first commit (Ligurian Support)

epitran/data/map/lij-Latn.csv

@@ -0,0 +1,58 @@
+Orth,Phon
+a,a
+á,a
+à,a
+â,aː
+ä,aː
+æ,ɛː
+æn,ɛn
+æñ,ɛŋ
+b,b
+c,k
+ç,s
+d,d
+e,e
+é,e
+ê,eː
+ë,eː
+è,ɛ
+éu,ew
+eu,ø
+f,f
+g,ɡ
+gl,d͡ʒ
+gn,ɲ
+h,
+i,i
+í,i
+ì,i
+î,iː
+ï,iː
+j,j
+l,l
+m,m
+n,n
+ñ,ŋ
+ò,ɔ
+ö,ɔː
+o,u
+ó,u
+ô,uː
+ou,ɔw
+p,p
+q,k
+qu,kw
+r,r
+s,s
+sch,sk
+t,t
+u,y
+ú,y
+ù,y
+û,yː
+ü,yː
+v,v
+z,z
+<dZ>,d͡ʒ
+<StS>,ʃt͡ʃ
+<tS>,t͡ʃ

epitran/data/post/lij-Latn.txt

@@ -0,0 +1,54 @@
+::voicedcons:: = b|d|d͡ʒ|ɡ|l|m|n|ŋ|r|v
+::vowels:: = a|ɛ|e|i|ɔ|ø|u|y
+::consonant:: = b|ʃ|d|d͡ʒ|f|ɡ|j|k|l|m|n|ŋ|ɲ|p|r|s|t|t͡ʃ|v|w|z|ʒ
+
+% Treatment of <ao> endings
+au -> ɔw / _ #
+
+% Treatment of <sc>
+st͡ʃ -> ʃ / _
+
+% Treatment of <n>
+n -> ŋ / _ #
+n -> ŋ / _ (::consonant::)
+[nm] -> ŋ / _ [bp]
+
+% Treatment of <s>
+s -> z / _ (::voicedcons::)
+s -> z / (a|ɛː?|eː?|i|ɔ|øː?|uː?|yː?) _ (::vowels::)
+s -> z / [ai]ː _ (?!e#)
+
+% Treatment of <gu>
+ɡy -> ɡw / (?<!fi) _ (::vowels::)
+seɡw -> seɡy / _ (::vowels::)
+yŋna -> ina / _ #
+
+% Plural endings
+z?iuŋ -> sjuŋ / _
+zi?uiŋ -> swiŋ / _
+i?uiŋ -> wiŋ / _
+
+% Treatment of double consonants
+b+ -> b / _
+ʃ+ -> ʃ / _
+d+ -> d / _
+ɡ*d͡ʒ+ -> d͡ʒ / _
+f+ -> f / _
+ɡ+ -> ɡ / _
+j+ -> j / _
+k+ -> k / _
+l+ -> l / _
+m+ -> m / _
+n+ -> n / _
+ŋ*n+ -> n / _
+n*ŋ+ -> ŋ / _
+ɲ+ -> ɲ / _
+p+ -> p / _
+r+ -> r / _
+s+ -> s / _
+t+ -> t / _
+k*t͡ʃ+ -> t͡ʃ / _
+v+ -> v / _
+w+ -> w / _
+z+ -> z / _
+ʒ+ -> ʒ / _

epitran/data/pre/lij-Latn.txt

@@ -0,0 +1,11 @@
+scc -> <StS> / _
+cci -> <tS> / [^s] _ [aeou]
+cc -> <tS> / [^s] _ [ie]
+ci -> <tS> / _ [aeou]
+c -> <tS> / _ [ie]
+ggi -> <dZ> / _ [aeou]
+gg -> <dZ> / _ [ie]
+gi -> <dZ> / _ [aeou]
+gli -> <dZ> / _ [aeou]
+gl -> <dZ> / _ [ie]
+g -> <dZ> / _ [ie]

tests.ipynb

@@ -0,0 +1,102 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "'faːse zeneize fiɡyraːse fiɡya ɛːɡwa leŋɡwa ɡwɛːra iŋfiltraziuŋ iŋfiltraziuiŋ paʃuŋ paʃuiŋ ʃt͡ʃetu ʃt͡ʃøpu kaŋ kaŋpu teŋpu sykɔw datɔw datai'"
+      ]
+     },
+     "execution_count": 2,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "import epitran\n",
+    "epi = epitran.Epitran('lij-Latn')\n",
+    "\" \".join(epi.transliterate(w) for w in \"fâse zeneise figurâse figua ægua lengua guæra infiltraçion infiltraçioin pascion pascioin sccetto scceuppo can campo tempo succao dattao dattai\".split())"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "'yŋna paŋna'"
+      ]
+     },
+     "execution_count": 3,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "sentence = \"pròpio unna panna\"\n",
+    "\" \".join(epi.transliterate(w) for w in sentence.split())"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "'meʃt͡ʃiaː ʃabeku zɡød͡ʒa t͡ʃiɛːu aŋɡɛːzu'"
+      ]
+     },
+     "execution_count": 4,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "sentence = \"mescciâ sciabecco sgheuggia ciæo angæzo\"\n",
+    "\" \".join(epi.transliterate(w) for w in sentence.split())"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3.7.1 ('base')",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.7.1"
+  },
+  "orig_nbformat": 4,
+  "vscode": {
+   "interpreter": {
+    "hash": "8717662bf88969997ef180c9a6313e076396e402a442e83df9655bbeeedc9756"
+   }
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}