diff --git a/legacy/data/DivehiThaanaRomanization.cfg b/legacy/data/DivehiThaanaRomanization.cfg
new file mode 100644
index 0000000..cbaa8e1
--- /dev/null
+++ b/legacy/data/DivehiThaanaRomanization.cfg
@@ -0,0 +1,573 @@
+# Original table by Randall K. Barry, 28 February 2021
+# Updated by Randall K. Barry, 11 May 2023 to add script code
+
+[General]
+Name=Divehi (Thaana script)
+ScriptCode=Thaa
+
+# Script used by the Divehi language of the Maldives
+
+# UNICODE THAANA CHARACTER RANGE: U+0780-U+07BF
+# ISO 15924 4-alpha script code: Thaa
+
+Truncation=%
+
+[RomanToScript]
+FieldsIncluded=100 110 111 130 240 245 246 250 260 264 440 490 600 610 611 630 651 700 710 711 730 740 800 830
+SubfieldsAlwaysExcluded=uvxy0123456789
+OtherSubfieldsExcludedByTag=100/e 110/e 111/j 246/i 260/c 264/c 650/a 700/e 700/i 710/e 710/i 711/i 711/j 730/i
+
+# RDA boilerplate phrases not transliterated:
+Place of publication not identified=Place of publication not identified
+publisher not identified=publisher not identified
+date of publication not identified=date of publication not identified
+At head of title=At head of title
+Colophon=Colophon
+and others=and others
+and one other=and one other
+and two others=and two others
+and three others=and three others
+and four others=and four others
+and five others=and five others
+and six others=and six others
+and seven others=and seven others
+and eight others=and eight others
+and nine others=and nine others
+and ten others=and ten others
+and eleven others=and eleven others
+and twelve others=and twelve others
+and thirteen others=and thirteen others
+and fourteen others=and fourteen others
+and fifteen others=and fifteen others
+and sixteen others=and sixteen others
+and seventeen others=and seventeen others
+and eighteen others=and eighteen others
+and nineteen others=and nineteen others
+et al.=et al.
+i.e.=i.e.
+S.l.=S.l.
+s.n.=s.n.
+XLIX=XLIX
+XLIV=XLIV
+XLIII=XLIII
+XLII=XLII
+XLI=XLI
+XLVIII-XLVIII
+XLVII=XLVII
+XLVI=XLVI
+XLV=XLV
+XL=XL
+LXXXIX=LXXXIX
+LXXXIV=LXXXIV
+LXXXIII=LXXXIII
+LXXXII=LXXXII
+LXXXI=LXXXI
+LXXXVIII=LXXXVIII
+LXXXVII=LXXXVII
+LXXXVI=LXXXVI
+LXXXV=LXXXV
+LXXX=LXXX
+LXXIX=LXXIX
+LXXIV=LXXIV
+LXXIII=LXXIII
+LXXII=LXXII
+LXXI=LXXI
+LXXVIII=LXXVIII
+LXXVII=LXXVII
+LXXVI=LXXVI
+LXXV=LXXV
+LXX=LXX
+LXIX=XIX
+LXIV=LXIV
+LXIII=LXIII
+LXII=LXII
+LXI=LXI
+LXVIII=LXVIII
+LXVII=LXVII
+LXVI=LXVI
+LXV=LXV
+LX=LX
+LIX=LIX
+LIV=LIV
+LIII=LIII
+LII=LII
+LI=LI
+LVIII=LVIII
+LVII=LVII
+LVI=LVI
+LV=LV
+XXXIX=XXXIX
+XXXIV=XXXIV
+XXXIII=XXXIII
+XXXII=XXXII
+XXXI=XXXI
+XXXVIII=XXXVIII
+XXXVII=XXXVII
+XXXVI=XXXVI
+XXXV=XXXV
+XXX=XXX
+XXIX=XXIX
+XXIV=XXIV
+XXIII=XXIII
+XXII=XXII
+XXI=XXI
+XXVIII=XXVIII
+XXVII=XXVII
+XXVI=XXVI
+XXV=XXV
+XX=XX
+XIX=XIX
+XIV=XIV
+XIII=XIII
+XII=XII
+XI=XI
+XVIII=XVIII
+XVII=XVII
+XVI=XVI
+XV=XV
+IV=IV
+VIII=VIII
+VII=VII
+VI=VI
+IX=IX
+III=III
+II=II
+
+# COMMON COMBINING CHARACTERS (always follow a base letter)
+
+# combining acute U+0301
+# combining tilde U+0303
+# combining macron U+0304
+# combining dot above U+0307
+# combining dot below U+0323
+# combining diaeresis below U+0324
+# combining low line U+0332
+# combining breve below U+032E
+# ayn (spacing) U+02BB
+# apostrophe (spacing) U+0027
+
+# REGULAR LATIN ALPHABETIC CHARACTERS TO BE CONVERTED
+
+# THAANA LETTER ALIF APPEARING MEDIALLY WITH ANY VOWEL IS ROMANIZED WITH APOSTROPHE
+# ORIGINAL VOWEL ASSOCIATED WITH THE ALIF CANNOT BE REGENERATED DURING CONVERSION
+U+0027=U+0787
+
+# THAANA LETTER SHAVIYANI WITH SUKUN(BREVE 032E)
+HU+032E=U+0781U+07B0
+hh=U+0787U+07B0U+0780
+hU+032E=U+0781U+07B0
+
+# THAANA LETTER ALIFU FINAL WITH SUKUN (LOW LINE 0332)
+%HU+0332=U+0787U+07B0
+%hU+0332=U+0787U+07B0
+H=U+0780
+h=U+0780
+SU+0301=U+0781
+ss=U+0787U+07B0U+0781
+sU+0301=U+0781
+
+# THAANA LETTER MEDIAL NOONU WITHOUT SUKUN (DOT ABOVE 0307)
+MU+0307=U+0782
+mU+0307=U+0782
+mm=U+0787U+07B0U+0789
+N=U+0782
+nn=U+0787U+07B0U+0782
+n=U+0782
+R=U+0783
+rr=U+0787U+07B0U+0783
+r=U+0783
+B=U+0784
+bb=U+0787U+07B0U+0784
+b=U+0784
+
+# THAANA LETTER "L/l" WITH DOT BELOW (0323) 
+LU+0323=U+0785
+lU+0323=U+0785
+K=U+0786
+kk=U+0787U+07B0U+0786
+k=U+0786
+
+# THAANA LETTER ALIF--OFTEN OMITTED IN ROMANIZATION (SEE VOWEL CONVERSION BELOW)
+U+0027=U+0787
+
+V=U+0788
+vv=U+0787U+07B0U+0788
+v=U+0788
+M=U+0789
+m=U+0789
+F=U+078A
+ff=U+0787U+07B0U+078A
+f=U+078A
+# THAANA LETTER "D/d" WITH DOT BELOW (0323) 
+DU+0323=U+0791
+dU+0323=U+0791
+D=U+078B
+dd=U+0787U+07B0U+078B
+d=U+078B
+# THAANA LETTER "T/t" WITH DOT BELOW (0323) 
+TU+0323=U+0793
+tU+0323=U+0793
+TU+0324T=U+078CU+07B0U+078C
+TU+0324t=U+078CU+07B0U+078C
+tU+0324t=U+078CU+07B0U+078C
+tU+0324T=U+078CU+07B0U+078C
+T=U+078C
+tt=U+0787U+07B0U+078C
+t=U+078C
+L=U+078D
+ll=U+0787U+07B0U+078D
+l=U+078D
+G=U+078E
+gg=U+0787U+07B0U+078E
+g=U+078E
+# THAANA LETTER "N/n' WITH TILDE (0303)
+NU+0303=U+078F
+nU+0303=U+078F
+S=U+0790
+ss=U+0787U+07B0U+0790
+s=U+0790
+Z=U+0792
+zz=U+0787U+07B0U+0792
+z=U+0792
+Y=U+0794
+yy=U+0787U+07B0U+0794
+y=U+0794
+P=U+0795
+p=U+0795
+pp=U+0787U+07B0U+0795
+J=U+0796
+jj=U+0787U+07B0U+0796
+j=U+0796
+C=U+0797
+cc=U+0787U+07B0U+0797
+c=U+0797
+
+# THAANA EXTENSION FOR ARABIC LOAN WORDS AND NAMES
+
+# THAANA EXTENSION FOR ARABIC LETTER TTAA
+TH=U+0798
+Th=U+0798
+thth=U+0787U+07B0U+0798
+th=U+0798
+
+# THAANA EXTENSION FOR ARABIC LETTER HHAA
+HU+0323=U+0799
+hU+0323=U+0799
+
+# THAANA EXTENSION FOR ARABIC LETTER KHAA
+KH=U+079A
+Kh=U+079A
+khkh=U+0787U+07B0U+079A
+kh=U+079A
+
+# THAANA EXTENSION FOR ARABIC LETTER THAALU
+DH=U+079B
+Dh=U+079B
+dhdh=U+0787U+07B0U+079B
+dh=U+079B
+
+# THAANA EXTENSION FOR ARABIC LETTER ZAA (NEWER LETTER)
+ZU+0332=U+079C
+zU+0332=U+079C
+
+# THAANA EXTENSION FOR ARABIC LETTER SHEENU
+SH=U+079D
+Sh=U+079D
+shsh=U+0787U+07B0U+079D
+sh=U+079D
+
+# THAANA EXTENSION FOR ARABIC LETTER SAADHU
+SU+0323=U+079E
+sU+0323=U+079E
+
+# THAANA EXTENSION FOR ARABIC LETTER TO
+TU+0324=U+07A0
+tU+0324=U+07A0
+
+# THAANA EXTENSION FOR ARABIC LETTER ZO
+DU+0332=U+07A1
+dU+0332=U+07A1
+
+# THAANA EXTENSION FOR ARABIC LETTER AINU
+U+02BB=U+07A2
+
+# THAANA EXTENSION FOR ARABIC LETTER GHAINU
+GH=U+07A3
+Gh=U+07A3
+ghgh=U+0787U+07B0U+07A3
+gh=U+07A3
+
+# THAANA EXTENSION FOR ARABIC LETTER QAAFU 
+Q=U+07A4
+qq=U+0787U+07B0U+07A4
+q=U+07A4
+
+# THAANA EXTENSION FOR ARABIC LETTER WAAVU (NEWER LETTER)
+W=U+07A5
+ww=U+0787U+07B0U+07A5
+w=U+07A5
+
+# INITIAL (AND UPPERCASE) VOWELS THAT CONVERT
+# TO ALIF FOLLOWED BY VOWEL (ALIF OMITTED IN ROMANIZATION)
+AU+0304%=U+0787U+07A7
+A%=U+0787U+07A6
+U+0020aU+0304=U+0020U+0787U+07A7
+U+0020a=U+0020U+0787U+07A6
+EU+0304%=U+0787U+07AD
+E%=U+0787U+07AC
+U+0020eU+0304=U+0020U+0787U+07AD
+U+0020e=U+0020U+0787U+07AC
+IU+0304%=U+0787U+07A9
+I%=U+0787U+07A8
+U+0020iU+0304=U+0020U+0787U+07A9
+U+0020i=U+0020U+0787U+07A8
+OU+0304%=U+0787U+07AF
+O%=U+0787U+07AE
+U+0020oU+0304=U+0020U+0787U+07AF
+U+0020o=U+0020U+0787U+07AE
+UU+0304%=U+0787U+07AB
+U%=U+0787U+07AB
