Modernize codebase, unify variable names... (#172)

* Refactor

* Get rid of `striplastline` tests fixture

* Rewrite `tempfile.NamedTemporaryFile` in tests using context manager

* Remove uneeded casts to str in tests

* Refactor `capsys.readouterr` calls in tests

* Drop uneeded cast to str

* Refactor print comments

* Drop uneeded casts to `str` comparing `polib.POFile` with strings

* Improve test

* Add more tests and refactor function

* Allow to pass negative values to `wrapwidth` arguments.

* Rewrite incorrect statement from docs

* Remove some 'pragma: no cover' comments

* Drop `filehash` and `save_pofile_checking_file_changed` functions

* Fix empty indented code block line error in `po2md`

* Bump version

* Fix image links regression from v0.3.68

.bumpversion.cfg

@@ -1,5 +1,5 @@
 [bumpversion]
-current_version = 0.3.70
+current_version = 0.3.71
 
 [bumpversion:file:mdpo/__init__.py]
 

.pre-commit-config.yaml

@@ -10,6 +10,7 @@ repos:
     hooks:
       - id: trailing-whitespace
         name: trailing-whitespace
+        exclude: test/test_po2md/translate-examples/code-blocks.*
       - id: end-of-file-fixer
         name: end-of-file-fixer
       - id: double-quote-string-fixer

docs/pre-commit-hooks.rst

@@ -15,7 +15,7 @@ so you don't need to specify them.
 .. code-block:: yaml
 
    - repo: https://github.com/mondeja/mdpo
-     rev: v0.3.70
+     rev: v0.3.71
      hooks:
        - id: md2po
          args:
@@ -43,7 +43,7 @@ po2md
 .. code-block:: yaml
 
    - repo: https://github.com/mondeja/mdpo
-     rev: v0.3.70
+     rev: v0.3.71
      hooks:
        - id: po2md
          args:
@@ -56,7 +56,7 @@ po2md
 .. code-block:: yaml
 
    - repo: https://github.com/mondeja/mdpo
-     rev: v0.3.70
+     rev: v0.3.71
      hooks:
        - id: po2md
          files: README\.md
@@ -75,7 +75,7 @@ md2po2md
 .. code-block:: yaml
 
    - repo: https://github.com/mondeja/mdpo
-     rev: v0.3.70
+     rev: v0.3.71
      hooks:
        - id: md2po2md
          args:
@@ -88,7 +88,7 @@ md2po2md
 .. code-block:: yaml
 
    - repo: https://github.com/mondeja/mdpo
-     rev: v0.3.70
+     rev: v0.3.71
      hooks:
        - id: md2po2md
          files: README\.md

mdpo/__init__.py

@@ -6,7 +6,7 @@
 from mdpo.po2md import pofile_to_markdown
 
 
-__version__ = '0.3.70'
+__version__ = '0.3.71'
 __title__ = 'mdpo'
 __description__ = ('Markdown files translation using PO files.')
 __all__ = (

mdpo/event.py

@@ -36,7 +36,7 @@ def _block_msg(block, details):
     return block.name
 
 
-def debug_events(program):  # pragma: no cover
+def debug_events(program):
     """Debugging events for interfaces. Displays in STDOUT all event targets.
 
     Args:

mdpo/html.py

@@ -18,9 +18,9 @@ def html_attrs_tuple_to_string(attrs):
     """
     response = ''
     for i, (name, value) in enumerate(attrs):
-        response += '%s' % name
+        response += name
         if value is not None:
-            response += '="%s"' % value
+            response += f'="{value}"'
         if i < len(attrs) - 1:
             response += ' '
     return response

mdpo/io.py

@@ -1,9 +1,13 @@
 """mdpo I/O utilities."""
 
+import filecmp
 import glob
-import hashlib
 import os
 import re
+import tempfile
+
+
+MDPO_FILECMP_TEMPFILE = os.path.join(tempfile.gettempdir(), '__mpdo-filecmp')
 
 
 def filter_paths(filepaths, ignore_paths=[]):
@@ -85,18 +89,6 @@ def to_glob_or_content(value):
     return (True, parsed)
 
 
-def filehash(filepath):
-    """Compute the hash of a file.
-
-    Args:
-        filepath (str): Path to the file.
-    """
-    hasher = hashlib.md5()
-    with open(filepath, 'rb') as f:
-        hasher.update(f.read())
-    return hasher.hexdigest()
-
-
 def save_file_checking_file_changed(filepath, content, encoding='utf-8'):
     """Save a file checking if the content has changed.
 
