Utils: rewrite latex-preprocessing in python

utils/checker.sh

@@ -1,46 +0,0 @@
-#!/bin/sh
-# Utility that corrects the songs input to respect some LaTeX rules
-
-
-for i in songs/*/*.sg; 
-do 
-
-# oe inclusion
-sed -i "s/coeur/cœur/g" $i;
-sed -i "s/boeuf/bœuf/g" $i;
-sed -i "s/oeuvre/œuvre/g" $i;
-sed -i "s/oeuf/œuf/g" $i;
-sed -i "s/soeur/sœur/g" $i;
-sed -i "s/noeud/nœud/g" $i;
-sed -i "s/oeil/œil/g" $i;
-sed -i "s/\\\oe{}/œ/g" $i;
-#punctuation
-sed -i "s/’/'/g" $i;
-sed -i "s/Ca /Ça /g" $i;
-
-# dots rules
-sed -i "s/\\.\\.\\./\\\dots/g" $i;
-
-#chords
-sed -i "s/\\[Re\\]/\\[Ré\\]/g" $i;
-sed -i "s/b\\]/\\&\\]/g" $i;
-#Do
-sed -i "s/032010}/X32010}/g" $i;
-#La
-sed -i "s/002220}/X02220}/g" $i;
-sed -i "s/002020}/X02020}/g" $i;
-#Lam
-sed -i "s/002210}/X02210}/g" $i;
-#Ré
-sed -i "s/000232}/XX0232}/g" $i;
-sed -i "s/X00232}/XX0232}/g" $i;
-#Ré7
-sed -i "s/000212}/XX0212}/g" $i;
-#Rém
-sed -i "s/000231}/XX0231}/g" $i;
-sed -i "s/X00231}/XX0231}/g" $i;
-#Si7
-sed -i "s/021202}/X21202}/g" $i;
-
-
-done;

utils/latex-preprocessing.py

@@ -0,0 +1,63 @@
+#!/usr/bin/python
+# -*- coding: utf-8 -*-
+
+import re
+import glob
+import codecs
+ 
+def replace_words(text, word_dic):
+    """
+    take a text and <strong class="highlight">replace</strong> words
+    that match a key in a dictionary with the associated value, 
+    return the changed text
+    """
+    rc = re.compile('|'.join(map(re.escape, word_dic)))
+    def translate(match):
+        return word_dic[match.group(0)]
+    return rc.sub(translate, text)
+ 
+# the dictionary has target_word:replacement_word pairs
+word_dic = {
+#oe inclusion
+"coeur": "cœur",
+"boeuf": "bœuf",
+"oeuvre": "œuvre",
+"soeur": "sœur",
+"noeud": "nœud",
+"oeil": "œil",
+"oe{}": "œ",
+#punctuation
+"’": "'",
+"Ca ": "Ça ",
+"...": "\\dots ",
+#Chords
+"\\[Re]": "\\[Ré]",
+"b]": "&]",
+#Do
+"032010": "X32010",
+#La
+"002220": "X02220",
+"002020": "X02020",
+"002210": "X02210",
+#Ré
+"000232": "XX0232",
+"X00232": "XX0232",
+"000212": "XX0212",
+"000231": "XX0231",
+"X00231": "XX0231",
+#Si
+"021202": "X21202",
+}
+ 
+# Process song files
+songfiles = glob.glob('songs/*/*.sg')
+for file in songfiles:
+    songfile = codecs.open(file, "r", "utf-8")
+    data = songfile.read().encode("utf-8")
+    data = replace_words(data, word_dic)
+    songfile.close()
+    songfile = open(file, "w")
+    songfile.write(data)
+    songfile.close()
+ 
+