+U+0020uU+0304=U+0020U+0787U+07AB
+U+0020u=U+0020U+0787U+07AB
+
+# THAANA MEDIAL OR FINAL VOWELS OVER ANY CONSONANT
+# (THIS ASSUMES NO UPPERCASE VOWELS REMAIN)
+
+aU+0304=U+07A7
+a=U+07A6
+eU+0304=U+07AD
+e=U+07AC
+iU+0304=U+07A9
+i=U+07A8
+oU+0304=U+07AF
+o=U+07AE
+uU+0304=U+07AB
+u=U+07AB
+
+# THAANA SUKUN (SILENCE) MARK; ONLY GENERATED IN OTHER COMBINATIONS
+# =U+07B0
+
+[ScriptToRoman]
+FieldsIncluded=100 110 111 130 240 245 246 250 260 264 440 490 600 610 611 630 651 700 710 711 730 740 800 830
+SubfieldsAlwaysExcluded=uvxy0123456789
+OtherSubfieldsExcludedByTag=100/e 110/e 111/j 246/i 260/c 264/c 650/a 700/e 700/i 710/e 710/i 711/i 711/j 730/i
+
+# RDA boilerplate phrases not transliterated:
+Place of publication not identified=Place of publication not identified
+publisher not identified=publisher not identified
+date of publication not identified=date of publication not identified
+At head of title=At head of title
+
+# THAANA LETTER NOONU WITHOUT SUKUN FOLLOWED BY A CONSONANT
+# IS ROMANIZED AS "m"+ DOT ABOVE (0307) THEN THE CONSONANT
+# OTHERWISE LETTER NOONU MAPS TO "n"
+U+0782U+0780=mU+0307U+0780
+U+0782U+0781=mU+0307U+0781
+U+0782U+0783=mU+0307U+0783
+U+0782U+0784=mU+0307U+0784
+U+0782U+0785=mU+0307U+0785
+U+0782U+0786=mU+0307U+0786
+U+0782U+0788=mU+0307U+0788
+U+0782U+0789=mU+0307U+0789
+U+0782U+078A=mU+0307U+078A
+U+0782U+0791=mU+0307U+0791
+U+0782U+078B=mU+0307U+078B
+U+0782U+078C=mU+0307U+078C
+U+0782U+078D=mU+0307U+078D
+U+0782U+078E=mU+0307U+078E
+U+0782U+078F=mU+0307U+078F
+U+0782U+0790=mU+0307U+0790
+U+0782U+0792=mU+0307U+0792
+U+0782U+0794=mU+0307U+0794
+U+0782U+0795=mU+0307U+0795
+U+0782U+0796=mU+0307U+0796
+U+0782U+0797=mU+0307U+0797
+U+0782U+0798=mU+0307U+0798
+U+0782U+0799=mU+0307U+0799
+U+0782U+079A=mU+0307U+079A
+U+0782U+079B=mU+0307U+079B
+U+0782U+079C=mU+0307U+079C
+U+0782U+079D=mU+0307U+079D
+U+0782U+079E=mU+0307U+079E
+U+0782U+07A0=mU+0307U+07A0
+U+0782U+07A1=mU+0307U+07A1
+U+0782U+07A2=mU+0307U+07A2
+U+0782U+07A3=mU+0307U+07A3
+U+0782U+07A4=mU+0307U+07A4
+U+0782U+07A5=mU+0307U+07A5
+
+# THAANA FINAL ALIFU WITH SUKUN (SILENCE) MARK
+# IS ROMANIZED WITH "h"+LOW LINE (0332)
+U+0787U+07B0U+0020=hU+0332U+0020
+
+# THAANA SHAVIYANI WITH SUKUN (SILENCE) MARK
+# IS ROMANIZED WITH "h"+BREVE BELOW
+U+0781U+07B0=hU+032E
+
+U+0787U+07B0U+0780=hh
+U+0780=h
+
+# THAANA ALIF WITH SUKUN AND SHAVIYANI
+U+0787U+07B0U+0781=sU+0301sU+0301
+
+U+0787U+07B0=hU+0332
+U+0781=sU+0301
+
+U+0787U+07B0U+0782=nn
+U+0782=n
+
+U+0787U+07B0U+0783=rr
+U+0783=r
+
+U+0787U+07B0U+0784=bb
+U+0784=b
+
+U+0787U+07B0U+0785=lU+0323
+U+0785=lU+0323
+
+U+0787U+07B0U+0786=kk
+U+0786=k
+
+U+0787U+07B0U+0788=vv
+U+0788=v
+
+U+0787U+07B0U+0789=mm
+U+0789=m
+
+U+0787U+07B0U+078A=ff
+U+078A=f
+
+# THAANA LETTER "D/d" WITH DOT BELOW (0323) 
+U+0787U+07B0U+0791=dU+0323dU+0323
+U+0791=dU+0323
+
+U+0787U+07B0U+078B=dd
+U+078B=d
+
+# THAANA LETTER "T/t" WITH DOT BELOW (0323)
+U+078CU+07B0U+078C=tU+0324t
+U+0787U+07B0U+078C=tt
+U+0793=tU+0323
+U+078C=t
+
+U+0787U+07B0U+078D=ll
+U+078D=l
+
+U+0787U+07B0U+078E=gg
+U+078E=g
+
+# THAANA LETTER "N/n' WITH TILDE (0303)
+U+0787U+07B0U+078F=nU+0303nU+0303
+U+078F=nU+0303
+
+U+0787U+07B0U+0790=ss
+U+0790=s
+
+U+0787U+07B0U+0792=zz
+U+0792=z
+
+U+0787U+07B0U+0794=yy
+U+0794=y
+
+U+0787U+07B0U+0795=pp
+U+0795=p
+
+U+0787U+07B0U+0796=jj
+U+0796=j
+
+U+0787U+07B0U+0797=cc
+U+0797=c
+
+# THAANA EXTENSION FOR ARABIC LETTER TTAA
+U+0787U+07B0U+0798=thth
+U+0798=th
+
+# THAANA EXTENSION FOR ARABIC LETTER HHAA
+U+0787U+07B0U+0799=hU+0323hU+0323
+U+0799=hU+0323
+
+# THAANA EXTENSION FOR ARABIC LETTER KHAA
+U+0787U+07B0U+079A=khkh
+U+079A=kh
+
+# THAANA EXTENSION FOR ARABIC LETTER THAALU
+U+0787U+07B0U+079B=dhdh
+U+079B=dh
+
+# THAANA EXTENSION FOR ARABIC LETTER ZAA (NEWER LETTER)
+U+0787U+07B0U+079C=zU+0332zU+0332
+U+079C=zU+0332
+
+# THAANA EXTENSION FOR ARABIC LETTER SHEENU
+U+0787U+07B0U+079D=shsh
+U+079D=sh
+
+# THAANA EXTENSION FOR ARABIC LETTER SAADHU
+U+0787U=07B0U+079E=sU+0323sU+0323
+U+079E=sU+0323
+
+# THAANA EXTENSION FOR ARABIC LETTER TO
+U+0787U=07B0U+07A0=tU+0324tU+0324
+U+07A0=tU+0324
+
+# THAANA EXTENSION FOR ARABIC LETTER ZO
+U+0787U=07B0U+07A1=dU+0332dU+0332
+U+07A1=dU+0332
+
+# THAANA EXTENSION FOR ARABIC LETTER AINU
+U+0787U=07B0U+07A2=U+02BBU+02BB
+U+07A2=U+02BB
+
+# THAANA EXTENSION FOR ARABIC LETTER GHAINU
+U+0787U+07B0U+07A3=ghgh
+U+07A3=gh
+
+# THAANA EXTENSION FOR ARABIC LETTER QAAFU 
+U+0787U+07B0U+07A4=qq
+U+07A4=q
+
+# THAANA EXTENSION FOR ARABIC LETTER WAAVU (NEWER LETTER)
+U+0787U+07B0U+07A5=ww
+U+07A5=w
+
+# INITIAL VOWELS FOLLOWING ALIF (ALIF OMITTED IN ROMANIZATION)
+U+0020U+0787U+07A7=U+0020aU+0304
+U+0020U+0787U+07A6=U+0020a
+U+0020U+0787U+07AD=U+0020eU+0304
+U+0020U+0787U+07AC=U+0020e
+U+0020U+0787U+07A9=U+0020iU+0304
+U+0020U+0787U+07A8=U+0020i
+U+0020U+0787U+07AF=U+0020oU+0304
+U+0020U+0787U+07AE=U+0020o
+U+0020U+0787U+07AB=U+0020uU+0304
+U+0020U+0787U+07AB=U+0020u
+
+# THAANA ALIF APPEARING MEDIALLY WITH ANY VOWEL
+# IS ROMANIZED WITH APOSTROPHE FOLLOWED BY THE SAME VOWEL
+
+U+0787U+07A7=U+0027aU+0304
+U+0787U+07A6=U+0027a
+U+0787U+07AD=U+0027eU+0304
+U+0787U+07AC=U+0027e
+U+0787U+07A9=U+0027iU+0304
+U+0787U+07A8=U+0027i
+U+0787U+07AF=U+0027oU+0304
+U+0787U+07AE=U+0027o
+U+0787U+07AB=U+0027uU+0304
+U+0787U+07AB=U+0027u
+
+# THAANA MEDIAL OR FINAL VOWELS OVER CONSONANTS EXCEPT ALIF
+# THIS PRODUCES NO UPPERCASE UPPERCASE VOWELS
+U+07A7=aU+0304
+U+07A6=a
+U+07AD=eU+0304
+U+07AC=e
+U+07A9=iU+0304
+U+07A8=i
+U+07AF=oU+0304
+U+07AE=o
+U+07AB=uU+0304
+U+07AB=u
+
+# THAANA LETTER ALIF--ANY REMAINING AFTER CONVERSION MAP TO APOSTROPHE
+U+0787=U+0027
diff --git a/legacy/data/KurdishRomanization.cfg b/legacy/data/KurdishRomanization.cfg
new file mode 100644
index 0000000..76e0ba1
--- /dev/null
+++ b/legacy/data/KurdishRomanization.cfg
@@ -0,0 +1,125 @@
+# version 0.9.1
+# Original table by William Kopycki
+# Last updated 08 July 2009
+
+[General]
+Name=Kurdish
+ScriptCode=(3
+Truncation=%
+
+[RomanToScript]
+FieldsIncluded=100 110 111 130 240 245 246 250 260 440 490 600 610 611 630 651 700 710 711 730 740 800 830
+SubfieldsAlwaysExcluded=uvxy0123456789
+OtherSubfieldsExcludedByTag=650/a 260/c 246/i
+IncludeFormattingCharactersLcPattern=True
+
+# "Authorized" names:
+
+# Punctuation marks:
+# %=U+066A ; cannot transliterate the truncation character
+*=U+066D
+,=U+060C
+;=U+061B
+?=U+061F
+
+# Numbers (these should be Arabic-Indic digits from 0660-0669. We will use 06F0-06F9 for Persian and Urdu--WK)
+0=U+0660
+1=U+0661
+2=U+0662
+3=U+0663
+4=U+0664
+5=U+0665
+6=U+0666
+7=U+0667
+8=U+0668
+9=U+0669
+
+# Vowels and vowel/consonant combinations
+U+02BBE=U+0639U+0647U+200C
+U+02BBe=U+0639U+0647U+200C
+A=U+0626U+0627
+a=U+0627
+EU+0302=U+0626U+0647U+200C
+eU+0302=U+06CE
+E=U+0626U+0647U+200C
+e=U+0647U+200C
+IU+0302=U+0626U+064A
+
+# here is the "alif maksura" which otherwise serves as the "Persian yah U+06CC which is not valid in MARC-8 character set. 