@@ -112,9 +104,9 @@ def save_file_checking_file_changed(filepath, content, encoding='utf-8'):
             f.write(content)
         return True
 
-    pre_hash = filehash(filepath)
+    with open(MDPO_FILECMP_TEMPFILE, 'w') as temp_f, open(filepath) as f:
+        temp_f.write(f.read())
     with open(filepath, 'w', encoding=encoding) as f:
         f.write(content)
-    post_hash = filehash(filepath)
 
-    return pre_hash != post_hash
+    return filecmp.cmp(MDPO_FILECMP_TEMPFILE, filepath)

mdpo/md.py

@@ -46,9 +46,8 @@ def escape_links_titles(text, link_start_string='[', link_end_string=']'):
 
     for match in re.findall(regex, text):
         original_string = match[0] + match[1]
-        target_string = match[0] + '"%s"' % (
-            match[1][1:-1].replace('"', '\\"')
-        )
+        escaped_title = match[1][1:-1].replace('"', '\\"')
+        target_string = f'{match[0]}"{escaped_title}"'
         text = text.replace(original_string, target_string)
     return text
 
@@ -418,10 +417,10 @@ def solve_link_reference_targets(translations):
         # store in solutions
         solutions[new_msgid] = new_msgstr
 
-        # print("----> new_msgid", new_msgid)
-        # print("----> new_msgstr", new_msgstr)
+        # print('----> new_msgid', new_msgid)
+        # print('----> new_msgstr', new_msgstr)
 
-    # print("----> link_references_text_targets", link_references_text_targets)
-    # print("----> msgid_msgstrs_with_links", msgid_msgstrs_with_links)
-    # print("----> solutions", solutions)
+    # print('----> link_references_text_targets', link_references_text_targets)
+    # print('----> msgid_msgstrs_with_links', msgid_msgstrs_with_links)
+    # print('----> solutions', solutions)
     return solutions

mdpo/md2po/__init__.py

@@ -10,7 +10,11 @@
     parse_mdpo_html_command,
 )
 from mdpo.event import debug_events, raise_skip_event
-from mdpo.io import filter_paths, to_glob_or_content
+from mdpo.io import (
+    filter_paths,
+    save_file_checking_file_changed,
+    to_glob_or_content,
+)
 from mdpo.md import parse_link_references
 from mdpo.md4c import (
     DEFAULT_MD4C_GENERIC_PARSER_EXTENSIONS,
@@ -21,7 +25,6 @@
     mark_not_found_entries_as_obsoletes,
     po_escaped_string,
     remove_not_found_entries,
-    save_pofile_checking_file_changed,
 )
 from mdpo.text import min_not_max_chars_in_a_row, parse_wrapwidth_argument
 
@@ -172,7 +175,7 @@ def __init__(self, glob_or_content, **kwargs):
         self._include_next_codeblock = False
         self._disable_next_codeblock = False
 
-        self._saved_files_changed = (  # pragma: no cover
+        self._saved_files_changed = (
             False if kwargs.get('_check_saved_files_changed') else None
         )
 
@@ -781,16 +784,30 @@ def leave_span(self, span, details):
                     )
                 self._codespan_backticks = None
             elif span is md4c.SpanType.IMG:
-                self._current_msgid += '![{}]({}'.format(
-                    self._current_imgspan['text'],
-                    self._current_imgspan['src'],
-                )
-                if self._current_imgspan['title']:
-                    self._current_msgid += ' "%s"' % (
-                        self._current_imgspan['title']
+                # TODO: refactor with getattr? Currently getting next error
+                # getattr(self, target_varname) += '![{}]({}'.format(
+                # SyntaxError: cannot assign to function call
+
+                if not self._inside_aspan:
+                    self._current_msgid += '![{}]({}'.format(
+                        self._current_imgspan['text'],
+                        self._current_imgspan['src'],
+                    )
+                    title = self._current_imgspan['title']
+                    if title:
+                        self._current_msgid += f' "{title}"'
+                    self._current_msgid += ')'
+                    self._current_imgspan = {}
+                else:
+                    self._current_aspan_text += '![{}]({}'.format(
+                        self._current_imgspan['text'],
+                        self._current_imgspan['src'],
                     )
-                self._current_msgid += ')'
-                self._current_imgspan = {}
+                    title = self._current_imgspan['title']
+                    if title:
+                        self._current_aspan_text += f' "{title}"'
+                    self._current_aspan_text += ')'
+                    self._current_imgspan = {}
             elif span is md4c.SpanType.U:
                 self._inside_uspan = False
 
