+ if current_token[0] == next_token[0] == token.STRING:
+ yield "{file_path}:{line_number}:\t{start} and {end}\n".format(
+ file_path=file_path,
+ line_number=current_token[2][0],
+ start=current_token[1],
+ end=next_token[1],
+ )
 
 
 if __name__ == "__main__":
@@ -70,19 +64,23 @@ def is_concatenated(file_path):
  if not os.path.exists(path):
  raise ValueError("Please enter a valid path, to a file/directory.")
 
+ failed = False
+
  if os.path.isfile(path):
- # Means that the given path is of a single file.
- sys.exit(is_concatenated(path))
+ for msg in strings_to_concatenate(path):
+ if msg:
+ failed = True
+ print(msg)
 
- failures = 0
- # Means that the given path is of a directory.
  for subdir, _, files in os.walk(path):
  for file_name in files:
  if any(
  file_name.endswith(extension) for extension in FILE_EXTENSIONS_TO_CHECK
  ):
  file_extension = os.path.join(subdir, file_name)
- failures += is_concatenated(os.path.join(subdir, file_name))
 
- exit_code = 1 if failures >= 1 else 0
- sys.exit(exit_code)
+ for msg in strings_to_concatenate(os.path.join(subdir, file_name)):
+ if msg:
+ failed = True
+ print(msg)
+ sys.exit(failed)