+
+# THIS NEEDS TO BE ADJUSTED FOR "i[circumflex]y" and probably "e[circumflex]y combinations to = U+0649
+
+iU+0302U+0020=U+0649
+iU+0302=U+064A
+I=
+i=
+O=U+06C6
+o=U+06C6
+uU+0302=U+0648U+0648
+U=U+0626U+0648
+u=U+0648
+
+# Consonants:
+B=U+0628
+b=U+0628
+CU+0327=U+0686
+cU+0327=U+0686
+C=U+062C
+c=U+062C
+DU+0323=U+0636
+dU+0323=U+0636
+D=U+062F
+d=U+062F
+F=U+0641
+f=U+0641
+G=U+06AF
+g=U+06AF
+HU+0308=U+062D
+hU+0308=U+062D
+H=U+0647
+h=U+0647
+J=U+0698
+j=U+0698
+K=U+06A9
+k=U+06A9
+#L and l with stroke
+U+0141=U+06B5
+U+0142=U+06B5
+L=U+0644
+l=U+0644
+M=U+0645
+m=U+0645
+N=U+0646
+n=U+0646
+P=U+067E
+p=U+067E
+Q=U+0642
+q=U+0642
+RU+0304=U+0695
+rU+0304=U+0695
+R=U+0631
+r=U+0631
+SU+0323=U+0635
+sU+0323=U+0635
+SU+0327=U+0634
+sU+0327=U+0634
+S=U+0633
+s=U+0633
+TU+0323=U+0637
+tU+0323=U+0637
+T=U+062A
+t=U+062A
+V=U+06A8
+v=U+06A8
+W=U+0648
+w=U+0648
+XU+0308=U+063A
+xU+0308=U+063A
+X=U+062E
+x=U+062E
+Y=U+064A
+y=U+064A
+Z=U+0632
+z=U+0632
+
+[ScriptToRoman]
diff --git a/legacy/data/PersianRomanization.cfg b/legacy/data/PersianRomanization.cfg
index 5ddecbe..f647b64 100644
--- a/legacy/data/PersianRomanization.cfg
+++ b/legacy/data/PersianRomanization.cfg
@@ -5,13 +5,13 @@
 
 [General]
 Name=Persian
+ScriptCode=(3
 Truncation=%
 
 [RomanToScript]
 FieldsIncluded=100 110 111 130 240 245 246 250 260 264 440 490 600 610 611 630 651 700 710 711 730 740 800 830
 SubfieldsAlwaysExcluded=uvxy0123456789
 OtherSubfieldsExcludedByTag=100/e 110/e 111/j 246/i 260/c 264/c 650/a 700/e 700/i 710/e 710/i 711/i 711/j 730/i
-Subfield6Code=(3
 IncludeFormattingCharactersLcPattern=True
 
 # RDA boilerplate phrases not transliterated:
diff --git a/legacy/data/PushtoRomanization.cfg b/legacy/data/PushtoRomanization.cfg
index ae425f1..bc17bf4 100644
--- a/legacy/data/PushtoRomanization.cfg
+++ b/legacy/data/PushtoRomanization.cfg
@@ -5,13 +5,13 @@
 
 [General]
 Name=Pushto
+ScriptCode=(3
 Truncation=%
 
 [RomanToScript]
 FieldsIncluded=100 110 111 130 245 246 250 260 264 440 490 505 600 610 611 630 651 700 710 711 730 740 800 830
 SubfieldsAlwaysExcluded=uvxy0123456789
 OtherSubfieldsExcludedByTag=100/e 110/e 111/j 246/i 260/c 264/c 650/a 700/e 700/i 710/e 710/i 711/i 711/j 730/i
-Subfield6Code=(3
 IncludeFormattingCharactersLcPattern=True
 
 # RDA boilerplate phrases not transliterated:
diff --git a/legacy/data/UrduRomanization.cfg b/legacy/data/UrduRomanization.cfg
index 58c088a..a1e3d38 100644
--- a/legacy/data/UrduRomanization.cfg
+++ b/legacy/data/UrduRomanization.cfg
@@ -5,13 +5,13 @@
 
 [General]
 Name=Urdu
+ScriptCode=(3
 Truncation=%
 
 [RomanToScript]
 FieldsIncluded=100 110 111 130 240 245 246 250 260 264 440 490 505 600 610 611 630 651 700 710 711 730 740 800 830
 SubfieldsAlwaysExcluded=uvxy0123456789
 OtherSubfieldsExcludedByTag=100/e 110/e 111/j 246/i 260/c 264/c 650/a 700/e 700/i 710/e 710/i 711/i 711/j 730/i
-Subfield6Code=(3
 IncludeFormattingCharactersLcPattern=True
 
 # RDA boilerplate phrases not transliterated:
diff --git a/scriptshifter/tables/data/divehi_thaana.yml b/scriptshifter/tables/data/divehi_thaana.yml
new file mode 100644
index 0000000..24c513f
--- /dev/null
+++ b/scriptshifter/tables/data/divehi_thaana.yml
@@ -0,0 +1,435 @@
+general:
+  name: Divehi (Thaana)
+
+roman_to_script:
+  map:
+    # COMMON COMBINING CHARACTERS (always follow a base letter)
+
+    # combining acute \u0301
+    # combining tilde \u0303
+    # combining macron \u0304
+    # combining dot above \u0307
+    # combining dot below \u0323
+    # combining diaeresis below \u0324
+    # combining low line \u0332
+    # combining breve below \u032E
+    # ayn (spacing) \u02BB
+    # apostrophe (spacing) \u0027
+
+    # REGULAR LATIN ALPHABETIC CHARACTERS TO BE CONVERTED
+
+    # THAANA LETTER ALIF APPEARING MEDIALLY WITH ANY VOWEL IS ROMANIZED WITH APOSTROPHE
+    # ORIGINAL VOWEL ASSOCIATED WITH THE ALIF CANNOT BE REGENERATED DURING CONVERSION
+    "\u0027": "\u0787"
+
+    # THAANA LETTER SHAVIYANI WITH SUKUN(BREVE 032E)
+    "H\u032E": "\u0781\u07B0"
+    "hh": "\u0787\u07B0\u0780"
+    "h\u032E": "\u0781\u07B0"
+
+    # THAANA LETTER ALIFU FINAL WITH SUKUN (LOW LINE 0332)
+    "%H\u0332": "\u0787\u07B0"
+    "%h\u0332": "\u0787\u07B0"
+    "H": "\u0780"
+    "h": "\u0780"
+    "S\u0301": "\u0781"
+    "ss": "\u0787\u07B0\u0781"
+    "s\u0301": "\u0781"
+
+    # THAANA LETTER MEDIAL NOONU WITHOUT SUKUN (DOT ABOVE 0307)
+    "M\u0307": "\u0782"
+    "m\u0307": "\u0782"
+    "mm": "\u0787\u07B0\u0789"
+    "N": "\u0782"
+    "nn": "\u0787\u07B0\u0782"
+    "n": "\u0782"
+    "R": "\u0783"
+    "rr": "\u0787\u07B0\u0783"
+    "r": "\u0783"
+    "B": "\u0784"
+    "bb": "\u0787\u07B0\u0784"
+    "b": "\u0784"
+
+    # THAANA LETTER "L/l" WITH DOT BELOW (0323) 
+    "L\u0323": "\u0785"
+    "l\u0323": "\u0785"
+    "K": "\u0786"
+    "kk": "\u0787\u07B0\u0786"
+    "k": "\u0786"
+
+    # THAANA LETTER ALIF--OFTEN OMITTED IN ROMANIZATION (SEE VOWEL CONVERSION BELOW)
+    "\u0027": "\u0787"
+
+    "V": "\u0788"
+    "vv": "\u0787\u07B0\u0788"
+    "v": "\u0788"
+    "M": "\u0789"
+    "m": "\u0789"
+    "F": "\u078A"
+    "ff": "\u0787\u07B0\u078A"
+    "f": "\u078A"
+    # THAANA LETTER "D/d" WITH DOT BELOW (0323) 
+    "D\u0323": "\u0791"
+    "d\u0323": "\u0791"
+    "D": "\u078B"
+    "dd": "\u0787\u07B0\u078B"
+    "d": "\u078B"
+    # THAANA LETTER "T/t" WITH DOT BELOW (0323) 
+    "T\u0323": "\u0793"
+    "t\u0323": "\u0793"
+    "T\u0324T": "\u078C\u07B0\u078C"
+    "T\u0324t": "\u078C\u07B0\u078C"
+    "t\u0324t": "\u078C\u07B0\u078C"
+    "t\u0324T": "\u078C\u07B0\u078C"
+    "T": "\u078C"
+    "tt": "\u0787\u07B0\u078C"
+    "t": "\u078C"
+    "L": "\u078D"
+    "ll": "\u0787\u07B0\u078D"
+    "l": "\u078D"
+    "G": "\u078E"
+    "gg": "\u0787\u07B0\u078E"
+    "g": "\u078E"
+    # THAANA LETTER "N/n' WITH TILDE (0303)
+    "N\u0303": "\u078F"
+    "n\u0303": "\u078F"
+    "S": "\u0790"
+    "ss": "\u0787\u07B0\u0790"
+    "s": "\u0790"
+    "Z": "\u0792"
+    "zz": "\u0787\u07B0\u0792"
+    "z": "\u0792"
+    "Y": "\u0794"
+    "yy": "\u0787\u07B0\u0794"
+    "y": "\u0794"
+    "P": "\u0795"
+    "p": "\u0795"
+    "pp": "\u0787\u07B0\u0795"
+    "J": "\u0796"
+    "jj": "\u0787\u07B0\u0796"
+    "j": "\u0796"
+    "C": "\u0797"
+    "cc": "\u0787\u07B0\u0797"
+    "c": "\u0797"
+
+    # THAANA EXTENSION FOR ARABIC LOAN WORDS AND NAMES
+
+    # THAANA EXTENSION FOR ARABIC LETTER TTAA
+    "TH": "\u0798"
+    "Th": "\u0798"
+    "thth": "\u0787\u07B0\u0798"
+    "th": "\u0798"
+
+    # THAANA EXTENSION FOR ARABIC LETTER HHAA
+    "H\u0323": "\u0799"
+    "h\u0323": "\u0799"
+
+    # THAANA EXTENSION FOR ARABIC LETTER KHAA
+    "KH": "\u079A"
+    "Kh": "\u079A"
+    "khkh": "\u0787\u07B0\u079A"
+    "kh": "\u079A"
+
+    # THAANA EXTENSION FOR ARABIC LETTER THAALU
+    "DH": "\u079B"
+    "Dh": "\u079B"
+    "dhdh": "\u0787\u07B0\u079B"
+    "dh": "\u079B"
+
+    # THAANA EXTENSION FOR ARABIC LETTER ZAA (NEWER LETTER)
+    "Z\u0332": "\u079C"
+    "z\u0332": "\u079C"
+
+    # THAANA EXTENSION FOR ARABIC LETTER SHEENU
+    "SH": "\u079D"
+    "Sh": "\u079D"
+    "shsh": "\u0787\u07B0\u079D"
+    "sh": "\u079D"
+
+    # THAANA EXTENSION FOR ARABIC LETTER SAADHU
+    "S\u0323": "\u079E"
+    "s\u0323": "\u079E"
+
+    # THAANA EXTENSION FOR ARABIC LETTER TO