@@ -968,9 +985,10 @@ def _parse(content):
 
         if save and po_filepath:
             if self._saved_files_changed is False:  # pragma: no cover
-                self._saved_files_changed = save_pofile_checking_file_changed(
-                    self.pofile,
+                self._saved_files_changed = save_file_checking_file_changed(
                     po_filepath,
+                    str(self.pofile),
+                    encoding=self.pofile.encoding,
                 )
             else:
                 self.pofile.save(fpath=po_filepath)
@@ -1034,8 +1052,8 @@ def markdown_to_pofile(
                 for you. It depends on the use you are going to give to
                 this library activate this mode (``plaintext=False``) or not.
         wrapwidth (int): Wrap width for po file indicated at ``po_filepath``
-            parameter. Only useful when the ``-w`` option was passed
-            to xgettext.
+            parameter. If negative, 0, 'inf' or 'math.inf' the content won't
+            be wrapped.
         mark_not_found_as_obsolete (bool): The strings extracted from markdown
             that will not be found inside the provided pofile will be marked
             as obsolete.

mdpo/md2po/__main__.py

@@ -76,9 +76,8 @@ def build_parser():
         '-w', '--wrapwidth', dest='wrapwidth', metavar='N/inf', type=str,
         default='78',
         help='Wrap width for po file indicated at \'-po/--po-filepath\''
-             ' parameter. Only useful when the \'-w\' option was passed to'
-             ' xgettext. You can use the values \'0\' and \'inf\' for infinite'
-             ' width.',
+             ' parameter. If negative, \'0\' or \'inf\', the PO file content'
+             ' will not be wrapped.',
     )
     parser.add_argument(
         '-m', '--merge-po-files', '--merge-pofiles',
@@ -204,12 +203,10 @@ def run(args=[]):
         pofile = md2po.extract(**extract_kwargs)
 
         if not opts.quiet:
-            sys.stdout.write(pofile.__unicode__() + '\n')
+            sys.stdout.write(f'{pofile}\n')
 
         # pre-commit mode
-        if (  # pragma: no cover
-            opts.check_saved_files_changed and md2po._saved_files_changed
-        ):
+        if opts.check_saved_files_changed and md2po._saved_files_changed:
             return (pofile, 1)
 
     return (pofile, 0)

mdpo/md2po2md/__init__.py

@@ -108,9 +108,9 @@ def markdown_to_pofile_to_markdown(
 
             os.makedirs(os.path.abspath(po_dirpath), exist_ok=True)
             if os.path.isdir(po_filepath):
-                po_filepath = (
-                    po_filepath.rstrip(os.sep) + os.sep +
-                    os.path.basename(filepath) + '.po'
+                po_filepath = os.path.join(
+                    po_filepath.rstrip(os.sep),
+                    f'{os.path.basename(filepath)}.po',
                 )
             if not po_filepath.endswith('.po'):
                 po_filepath += '.po'
@@ -139,9 +139,7 @@ def markdown_to_pofile_to_markdown(
                 po_encoding=po_encoding,
                 md_encoding=md_encoding,
             )
-            if (  # pragma: no cover
-                _check_saved_files_changed and _saved_files_changed is False
-            ):
+            if _check_saved_files_changed and _saved_files_changed is False:
                 _saved_files_changed = md2po._saved_files_changed
 
             # po2md
@@ -158,9 +156,7 @@ def markdown_to_pofile_to_markdown(
                 save=md_filepath,
                 md_encoding=md_encoding,
             )
-            if (  # pragma: no cover
-                _check_saved_files_changed and _saved_files_changed is False
-            ):
+            if _check_saved_files_changed and _saved_files_changed is False:
                 _saved_files_changed = po2md._saved_files_changed
 
     return _saved_files_changed

mdpo/md2po2md/__main__.py

@@ -112,9 +112,7 @@ def run(args=[]):
             _check_saved_files_changed=opts.check_saved_files_changed,
             **kwargs,
         )
-        if (  # pragma: no cover
-            opts.check_saved_files_changed and _saved_files_changed
-        ):
+        if opts.check_saved_files_changed and _saved_files_changed:
             exitcode = 1
     return exitcode