Used isort to sort imports. Also added to the git pre-commit hook.

[python_utils.git] / file_utils.py

diff --git a/file_utils.py b/file_utils.py
index 5d9a0be3b272bbc51eebf7894baac7a25fe11179..905e23b16df017d3cab6c6ef27ce61ee8596c112 100644 (file)
--- a/file_utils.py
+++ b/file_utils.py
@@ -4,22 +4,52 @@
 
 import datetime
 import errno
+import glob
 import hashlib
+import io
 import logging
 import os
-import io
 import pathlib
+import re
 import time
-from typing import Optional
-import glob
-from os.path import isfile, join, exists
-from typing import List
+from os.path import exists, isfile, join
+from typing import List, Optional, TextIO
 from uuid import uuid4
 
-
 logger = logging.getLogger(__name__)
 
 
+def remove_newlines(x):
+    return x.replace('\n', '')
+
+
+def strip_whitespace(x):
+    return x.strip()
+
+
+def remove_hash_comments(x):
+    return re.sub(r'#.*$', '', x)
+
+
+def slurp_file(
+    filename: str,
+    *,
+    skip_blank_lines=False,
+    line_transformers=[],
+):
+    ret = []
+    if not file_is_readable(filename):
+        raise Exception(f'{filename} can\'t be read.')
+    with open(filename) as rf:
+        for line in rf:
+            for transformation in line_transformers:
+                line = transformation(line)
+            if skip_blank_lines and line == '':
+                continue
+            ret.append(line)
+    return ret
+
+
 def remove(path: str) -> None:
     """Deletes a file.  Raises if path refers to a directory or a file
     that doesn't exist.
@@ -184,6 +214,18 @@ def does_file_exist(filename: str) -> bool:
     return os.path.exists(filename) and os.path.isfile(filename)
 
 
+def file_is_readable(filename: str) -> bool:
+    return does_file_exist(filename) and os.access(filename, os.R_OK)
+
+
+def file_is_writable(filename: str) -> bool:
+    return does_file_exist(filename) and os.access(filename, os.W_OK)
+
+
+def file_is_executable(filename: str) -> bool:
+    return does_file_exist(filename) and os.access(filename, os.X_OK)
+
+
 def does_directory_exist(dirname: str) -> bool:
     """Returns True if a file exists and is a directory.
 
@@ -288,11 +330,13 @@ def get_file_md5(filename: str) -> str:
 
 def set_file_raw_atime(filename: str, atime: float):
     mtime = get_file_raw_mtime(filename)
+    assert mtime
     os.utime(filename, (atime, mtime))
 
 
 def set_file_raw_mtime(filename: str, mtime: float):
     atime = get_file_raw_atime(filename)
+    assert atime
     os.utime(filename, (atime, mtime))
 
 
@@ -390,8 +434,8 @@ def describe_file_mtime(filename: str, *, brief=False) -> Optional[str]:
     return describe_file_timestamp(filename, lambda x: x.st_mtime, brief=brief)
 
 
-def touch_file(filename: str, *, mode: Optional[int] = 0o666) -> bool:
-    return pathlib.Path(filename, mode=mode).touch()
+def touch_file(filename: str, *, mode: Optional[int] = 0o666):
+    pathlib.Path(filename, mode=mode).touch()
 
 
 def expand_globs(in_filename: str):
@@ -426,14 +470,14 @@ class FileWriter(object):
         self.filename = filename
         uuid = uuid4()
         self.tempfile = f'{filename}-{uuid}.tmp'
-        self.handle = None
+        self.handle: Optional[TextIO] = None
 
-    def __enter__(self) -> io.TextIOWrapper:
+    def __enter__(self) -> TextIO:
         assert not does_path_exist(self.tempfile)
         self.handle = open(self.tempfile, mode="w")
         return self.handle
 
-    def __exit__(self, exc_type, exc_val, exc_tb) -> bool:
+    def __exit__(self, exc_type, exc_val, exc_tb) -> Optional[bool]:
         if self.handle is not None:
             self.handle.close()
             cmd = f'/bin/mv -f {self.tempfile} {self.filename}'