+    "T\u0324": "\u07A0"
+    "t\u0324": "\u07A0"
+
+    # THAANA EXTENSION FOR ARABIC LETTER ZO
+    "D\u0332": "\u07A1"
+    "d\u0332": "\u07A1"
+
+    # THAANA EXTENSION FOR ARABIC LETTER AINU
+    "\u02BB": "\u07A2"
+
+    # THAANA EXTENSION FOR ARABIC LETTER GHAINU
+    "GH": "\u07A3"
+    "Gh": "\u07A3"
+    "ghgh": "\u0787\u07B0\u07A3"
+    "gh": "\u07A3"
+
+    # THAANA EXTENSION FOR ARABIC LETTER QAAFU 
+    "Q": "\u07A4"
+    "qq": "\u0787\u07B0\u07A4"
+    "q": "\u07A4"
+
+    # THAANA EXTENSION FOR ARABIC LETTER WAAVU (NEWER LETTER)
+    "W": "\u07A5"
+    "ww": "\u0787\u07B0\u07A5"
+    "w": "\u07A5"
+
+    # INITIAL (AND UPPERCASE) VOWELS THAT CONVERT
+    # TO ALIF FOLLOWED BY VOWEL (ALIF OMITTED IN ROMANIZATION)
+    "A\u0304%": "\u0787\u07A7"
+    "A%": "\u0787\u07A6"
+    "\u0020a\u0304": "\u0020\u0787\u07A7"
+    "\u0020a": "\u0020\u0787\u07A6"
+    "E\u0304%": "\u0787\u07AD"
+    "E%": "\u0787\u07AC"
+    "\u0020e\u0304": "\u0020\u0787\u07AD"
+    "\u0020e": "\u0020\u0787\u07AC"
+    "I\u0304%": "\u0787\u07A9"
+    "I%": "\u0787\u07A8"
+    "\u0020i\u0304": "\u0020\u0787\u07A9"
+    "\u0020i": "\u0020\u0787\u07A8"
+    "O\u0304%": "\u0787\u07AF"
+    "O%": "\u0787\u07AE"
+    "\u0020o\u0304": "\u0020\u0787\u07AF"
+    "\u0020o": "\u0020\u0787\u07AE"
+    "U\u0304%": "\u0787\u07AB"
+    "U%": "\u0787\u07AB"
+    "\u0020u\u0304": "\u0020\u0787\u07AB"
+    "\u0020u": "\u0020\u0787\u07AB"
+
+    # THAANA MEDIAL OR FINAL VOWELS OVER ANY CONSONANT
+    # (THIS ASSUMES NO UPPERCASE VOWELS REMAIN)
+
+    "a\u0304": "\u07A7"
+    "a": "\u07A6"
+    "e\u0304": "\u07AD"
+    "e": "\u07AC"
+    "i\u0304": "\u07A9"
+    "i": "\u07A8"
+    "o\u0304": "\u07AF"
+    "o": "\u07AE"
+    "u\u0304": "\u07AB"
+    "u": "\u07AB"
+
+    # THAANA SUKUN (SILENCE) MARK; ONLY GENERATED IN OTHER COMBINATIONS
+    # "": "\u07B0"
+
+script_to_roman:
+  map:
+
+    # THAANA LETTER NOONU WITHOUT SUKUN FOLLOWED BY A CONSONANT
+    # IS ROMANIZED AS "m"+ DOT ABOVE (0307) THEN THE CONSONANT
+    # OTHERWISE LETTER NOONU MAPS TO "n"
+    "\u0782\u0780": "m\u0307\u0780"
+    "\u0782\u0781": "m\u0307\u0781"
+    "\u0782\u0783": "m\u0307\u0783"
+    "\u0782\u0784": "m\u0307\u0784"
+    "\u0782\u0785": "m\u0307\u0785"
+    "\u0782\u0786": "m\u0307\u0786"
+    "\u0782\u0788": "m\u0307\u0788"
+    "\u0782\u0789": "m\u0307\u0789"
+    "\u0782\u078A": "m\u0307\u078A"
+    "\u0782\u0791": "m\u0307\u0791"
+    "\u0782\u078B": "m\u0307\u078B"
+    "\u0782\u078C": "m\u0307\u078C"
+    "\u0782\u078D": "m\u0307\u078D"
+    "\u0782\u078E": "m\u0307\u078E"
+    "\u0782\u078F": "m\u0307\u078F"
+    "\u0782\u0790": "m\u0307\u0790"
+    "\u0782\u0792": "m\u0307\u0792"
+    "\u0782\u0794": "m\u0307\u0794"
+    "\u0782\u0795": "m\u0307\u0795"
+    "\u0782\u0796": "m\u0307\u0796"
+    "\u0782\u0797": "m\u0307\u0797"
+    "\u0782\u0798": "m\u0307\u0798"
+    "\u0782\u0799": "m\u0307\u0799"
+    "\u0782\u079A": "m\u0307\u079A"
+    "\u0782\u079B": "m\u0307\u079B"
+    "\u0782\u079C": "m\u0307\u079C"
+    "\u0782\u079D": "m\u0307\u079D"
+    "\u0782\u079E": "m\u0307\u079E"
+    "\u0782\u07A0": "m\u0307\u07A0"
+    "\u0782\u07A1": "m\u0307\u07A1"
+    "\u0782\u07A2": "m\u0307\u07A2"
+    "\u0782\u07A3": "m\u0307\u07A3"
+    "\u0782\u07A4": "m\u0307\u07A4"
+    "\u0782\u07A5": "m\u0307\u07A5"
+
+    # THAANA FINAL ALIFU WITH SUKUN (SILENCE) MARK
+    # IS ROMANIZED WITH "h"+LOW LINE (0332)
+    "\u0787\u07B0\u0020": "h\u0332\u0020"
+
+    # THAANA SHAVIYANI WITH SUKUN (SILENCE) MARK
+    # IS ROMANIZED WITH "h"+BREVE BELOW
+    "\u0781\u07B0": "h\u032E"
+
+    "\u0787\u07B0\u0780": "hh"
+    "\u0780": "h"
+
+    # THAANA ALIF WITH SUKUN AND SHAVIYANI
+    "\u0787\u07B0\u0781": "s\u0301s\u0301"
+
+    "\u0787\u07B0": "h\u0332"
+    "\u0781": "s\u0301"
+
+    "\u0787\u07B0\u0782": "nn"
+    "\u0782": "n"
+
+    "\u0787\u07B0\u0783": "rr"
+    "\u0783": "r"
+
+    "\u0787\u07B0\u0784": "bb"
+    "\u0784": "b"
+
+    "\u0787\u07B0\u0785": "l\u0323"
+    "\u0785": "l\u0323"
+
+    "\u0787\u07B0\u0786": "kk"
+    "\u0786": "k"
+
+    "\u0787\u07B0\u0788": "vv"
+    "\u0788": "v"
+
+    "\u0787\u07B0\u0789": "mm"
+    "\u0789": "m"
+
+    "\u0787\u07B0\u078A": "ff"
+    "\u078A": "f"
+
+    # THAANA LETTER "D/d" WITH DOT BELOW (0323) 
+    "\u0787\u07B0\u0791": "d\u0323d\u0323"
+    "\u0791": "d\u0323"
+
+    "\u0787\u07B0\u078B": "dd"
+    "\u078B": "d"
+
+    # THAANA LETTER "T/t" WITH DOT BELOW (0323)
+    "\u078C\u07B0\u078C": "t\u0324t"
+    "\u0787\u07B0\u078C": "tt"
+    "\u0793": "t\u0323"
+    "\u078C": "t"
+
+    "\u0787\u07B0\u078D": "ll"
+    "\u078D": "l"
+
+    "\u0787\u07B0\u078E": "gg"
+    "\u078E": "g"
+
+    # THAANA LETTER "N/n' WITH TILDE (0303)
+    "\u0787\u07B0\u078F": "n\u0303n\u0303"
+    "\u078F": "n\u0303"
+
+    "\u0787\u07B0\u0790": "ss"
+    "\u0790": "s"
+
+    "\u0787\u07B0\u0792": "zz"
+    "\u0792": "z"
+
+    "\u0787\u07B0\u0794": "yy"
+    "\u0794": "y"
+
+    "\u0787\u07B0\u0795": "pp"
+    "\u0795": "p"
+
+    "\u0787\u07B0\u0796": "jj"
+    "\u0796": "j"
+
+    "\u0787\u07B0\u0797": "cc"
+    "\u0797": "c"
+
+    # THAANA EXTENSION FOR ARABIC LETTER TTAA
+    "\u0787\u07B0\u0798": "thth"
+    "\u0798": "th"
+
+    # THAANA EXTENSION FOR ARABIC LETTER HHAA
+    "\u0787\u07B0\u0799": "h\u0323h\u0323"
+    "\u0799": "h\u0323"
+
+    # THAANA EXTENSION FOR ARABIC LETTER KHAA
+    "\u0787\u07B0\u079A": "khkh"
+    "\u079A": "kh"
+
+    # THAANA EXTENSION FOR ARABIC LETTER THAALU
+    "\u0787\u07B0\u079B": "dhdh"
+    "\u079B": "dh"
+
+    # THAANA EXTENSION FOR ARABIC LETTER ZAA (NEWER LETTER)
+    "\u0787\u07B0\u079C": "z\u0332z\u0332"
+    "\u079C": "z\u0332"
+
+    # THAANA EXTENSION FOR ARABIC LETTER SHEENU
+    "\u0787\u07B0\u079D": "shsh"
+    "\u079D": "sh"
+
+    # THAANA EXTENSION FOR ARABIC LETTER SAADHU
+    "\u0787U": "07B0\u079E=s\u0323s\u0323"
+    "\u079E": "s\u0323"
+
+    # THAANA EXTENSION FOR ARABIC LETTER TO
+    "\u0787U": "07B0\u07A0=t\u0324t\u0324"
+    "\u07A0": "t\u0324"
+
+    # THAANA EXTENSION FOR ARABIC LETTER ZO
+    "\u0787U": "07B0\u07A1=d\u0332d\u0332"
+    "\u07A1": "d\u0332"
+
+    # THAANA EXTENSION FOR ARABIC LETTER AINU
+    "\u0787U": "07B0\u07A2=\u02BB\u02BB"
+    "\u07A2": "\u02BB"
+
+    # THAANA EXTENSION FOR ARABIC LETTER GHAINU
+    "\u0787\u07B0\u07A3": "ghgh"
+    "\u07A3": "gh"
+
+    # THAANA EXTENSION FOR ARABIC LETTER QAAFU 
+    "\u0787\u07B0\u07A4": "qq"
+    "\u07A4": "q"
+
+    # THAANA EXTENSION FOR ARABIC LETTER WAAVU (NEWER LETTER)
+    "\u0787\u07B0\u07A5": "ww"
+    "\u07A5": "w"
+
+    # INITIAL VOWELS FOLLOWING ALIF (ALIF OMITTED IN ROMANIZATION)
+    "\u0020\u0787\u07A7": "\u0020a\u0304"
+    "\u0020\u0787\u07A6": "\u0020a"
+    "\u0020\u0787\u07AD": "\u0020e\u0304"
+    "\u0020\u0787\u07AC": "\u0020e"
+    "\u0020\u0787\u07A9": "\u0020i\u0304"
+    "\u0020\u0787\u07A8": "\u0020i"
+    "\u0020\u0787\u07AF": "\u0020o\u0304"
+    "\u0020\u0787\u07AE": "\u0020o"
+    "\u0020\u0787\u07AB": "\u0020u\u0304"
+    "\u0020\u0787\u07AB": "\u0020u"
+
+    # THAANA ALIF APPEARING MEDIALLY WITH ANY VOWEL
+    # IS ROMANIZED WITH APOSTROPHE FOLLOWED BY THE SAME VOWEL
+
+    "\u0787\u07A7": "\u0027a\u0304"
+    "\u0787\u07A6": "\u0027a"
+    "\u0787\u07AD": "\u0027e\u0304"
+    "\u0787\u07AC": "\u0027e"
+    "\u0787\u07A9": "\u0027i\u0304"
