Moves map to config

Adds suggested characters in terminal
Adds ignored list
Some cosmetic changes

README.md

@@ -5,12 +5,22 @@ Simple script to:
 - convert line endings to UNIX standard
 - fix broken characters
 
+## Configuration
+
+You can place your own map in `characters` section of `config.yaml` file.
+Default configuration includes polish characters.
+When run in terminal, script lists all characters found in the subtitles file that are not in ASCII or ignored list.
+
 ## Usage
 
 ```bash
-./subtitles-fixer.py /path/to/subtitles.ext
+$ python3 /path/to/subtitles-fixer.py /path/to/subtitles.ext
 ```
 
 ## Bonus
 
 Nemo action for context menu. Place `fix-subtitles.nemo_action` in `~/.local/share/nemo/actions/` and restart nemo.
+
+## To do
+
+- Translation?

config.yaml

@@ -0,0 +1,33 @@
+characters:
+  ¿: ż
+  ³: ł
+  æ: ć
+  ¹: ą
+  ê: ę
+  ñ: ń
+  £: Ł
+  ¯: Ż
+  œ: ś
+  Ÿ: ź
+  Œ: Ś
+
+ignored:
+    - ą
+    - ć
+    - ę
+    - ł
+    - ń
+    - ó
+    - ś
+    - ź
+    - ż
+    - Ą
+    - Ć
+    - Ę
+    - Ł
+    - Ń
+    - Ó
+    - Ś
+    - Ź
+    - Ż
+

fix-subtitles.nemo_action

@@ -3,7 +3,7 @@
 Active=true
 Name=Napraw napisy
 Comment=Odpala skrypt do naprawiania napisów
-Exec=/path/to/subtitles-fixer.py %F
+Exec=python3 /path/to/subtitles-fixer.py %F
 Icon-Name=edit
 Selection=s
 Extensions=txt;srt;sub;

subtitles-fixer.py

@@ -4,6 +4,7 @@ import sys
 import os.path
 import subprocess
 import chardet
+import yaml
 
 def notify(header, body, exitCode):
     subprocess.run(['/usr/bin/notify-send', header, body])
@@ -24,58 +25,10 @@ filepath = sys.argv[1]
 if not os.path.isfile(filepath):
     notify('Błąd', 'Aby naprawić napisy, musisz wskazać plik', 1)
 
-map = {
-    '': '',
-    'Ä…': 'ą',
-    '¹': 'ą',
-    'š': 'ą',
-    'ć': 'ć',
-    'ĂŚ': 'ć',
-    'æ': 'ć',
-    'ć': 'ć',
-    'Ä™': 'ę',
-    'ĂŞ': 'ę',
-    'ê': 'ę',
-    'ę': 'ę',
-    'Ĺ‚': 'ł',
-    'Âł': 'ł',
-    '³': 'ł',
-    'ł': 'ł',
-    'Ĺ„': 'ń',
-    'Ăą': 'ń',
-    'ñ': 'ń',
-    'ń': 'ń',
-    'Ăł': 'ó',
-    'Ĺ›': 'ś',
-    'œ': 'ś',
-    'œ': 'ś',
-    'œ': 'ś',
-    'Ĺş': 'ź',
-    'Ÿ': 'ź',
-    'ĹĽ': 'ż',
-    'Âż': 'ż',
-    '¿': 'ż',
-    'Ĺź': 'ź',
-    'Ÿ': 'ź',
-    'Ÿ': 'ź',
-    'Ä„': 'Ą',
-    'Ć': 'Ć',
-    'Æ': 'Ć',
-    'Ę': 'Ę',
-    'Ê': 'Ę',
-    'Ł': 'Ł',
-    '£': 'Ł',
-    #'': 'Ń',
-    'Ñ': 'Ń',
-    'Ń': 'Ń',
-    #'': 'Ó',
-    'Ĺš': 'Ś',
-    'Œ': 'Ś',
-    'Œ': 'Ś',
-    '': 'Ź',
-    'Ĺ»': 'Ż',
-    '¯': 'Ż',
-}
+with open('config.yaml', 'r') as configFile:
+    config = yaml.safe_load(configFile)
+    configFile.close()
+    map = config['characters']
 
 try:
     encoding = detectEncodingOfAFile(filepath)
@@ -89,6 +42,16 @@ try:
     for badChar, goodChar in map.items():
         text = text.replace(badChar, goodChar)
 
+    badChars = []
+    for c in text:
+        if ord(c) > 127 and not c in badChars:
+            badChars.append(c)
+    for char in config['ignored']:
+        if char in badChars:
+            badChars.remove(char)
+    if len(badChars):
+        print('Znalezione znaki, które można dodać do mapy: ' + badChars.__str__())
+
     with open(filepath, 'w', newline='\n') as f:
         f.write(text)
         f.close()