+    "\u0787\u07A8": "\u0027i"
+    "\u0787\u07AF": "\u0027o\u0304"
+    "\u0787\u07AE": "\u0027o"
+    "\u0787\u07AB": "\u0027u\u0304"
+    "\u0787\u07AB": "\u0027u"
+
+    # THAANA MEDIAL OR FINAL VOWELS OVER CONSONANTS EXCEPT ALIF
+    # THIS PRODUCES NO UPPERCASE UPPERCASE VOWELS
+    "\u07A7": "a\u0304"
+    "\u07A6": "a"
+    "\u07AD": "e\u0304"
+    "\u07AC": "e"
+    "\u07A9": "i\u0304"
+    "\u07A8": "i"
+    "\u07AF": "o\u0304"
+    "\u07AE": "o"
+    "\u07AB": "u\u0304"
+    "\u07AB": "u"
+
+    # THAANA LETTER ALIF--ANY REMAINING AFTER CONVERSION MAP TO APOSTROPHE
+    "\u0787": "\u0027"
diff --git a/scriptshifter/tables/data/dogri_devanagari.yml b/scriptshifter/tables/data/dogri_devanagari.yml
new file mode 100644
index 0000000..318dd4f
--- /dev/null
+++ b/scriptshifter/tables/data/dogri_devanagari.yml
@@ -0,0 +1,16 @@
+general:
+  name: Dogri (Devanagari)
+
+script_to_roman:
+  hooks:
+    post_config:
+      -
+        - aksharamukha.romanizer.s2r_post_config
+        - src_script: "Devanagari"
+
+roman_to_script:
+  hooks:
+    post_config:
+      -
+        - aksharamukha.romanizer.r2s_post_config
+        - dest_script: "Devanagari"
diff --git a/scriptshifter/tables/data/index.yml b/scriptshifter/tables/data/index.yml
index 99aa976..59ee728 100644
--- a/scriptshifter/tables/data/index.yml
+++ b/scriptshifter/tables/data/index.yml
@@ -44,6 +44,10 @@ chuvash_cyrillic:
   name: Chuvash (Cyrillic)
 devanagari:
   name: Devanagari
+divehi_thaana:
+  name: Divehi (Thaana)
+dogri_devanagari:
+  name: Dogri (Devanagari)
 dungan_cyrillic:
   name: Dungan (Cyrillic)
 ethiopic:
@@ -96,13 +100,15 @@ korean_names:
   description: Korean S2R for strings ONLY containing personal names formatted as last + first name. Separate multiple names with a comma or a center-dot (U+00B7).
 koryak_cyrillic:
   name: Koryak (Cyrillic)
+kurdish:
+  name: Kurdish
 kyrgyz_cyrillic:
   name: Kyrgyz (Cyrillic)
 lithuanian_cyrillic:
   name: Lithuanian (Cyrillic)
 macedonian:
   name: Macedonian
-marathi:
+marathi_devanagari:
   name: Marathi (Devanagari)
 mansi_cyrillic:
   name: Mansi (Cyrillic)
@@ -118,6 +124,10 @@ mordvin_cyrillic:
   name: Mordvin (Cyrillic)
 nenets_cyrillic:
   name: Nenets (Cyrillic)
+newari_devanagari:
+  name: Newari (Devanagari)
+nepali_devanagari:
+  name: Nepali (Devanagari)
 oriya:
   name: Oriya
 ossetic_cyrillic:
@@ -126,17 +136,23 @@ pali:
   name: Pali
 panjabi:
   name: Panjabi
-prakrit:
+persian:
+  name: Persian
+prakrit_devanagari:
   name: Prakrit (Devanagari)
 pulaar:
   name: Pulaar (Adlam)
+pushto:
+  name: Pushto
+rajasthani_devanagari:
+  name: Rajasthani (Devanagari)
 gurmukhi:
   name: Punjabi (Gurmukhi)
 romani_cyrillic:
   name: Romani (Cyrillic)
 russian:
   name: Russian
-sanskrit:
+sanskrit_devanagari:
   name: Sanskrit (Devanagari)
 serbian:
   name: Serbian
@@ -174,6 +190,8 @@ uighur_cyrillic:
   name: Uighur (Cyrillic)
 ukrainian:
   name: Ukrainian
+urdu:
+  name: Urdu
 uzbek_cyrillic:
   name: Uzbek (Cyrillic)
 yakut_cyrillic:
diff --git a/scriptshifter/tables/data/kurdish.yml b/scriptshifter/tables/data/kurdish.yml
new file mode 100644
index 0000000..1dfbea0
--- /dev/null
+++ b/scriptshifter/tables/data/kurdish.yml
@@ -0,0 +1,113 @@
+general:
+  name: Kurdish
+
+roman_to_script:
+  map:
+    # "Authorized" names:
+
+    # Punctuation marks:
+    "# %": "\u066A ; cannot transliterate the truncation character"
+    "*": "\u066D"
+    ",": "\u060C"
+    ";": "\u061B"
+    "?": "\u061F"
+
+    # Numbers (these should be Arabic-Indic digits from 0660-0669. We will use 06F0-06F9 for Persian and Urdu--WK)
+    "0": "\u0660"
+    "1": "\u0661"
+    "2": "\u0662"
+    "3": "\u0663"
+    "4": "\u0664"
+    "5": "\u0665"
+    "6": "\u0666"
+    "7": "\u0667"
+    "8": "\u0668"
+    "9": "\u0669"
+
+    # Vowels and vowel/consonant combinations
+    "\u02BBE": "\u0639\u0647\u200C"
+    "\u02BBe": "\u0639\u0647\u200C"
+    "A": "\u0626\u0627"
+    "a": "\u0627"
+    "E\u0302": "\u0626\u0647\u200C"
+    "e\u0302": "\u06CE"
+    "E": "\u0626\u0647\u200C"
+    "e": "\u0647\u200C"
+    "I\u0302": "\u0626\u064A"
+
+    # here is the "alif maksura" which otherwise serves as the "Persian yah \u06CC which is not valid in MARC-8 character set. 
+
+    # THIS NEEDS TO BE ADJUSTED FOR "i[circumflex]y" and probably "e[circumflex]y combinations to = \u0649"
+
+    "i\u0302\u0020": "\u0649"
+    "i\u0302": "\u064A"
+    "I": ""
+    "i": ""
+    "O": "\u06C6"
+    "o": "\u06C6"
+    "u\u0302": "\u0648\u0648"
+    "U": "\u0626\u0648"
+    "u": "\u0648"
+
+    # Consonants:
+    "B": "\u0628"
+    "b": "\u0628"
+    "C\u0327": "\u0686"
+    "c\u0327": "\u0686"
+    "C": "\u062C"
+    "c": "\u062C"
+    "D\u0323": "\u0636"
+    "d\u0323": "\u0636"
+    "D": "\u062F"
+    "d": "\u062F"
+    "F": "\u0641"
+    "f": "\u0641"
+    "G": "\u06AF"
+    "g": "\u06AF"
+    "H\u0308": "\u062D"
+    "h\u0308": "\u062D"
+    "H": "\u0647"
+    "h": "\u0647"
+    "J": "\u0698"
+    "j": "\u0698"
+    "K": "\u06A9"
+    "k": "\u06A9"
+    #L and l with stroke
+    "\u0141": "\u06B5"
+    "\u0142": "\u06B5"
+    "L": "\u0644"
+    "l": "\u0644"
+    "M": "\u0645"
+    "m": "\u0645"
+    "N": "\u0646"
+    "n": "\u0646"
+    "P": "\u067E"
+    "p": "\u067E"
+    "Q": "\u0642"
+    "q": "\u0642"
+    "R\u0304": "\u0695"
+    "r\u0304": "\u0695"
+    "R": "\u0631"
+    "r": "\u0631"
+    "S\u0323": "\u0635"
+    "s\u0323": "\u0635"
+    "S\u0327": "\u0634"
+    "s\u0327": "\u0634"
+    "S": "\u0633"
+    "s": "\u0633"
+    "T\u0323": "\u0637"
+    "t\u0323": "\u0637"
+    "T": "\u062A"
+    "t": "\u062A"
+    "V": "\u06A8"
+    "v": "\u06A8"
+    "W": "\u0648"
+    "w": "\u0648"
+    "X\u0308": "\u063A"
+    "x\u0308": "\u063A"
+    "X": "\u062E"
+    "x": "\u062E"
+    "Y": "\u064A"
+    "y": "\u064A"
+    "Z": "\u0632"
+    "z": "\u0632"
diff --git a/scriptshifter/tables/data/marathi.yml b/scriptshifter/tables/data/marathi_devanagari.yml
similarity index 100%
rename from scriptshifter/tables/data/marathi.yml
rename to scriptshifter/tables/data/marathi_devanagari.yml
diff --git a/scriptshifter/tables/data/nepali_devanagari.yml b/scriptshifter/tables/data/nepali_devanagari.yml
new file mode 100644
index 0000000..2aa20a2
--- /dev/null
+++ b/scriptshifter/tables/data/nepali_devanagari.yml
@@ -0,0 +1,16 @@
+general:
+  name: Nepali (Devanagari)
+
+script_to_roman:
+  hooks:
+    post_config:
+      -
+        - aksharamukha.romanizer.s2r_post_config
+        - src_script: "Devanagari"
+
+roman_to_script:
+  hooks:
+    post_config:
+      -
+        - aksharamukha.romanizer.r2s_post_config
+        - dest_script: "Devanagari"
diff --git a/scriptshifter/tables/data/newari_devanagari.yml b/scriptshifter/tables/data/newari_devanagari.yml
new file mode 100644
index 0000000..3c0d7fe
--- /dev/null
+++ b/scriptshifter/tables/data/newari_devanagari.yml
@@ -0,0 +1,16 @@
+general:
+  name: Newari (Devanagari)
+
+script_to_roman:
+  hooks:
+    post_config:
+      -
+        - aksharamukha.romanizer.s2r_post_config
+        - src_script: "Devanagari"
+
+roman_to_script:
+  hooks:
+    post_config:
+      -
+        - aksharamukha.romanizer.r2s_post_config
+        - dest_script: "Devanagari"
diff --git a/scriptshifter/tables/data/persian.yml b/scriptshifter/tables/data/persian.yml
new file mode 100644
index 0000000..edc54dc
--- /dev/null
+++ b/scriptshifter/tables/data/persian.yml
@@ -0,0 +1,309 @@
+general:
+  name: Persian
+
+roman_to_script:
+  map:
+    # Punctuation marks:
+    "# %": "\u066A ; cannot transliterate the truncation character"
+    "*": "\u066D"
+    ",": "\u060C"
+    ";": "\u061B"
+    "?": "\u061F"
+
+    # Exceptions for specific words 
+    # Allah
+    "Alla\u0304h": "\u0627\u0644\u0644\u0647"
+
+    # Qur'an
+    "Qur\u02BCa\u0304n": "\u0642\u0631\u0622\u0646"
+
+    # lillah
+    "lilla\u0304h": "\u0644\u0644\u0647"
+
+    # billah
+    "billa\u0304h": "\u0628\u0644\u0644\u0647"
+
+    # Rahman
+    "Rah\u0323ma\u0304n": "\u0631\u062D\u0645\u0646"
+
+    # ta'lif
+
+    # Ibn when it appears in the middle of a name sequence
+    "ibn": "\u0628\u0646"
+
+    # Parsing "sh[dot below] as in "Ishaq [name]"
+    "%sh\u0323%": "\u0633\u062D"
+
+    # Edition statements with Latin number
+    "Cha\u0304p-i 1": "\u0686\u0627\u067E 1"
+    "Cha\u0304p-i 2": "\u0686\u0627\u067E 2"
+    "Cha\u0304p-i 3": "\u0686\u0627\u067E 3"
+    "Cha\u0304p-i 4": "\u0686\u0627\u067E 4"
+    "Cha\u0304p-i 5": "\u0686\u0627\u067E 5"
+    "Cha\u0304p-i 6": "\u0686\u0627\u067E 6"
+    "Cha\u0304p-i 7": "\u0686\u0627\u067E 7"
+    "Cha\u0304p-i 8": "\u0686\u0627\u067E 8"
+    "Cha\u0304p-i 9": "\u0686\u0627\u067E 9"
+
+    # Numbers (Extended Arabic \u06F0-06F9 for Persian)
+    # currently *not* valid MARC21 characters
+    "# 0": "\u06F0"
+    "# 1": "\u06F1"
+    "# 2": "\u06F2"
+    "# 3": "\u06F3"
+    "# 4": "\u06F4"
+    "# 5": "\u06F5"
+    "# 6": "\u06F6"
+    "# 7": "\u06F7"
+    "# 8": "\u06F8"
+    "# 9": "\u06F9"
+
+    # Use Basic Arabic \u0660-0669, instead:
+    "0": "\u0660"
+    "1": "\u0661"
+    "2": "\u0662"
+    "3": "\u0663"
+    "4": "\u0664"
+    "5": "\u0665"
+    "6": "\u0666"
+    "7": "\u0667"
+    "8": "\u0668"
+    "9": "\u0669"
+
+    "# prime ": " ZWNJ"
+    "\u02B9a\u0304": "\u200C\u0622"
+    "\u02B9a": "\u200C\u0627"
+    "%\u02B9i\u0304": "\u200C\u0627\u0649"
+    "\u02B9i\u0304": "\u200C\u0627\u064A"
+    "i\u0304\u02B9a\u0304": "\u0649\u200C\u0622"
+    "i\u0304\u02B9a": "\u0649\u200C\u0627"
+    "i\u0304\u02B9u\u0304": "\u0649\u200C\u0622"
+    "i\u0304\u02B9u": "\u0649\u200C\u0627"
+    "i\u0304\u02B9": "\u0649\u200C"
+    "\u02B9": "\u200C"
+
+    # Vowel and vowel/consonant combinations
+    # and hyphenated suffixes:
+    # izafah here
+    "%h-\u02BEi": "\u06C0"
+    "%h-\u02BCi": "\u06C0 "
+    "%-\u02BEi": "\u06C0"
+    "%-\u02BCi": "\u06C0"
+
+    "%h-yi": "\u0647\u200C\u0649"
+    "%-yi": "\u0649"
+    "%yi": "\u0649"
+    "%\u02BEi\u0304": "\u0649"
+    "%\u02BCi\u0304": "\u0649"
+    "%i\u0304-i": "\u0649"
+    "%i\u0304": "\u0649"
+    "%ayy-i": "\u0649"
+    "%ay": "\u0649"
+    "%al-i": "\u0644"
+    "%-i": ""
+
+    # Hyphenated prefixes:
+    "wa-": "\u0648"
+    "bi-": "\u0628"
+    "al-": "\u0627\u0644"
+    "lil-": "\u0644\u0644"
+    "li-": "\u0644"
+    "ka-": "\u0643"
+
+    # ayn combo
+    "%\u02BBa\u0304%": "\u0639\u0627"
+    # alif combos
+    # [final position]
+    "%a\u0304\u02BE": "\u0627\u0621"
+    "%a\u0304\u02BC": "\u0627\u0621"
+    "%a\u0304\u02BEi\u0304": "\u0627\u0626\u0649"
+    "%a\u0304\u02BCi\u0304": "\u0627\u0626\u0649"
+    # [initial position]
+    "A\u0304\u02BEi\u0304%": "\u0622\u0626\u064A"
+    "A\u0304\u02BCi\u0304%": "\u0622\u0626\u064A"
+    "a\u0304\u02BEi\u0304%": "\u0622\u0626\u064A"
+    "a\u0304\u02BCi\u0304%": "\u0622\u0626\u064A"
+    # [medial position]
+    "a\u0304\u02BEi\u0304": "\u0627\u0626\u064A"
+    "a\u0304\u02BCi\u0304": "\u0627\u0626\u064A"
+    "a\u0304\u02BEi": "\u0627\u0626"
+    "a\u0304\u02BCi": "\u0627\u0626"
+    "a\u02BEi\u0304": "\u0626\u064A"
+    "a\u02BCi\u0304": "\u0626\u064A"
+    "u\u0304\u02BEi\u0304": "\u0648\u0626\u0649"
+    "u\u0304\u02BCi\u0304": "\u0648\u0626\u0649"
+
+    # A
+    "\u02BBA\u0304": "\u0639\u0627"
+    "\u02BBa\u0304": "\u0639\u0627"
+    "\u02BBA%": "\u0639"
+    "\u02BBa": "\u0639"
+    "A\u02BB%": "\u0627\u0639"
+    "a\u02BB%": "\u0627\u0639"
+    "A\u02BB": "\u0623\u0639"
+    "a\u02BB": "\u0639"
+    "a\u02BE": "\u0623"
+    "a\u02BC": "\u0623"
+    "\u02BEa": "\u0623"
+    "\u02BCa": "\u0623"
+    "a\u0304\u02BE": "\u0621"
+    "a\u0304\u02BC": "\u0621"
+    "A\u0304%": "\u0622"
+    "a\u0304%": "\u0622"
+    # previously an alif:
+    "A\u0304": "\u0622"
+    "a\u0304": "\u0627"
+    "A\u0301": "\u0649"
+    "a\u0301": "\u0649"
+    "ayy": "\u064A"
+    "A%": "\u0627"
+    "a%": "\u0627"
+    "A": "\u0627"
+    "a": ""
+
+    # I
+    "\u02BBI\u0304": "\u0639\u064A"
+    "\u02BBi\u0304": "\u0639\u064A"
+    "I\u02BB%": "\u0627\u0639"
+    "i\u02BB%": "\u0627\u0639"
+    "i\u02BB": "\u0639"
+    "\u02BBI%": "\u0639"
+    "I\u02BE%": "\u0627\u0626"
+    "I\u02BC%": "\u0627\u0626"
+    "i\u02BE": "\u0626"
+    "i\u02BC": "\u0626"
+    "\u02BEi\u0304": "\u0626\u0649"
+    "\u02BCi\u0304": "\u0626\u0649"
+    "\u02BEi": "\u0626"
+    "\u02BCi": "\u0626"
+    "I\u0304%": "\u0627\u064A"
+    "i\u0304%": "\u0627\u064A"
+    "i\u0304y": "\u064A"
+    "I\u0304": "\u0627\u0649"
+    "i\u0304": "\u064A"
+    "I%": "\u0627"
+    "i%": "\u0627"
+    "I": "\u0627"
+    "i": ""
+
+    # U
+    "\u02BEu\u0304": "\u0626\u0648"
+    "\u02BCu\u0304": "\u0626\u0648"
+    "U\u02BE%": "\u0627\u0624"
+    "U\u02BC%": "\u0627\u0624"
+    "u\u02BE": "\u0624"
+    "u\u02BC": "\u0624"
+    "U\u0304w%": "\u0627\u0628"
+    "u\u0304w": "\u0628"
+    "U\u0304": "\u0627\u0648"
+    "u\u0304": "\u0648"
+    "U%": "\u0627"
+    "u%": "\u0627"
+    "U": "\u0627"
+    "u": ""
+
+    # Consonants:
+    "B": "\u0628"
+    "bb": "\u0628"
+    "b": "\u0628"
+    "P": "\u067E"
+    "pp": "\u067E"
+    "p": "\u067E"
+    "T\u0323": "\u0637"
+    "t\u0323t\u0323": "\u0637"
+    "t\u0323": "\u0637"
+    "T": "\u062A"
+    "tt": "\u062A"
+    "t": "\u062A"
+    "Sh": "\u0634"
+    "shsh": "\u0634"
+    "sh": "\u0634"
+    "S\u0323": "\u0635"
+    "s\u0323s\u0323": "\u0635"
+    "s\u0323": "\u0635"
+    "S\u0332": "\u062B"
+    "ss\u0332": "\u062B"
+    "s\u0332": "\u062B"
+    "S": "\u0633"
+    "ss": "\u0633"
+    "s": "\u0633"
+    "J": "\u062C"
+    "jj": "\u062C"
+    "j": "\u062C"
+    "Ch": "\u0686"
+    "chch": "\u0686"
+    "ch": "\u0686"
+    "H\u0323": "\u062D"
+    "h\u0323h\u0323": "\u062D"
+    "h\u0323": "\u062D"
+    "H": "\u0647"
+    "hh": "\u0647"
+    "h": "\u0647"
+    "Kh": "\u062E"
+    "khkh": "\u062E"
+    "kh": "\u062E"
+    "K": "\u06A9"
+    "kk": "\u06A9"
+    "k": "\u06A9"
+    # Arabic d with dot below:
+    "D\u0323": "\u0636"
+    "d\u0323d\u0323": "\u0636"
+    "d\u0323": "\u0636"
+    "D": "\u062F"
+    "dd": "\u062F"
+    "d": "\u062F"
+    "Zh": "\u0698"
+    "zhzh": "\u0698"
+    "zh": "\u0698"
+    "Z\u0323": "\u0638"
+    "z\u0323z\u0323": "\u0638"
+    "z\u0323": "\u0638"
+    "Z\u0324": "\u0636"
+    "z\u0324z\u0324": "\u0636"
+    "z\u0324": "\u0636"
+    "Z\u0332": "\u0630"
+    "zz\u0332": "\u0630"
+    "z\u0332": "\u0630"
+    "Z": "\u0632"
+    "zz": "\u0632"
+    "z": "\u0632"
+    "R": "\u0631"
+    "rr": "\u0631"
+    "r": "\u0631"
+    "Gh": "\u063A"
+    "ghgh": "\u063A"
+    "gh": "\u063A"
+    "G": "\u06AF"
+    "g": "\u06AF"
+    "F": "\u0641"
+    "ff": "\u0641"
+    "f": "\u0641"
+    "Q": "\u0642"
+    "qq": "\u0642"
+    "q": "\u0642"
+    "L": "\u0644"
+    "ll": "\u0644"
+    "l": "\u0644"
+    "M": "\u0645"
+    "mm": "\u0645"
+    "m": "\u0645"
+    "N": "\u0646"
+    "nn": "\u0646"
+    "n": "\u0646"
+    "V": "\u0648"
+    "vv": "\u0648"
+    "v": "\u0648"
+    "W": "\u0648"
+    "ww": "\u0648"
+    "w": "\u0648"
+    "Y": "\u064A"
+    "y": "\u064A"
+
+    # non-Arabic consonants:
+
+    # ain (alone)
+    "\u02BB": "\u0639"
+
+    # hamza (alone in final position)
+    "%\u02BE": "\u0621"
+    "%\u02BC": "\u0621"
diff --git a/scriptshifter/tables/data/prakrit.yml b/scriptshifter/tables/data/prakrit_devanagari.yml
similarity index 100%
rename from scriptshifter/tables/data/prakrit.yml
rename to scriptshifter/tables/data/prakrit_devanagari.yml
diff --git a/scriptshifter/tables/data/pushto.yml b/scriptshifter/tables/data/pushto.yml
new file mode 100644
index 0000000..1dce5ea
--- /dev/null
+++ b/scriptshifter/tables/data/pushto.yml
@@ -0,0 +1,390 @@
+general:
+  name: Pushto
+
+roman_to_script:
+  map:
+    # Punctuation marks:
+    # %": "\u066A # cannot transliterate the truncation character"
+    "*": "\u066D"
+    ",": "\u060C"
+    ";": "\u061B"
+    "?": "\u061F"
+
+    # Exceptions for specific words 
+    # Allah
+    "Alla\u0304h": "\u0627\u0644\u0644\u0647"
+
+    # Qur'an
+    "Qur\u02BCa\u0304n": "\u0642\u0631\u0622\u0646"
+
+    # lillah
+    "lilla\u0304h": "\u0644\u0644\u0647"
+
+    # billah
+    "billa\u0304h": "\u0628\u0644\u0644\u0647"
+
+    # Rahman
+    "Rah\u0323ma\u0304n": "\u0631\u062D\u0645\u0646"
+
+    # ta'lif
+
+    # Ibn when it appears in the middle of a name sequence
+    "ibn": "\u0628\u0646"
+
+    # Parsing "sh[dot below] as in "Ishaq [name]"
+    "%sh\u0323%": "\u0633\u062D"
+
+    # Edition statements with Latin number
+    "Cha\u0304p-i 1": "\u0686\u0627\u067E 1"
+    "Cha\u0304p-i 2": "\u0686\u0627\u067E 2"
+    "Cha\u0304p-i 3": "\u0686\u0627\u067E 3"
+    "Cha\u0304p-i 4": "\u0686\u0627\u067E 4"
+    "Cha\u0304p-i 5": "\u0686\u0627\u067E 5"
+    "Cha\u0304p-i 6": "\u0686\u0627\u067E 6"
+    "Cha\u0304p-i 7": "\u0686\u0627\u067E 7"
+    "Cha\u0304p-i 8": "\u0686\u0627\u067E 8"
+    "Cha\u0304p-i 9": "\u0686\u0627\u067E 9"
+
+    # Numbers (Extended Arabic \u06F0-06F9 for Persian)
+    # currently *not* valid MARC21 characters
+    "# 0": "\u06F0"
+    "# 1": "\u06F1"
+    "# 2": "\u06F2"
+    "# 3": "\u06F3"
+    "# 4": "\u06F4"
+    "# 5": "\u06F5"
+    "# 6": "\u06F6"
+    "# 7": "\u06F7"
+    "# 8": "\u06F8"
+    "# 9": "\u06F9"
+
+    # Use Basic Arabic \u0660-0669, instead:
+    "0": "\u0660"
+    "1": "\u0661"
+    "2": "\u0662"
+    "3": "\u0663"
+    "4": "\u0664"
+    "5": "\u0665"
+    "6": "\u0666"
+    "7": "\u0667"
+    "8": "\u0668"
+    "9": "\u0669"
+
+    "# prime ": " ZWNJ"
+    "\u02B9a\u0304": "\u200C\u0622"
+    "\u02B9a": "\u200C\u0627"
+    "%\u02B9i\u0304": "\u200C\u0627\u0649"
+    "\u02B9i\u0304": "\u200C\u0627\u064A"
+    "i\u0304\u02B9": "\u0649\u200C"
+    "a\u0323y\u02B9": "\u06D3\u200C"
+    "\u02B9": "\u200C"
+
+    # Vowel and vowel/consonant combinations
+    # and hyphenated suffixes:
+    # izafah here
+    "%h-\u02BEi": "\u06C0"
+    "%h-\u02BCi": "\u06C0 "
+    "%-\u02BEi": "\u06C0"
+    "%-\u02BCi": "\u06C0"
+
+    "%h-yi": "\u0647\u200C\u0649"
+    "%-yi": "\u0649"
+    "%yi": "\u0649"
+    "%\u02BEi\u0304": "\u0649"
+    "%\u02BCi\u0304": "\u0649"
+    "%i\u0304-i": "\u0649"
+    "%i\u0304": "\u0649"
+
+    "%a\u0323h": "\u06C0"
+    "%ayy-i": "\u0649"
+    "%a\u0304y": "\u0627\u0649"
+    "%a\u0301": "\u0649\u0670"
+    "%al-i": "\u0644"
+    "%-i": ""
+
+    # Hyphenated prefixes:
+    "wa-": "\u0648"
+    "bi-": "\u0628"
+    "al-": "\u0627\u0644"
+    "lil-": "\u0644\u0644"
+    "li-": "\u0644"
+    "ka-": "\u0643"
+
+    # Diphthongs here
+    "Ayy%": "\u0627\u064A"
+    "ayy%": "\u0627\u064A"
+    "%a\u0323y": "\u06D3"
+    "%ay": "\u0649"
+    "\u02BBAw": "\u0639\u0648"
+    "\u02BBaw": "\u0639\u0648"
+    "Aw": "\u0627\u0648"
+    "aw": "\u0648"
+
+    # ayn combo
+    "%\u02BBa\u0304%": "\u0639\u0627"
+
+    # alif combos
+    # [final position]
+    "%a\u0304\u02BE": "\u0627\u0621"
+    "%a\u0304\u02BC": "\u0627\u0621"
+    "%a\u0304\u02BEi\u0304": "\u0627\u0626\u0649"
+    "%a\u0304\u02BCi\u0304": "\u0627\u0626\u0649"
+    # [initial position]
+    "A\u0304\u02BEi\u0304%": "\u0622\u0626\u064A"
+    "A\u0304\u02BCi\u0304%": "\u0622\u0626\u064A"
+    "a\u0304\u02BEi\u0304%": "\u0622\u0626\u064A"
+    "a\u0304\u02BCi\u0304%": "\u0622\u0626\u064A"
+
+    # [medial position]
+    "a\u0304\u02BEi\u0304": "\u0627\u0626\u064A"
+    "a\u0304\u02BCi\u0304": "\u0627\u0626\u064A"
+    "a\u0304\u02BEi": "\u0627\u0626"
+    "a\u0304\u02BCi": "\u0627\u0626"
+    "a\u02BEi\u0304": "\u0626\u064A"
+    "a\u02BCi\u0304": "\u0626\u064A"
+
+    #a [macron] hamza followed by e
+    "a\u0304\u02BEe": "\u0627\u0626\u064A"
+    "a\u0304\u02BCe": "\u0627\u0626\u064A"
+
+    # s[dot below]h
+    "S\u0323h": "\u069A"
+    "s\u0323hs\u0323h": "\u069A"
+    "s\u0323h": "\u069A"
+
+    # A
+    "\u02BBA\u0304": "\u0639\u0627"
+    "\u02BBa\u0304": "\u0639\u0627"
+    "\u02BBA%": "\u0639"
+    "\u02BBa": "\u0639"
+    "A\u02BB": "\u0623\u0639"
+    "a\u02BB": "\u0639"
+    "a\u02BE": "\u0623"
+    "a\u02BC": "\u0623"
+    "\u02BEa": "\u0623"
+    "\u02BCa": "\u0623"
+    "a\u0304\u02BE": "\u0621"
+    "a\u0304\u02BC": "\u0621"
+    "A\u0304%": "\u0622"
+    "a\u0304%": "\u0622"
+    "A\u0304": "\u0627"
+    "a\u0304": "\u0627"
+    "A\u0301": "\u0649"
+    "a\u0301": "\u0649"
+    "ayy": "\u064A"
+
+    # heh hamzah at end
+    "a\u0323h": "\u06C0"
+
+    "A%": "\u0627"
+    "a%": "\u0627"
+    "A": ""
+    "a": ""
+
+    # E
+    "%e": "\u06D0"
+    "E%": "\u0627\u064A"
+    "e%": "\u0627\u064A"
+    "e": "\u06D0"
+
+    # I 
+    "I\u02BB%": "\u0627\u0639"
+    "i\u02BB%": "\u0627\u0639"
+    "i\u02BB": "\u0639"
+    "\u02BBI%": "\u0639"
+    "I\u02BE%": "\u0627\u0626"
+    "I\u02BC%": "\u0627\u0626"
+    "i\u02BE": "\u0626"
+    "i\u02BC": "\u0626"
+    "\u02BEi": "\u0626"
+    "\u02BCi": "\u0626"
+    "I\u0304%": "\u0627\u064A"
+    "i\u0304%": "\u0627\u064A"
+    "i\u0304y": "\u064A"
+    "i\u0304": "\u064A"
+    "I%": "\u0627"
+    "i%": "\u0627"
+    "I": ""
+    "i": ""
+
+    # O
+    "o%": "\u0627\u0648"
+    "O%": "\u0627\u0648"
+    "o": "\u0648"
+
+    # U
+    # (u [macron] hamza combos)
+    "u\u0304\u02BE": "\u0624"
+    "u\u0304\u02BC": "\u0624"
+
+    "\u02BEu\u0304": "\u0626\u0648"
+    "\u02BCu\u0304": "\u0626\u0648"
+    "U\u02BE%": "\u0627\u0624"
+    "U\u02BC%": "\u0627\u0624"
+    "u\u02BE": "\u0624"
+    "u\u02BC": "\u0624"
+    "U\u0304w%": "\u0627\u0628"
+    "u\u0304w": "\u0628"
+    "U\u0304": "\u0627\u0648"
+    "u\u0304": "\u0648"
+    "U%": "\u0627"
+    "u%": "\u0627"
+    "U": ""
+    "u": ""
+
+    # Consonants:
+    "B": "\u0628"
+    "bb": "\u0628"
+    "b": "\u0628"
+    "P": "\u067E"
+    "pp": "\u067E"
+    "p": "\u067E"
+
+    "T\u0324": "\u0637"
+    "t\u0324t\u0324": "\u0637"
+    "t\u0324": "\u0637"
+
+    "T\u0323": "\u067C"
+    "t\u0323t\u0323": "\u067C"
+    "t\u0323": "\u067C"
+
+    "T": "\u062A"
+    "tt": "\u062A"
+    "t": "\u062A"
+
+    # s[dot]h used to be here
+
+    "Sh": "\u0634"
+    "shsh": "\u0634"
+    "sh": "\u0634"
+
+    "S\u0323": "\u0635"
+    "s\u0323s\u0323": "\u0635"
+    "s\u0323": "\u0635"
+
+    "S\u0332": "\u062B"
+    "s\u0332s\u0332": "\u062B"
+    "s\u0332": "\u062B"
+
+    "S\u0307": "\u0685"
+    "s\u0307s\u0307": "\u0685"
+    "s\u0307": "\u0685"
+
+    "S": "\u0633"
+    "ss": "\u0633"
+    "s": "\u0633"
+    "J": "\u062C"
+    "jj": "\u062C"
+    "j": "\u062C"
+    "Ch": "\u0686"
+    "chch": "\u0686"
+    "ch": "\u0686"
+    "H\u0323": "\u062D"
+    "h\u0323h\u0323": "\u062D"
+    "h\u0323": "\u062D"
+    "H": "\u0647"
+    "hh": "\u0647"
+    "h": "\u0647"
+    "Kh": "\u062E"
+    "khkh": "\u062E"
+    "kh": "\u062E"
+
+    "K": "\u06A9"
+    "kk": "\u06A9"
+    "k": "\u06A9"
+
+    # particle da
+    "Da\u0020": "\u062F\u0020"
+    "da\u0020": "\u062F\u0020"
+
+    "D\u0323": "\u0689"
+    "d\u0323d\u0323": "\u0689"
+    "d\u0323": "\u0689"
+
+    "D": "\u062F"
+    "dd": "\u062F"
+    "d": "\u062F"
+
+    "Z\u0323h": "\u0696"
+    "z\u0323hz\u0323h": "\u0696"
+    "z\u0323h": "\u0696"
+
+    "Zh": "\u0698"
+    "zhzh": "\u0698"
+    "zh": "\u0698"
+
+    "Z\u0324": "\u0638"
+    "z\u0324z\u0324": "\u0638"
+    "z\u0324": "\u0638"
+
+    "Z\u0323": "\u0636"
+    "z\u0323z\u0323": "\u0636"
+    "z\u0323": "\u0636"
+
+    "Z\u0332": "\u0630"
+    "z\u0332z\u0332": "\u0630"
+    "z\u0332": "\u0630"
+
+    "Z\u0307": "\u0681"
+    "z\u0307z\u0307": "\u0681"
+    "z\u0307": "\u0681"
+
+    "Z": "\u0632"
+    "zz": "\u0632"
+    "z": "\u0632"
+
+    "R\u0323": "\u0693"
+    "r\u0323r\u0323": "\u0693"
+    "r\u0323": "\u0693"
+
+    "R": "\u0631"
+    "rr": "\u0631"
+    "r": "\u0631"
+
+    "Gh": "\u063A"
+    "ghgh": "\u063A"
+    "gh": "\u063A"
+
+    "G": "\u06AB"
+    "gg": "\u06AB"
+    "g": "\u06AB"
+
+    "F": "\u0641"
+    "ff": "\u0641"
+    "f": "\u0641"
+
+    "Q": "\u0642"
+    "qq": "\u0642"
+    "q": "\u0642"
+
+    "L": "\u0644"
+    "ll": "\u0644"
+    "l": "\u0644"
+
+    "M": "\u0645"
+    "mm": "\u0645"
+    "m": "\u0645"
+
+    "N\u0323": "\u06BC"
+    "n\u0323n\u0323": "\u06BC"
+    "n\u0323": "\u06BC"
+
+    "N": "\u0646"
+    "nn": "\u0646"
+    "n": "\u0646"
+
+    "W": "\u0648"
+    "ww": "\u0648"
+    "w": "\u0648"
+
+    "Y": "\u064A"
+    "y": "\u064A"
+
+    # non-Arabic consonants:
+
+    # ain (alone)
+    "\u02BB": "\u0639"
+
+    # hamza (alone in final position)
+    "%\u02BE": "\u0621"
+    "%\u02BC": "\u0621"
diff --git a/scriptshifter/tables/data/rajasthani_devanagari.yml b/scriptshifter/tables/data/rajasthani_devanagari.yml
new file mode 100644
index 0000000..061130e
--- /dev/null
+++ b/scriptshifter/tables/data/rajasthani_devanagari.yml
@@ -0,0 +1,16 @@
+general:
+  name: Rajasthani (Devanagari)
+
+script_to_roman:
+  hooks:
+    post_config:
+      -
+        - aksharamukha.romanizer.s2r_post_config
+        - src_script: "Devanagari"
+
+roman_to_script:
+  hooks:
+    post_config:
+      -
+        - aksharamukha.romanizer.r2s_post_config
+        - dest_script: "Devanagari"
diff --git a/scriptshifter/tables/data/sanskrit.yml b/scriptshifter/tables/data/sanskrit_devanagari.yml
similarity index 100%
rename from scriptshifter/tables/data/sanskrit.yml
rename to scriptshifter/tables/data/sanskrit_devanagari.yml
diff --git a/scriptshifter/tables/data/urdu.yml b/scriptshifter/tables/data/urdu.yml
new file mode 100644
index 0000000..ba50220
--- /dev/null
+++ b/scriptshifter/tables/data/urdu.yml
@@ -0,0 +1,463 @@
+general:
+  name: Urdi
+
+roman_to_script:
+  map:
+    # Punctuation marks:
+    # "%": "\u066A"; cannot transliterate the truncation character
+    "*": "\u066D"
+    ",": "\u060C"
+    ";": "\u061B"
+    "?": "\u061F"
+
+    # Exceptions for specific words 
+    # Allah
+    "Alla\u0304h": "\u0627\u0644\u0644\u0647"
+    "alla\u0304h": "\u0627\u0644\u0644\u0647"
+
+    # Qur'an
+    "Qur\u02BCa\u0304n": "\u0642\u0631\u0622\u0646"
+    "qur\u02BCa\u0304n": "\u0642\u0631\u0622\u0646"
+
+    # aur (with spaces)
+    " aur ": " \u0627\u0648\u0631 "
+    "Aur ": "\u0627\u0648\u0631 "
+
+    ####
+    # Abdurrahman
+    "\u02BBAbdurrah\u0323ma\u0301n": "\u0639\u0628\u062F\u0627\u0644\u0631\u062D\u0645\u0670\u0646"
+    "\u02BBAbdurrah\u0323ma\u0304n": "\u0639\u0628\u062F\u0627\u0644\u0631\u062D\u0645\u0670\u0646"
+
+    # Abd names
+    "\u02BBAbdul\u02BB": "\u0639\u0628\u062F\u0627\u0644\u0639"
+    "\u02BBAbdula": "\u0639\u0628\u062F\u0627\u0644"
+    "\u02BBAbdulb": "\u0639\u0628\u062F\u0627\u0644\u0628"
+    "\u02BBAbdulf": "\u0639\u0628\u062F\u0627\u0644\u0641"
+    "\u02BBAbdulg\u0332h\u0332": "\u0639\u0628\u062F\u0627\u0644\u063A"
+    "\u02BBAbdulh\u0323": "\u0639\u0628\u062F\u0627\u0644\u062D"
+    "\u02BBAbdulh": "\u0639\u0628\u062F\u0627\u0644\u0647 "
+    "\u02BBAbdulj": "\u0639\u0628\u062F\u0627\u0644\u062C"
+    "\u02BBAbdulk\u0332h\u0332": "\u0639\u0628\u062F\u0627\u0644\u062E"
+    "\u02BBAbdulk": "\u0639\u0628\u062F\u0627\u0644\u0643 "
+    "\u02BBAbdulm": "\u0639\u0628\u062F\u0627\u0644\u0645"
+    "\u02BBAbdulq": "\u0639\u0628\u062F\u0627\u0644\u0642"
+    "\u02BBAbdulv": "\u0639\u0628\u062F\u0627\u0644\u0648"
+    "\u02BBAbdunn": "\u0639\u0628\u062F\u0627\u0644\u0646"
+    "\u02BBAbdurr": "\u0639\u0628\u062F\u0627\u0644\u0631"
+    "\u02BBAbdus\u0323s\u0323": "\u0639\u0628\u062F\u0627\u0644\u0325"
+    "\u02BBAbduss": "\u0639\u0628\u062F\u0627\u0644\u0633"
+    "\u02BBAbdushsh": "\u0639\u0628\u062F\u0627\u0644\u0634"
+    "\u02BBAbdutt": "\u0639\u0628\u062F\u0627\u0644\u062A"
+    "\u02BBAbduz\u0323z\u0323": "\u0639\u0628\u062F\u0627\u0644\u0636"
+    "\u02BBAbduz\u0324z\u0324": "\u0639\u0628\u062F\u0627\u0644\u0638"
+
+    # Abu names
+    "Abu\u0304 ": "\u0627\u0628\u0648\u0020"
+    "Abu\u0304": "\u0627\u0628\u0648\u200C\u0627\u0644"
+
+    ####
+
+    #lillah
+    "lilla\u0304h": "\u0644\u0644\u0647"
+
+    #billah
+    "billa\u0304h": "\u0628\u0644\u0644\u0647"
+
+    # Rahman
+    "Rah\u0323ma\u0304n": "\u0631\u062D\u0645\u0646"
+
+    # Nuzhat
+    "Nuzhat": "\u0646\u0632\u0647\u062A"
+
+    # Uddin names
+    "%i\u0304uddi\u0304n": "\u0649\u200C\u0627\u0644\u062F\u0651\u064A\u0646"
+    "%uddi\u0304n": "\u200C\u0627\u0644\u062F\u0651\u064A\u0646"
+
+    # ta'lif
+
+    # Ibn when it appears in the middle of a name sequence
+    "ibn": "\u0628\u0646"
+
+    # Abbreviated name elements
+    "# Ae": "\u0627\u06D2"
+
+    # Parsing "sh[dot below] as in "Ishaq [name]"
+    "%sh\u0323%": "\u0633\u062D"
+
+    # Numbers (\u06F0-06F9 for Persian/Urdu)
+    # currently *not* valid MARC21 characters
+    "# 0": "\u06F0"
+    "# 1": "\u06F1"
+    "# 2": "\u06F2"
+    "# 3": "\u06F3"
+    "# 4": "\u06F4"
+    "# 5": "\u06F5"
+    "# 6": "\u06F6"
+    "# 7": "\u06F7"
+    "# 8": "\u06F8"
+    "# 9": "\u06F9"
+
+    # Postpositions
+
+    # Aspirates [06BE] vs. heh [062D] combinations
+    "bh\u0323": "\u0628\u062D"
+    "Bh": "\u0628\u06BE"
+    "bh": "\u0628\u06BE"
+
+    "ph\u0323": "\u067E\u062D"
+    "Ph": "\u067E\u06BE"
+    "ph": "\u067E\u06BE"
+
+    "th\u0323": "\u062A\u062D"
+    "Th": "\u062A\u06BE"
+    "th": "\u062A\u06BE"
+
+    "t\u0323h\u0323": "\u0679\u062D"
+    "T\u0323h": "\u0679\u06BE"
+    "t\u0323h": "\u0679\u06BE"
+
+    "jh\u0323": "\u062C\u062D"
+    "Jh": "\u062C\u06BE"
+    "jh": "\u062C\u06BE"
+
+    "ch\u0323": "\u0686\u062D"
+    "Ch": "\u0686\u06BE"
+    "ch": "\u0686\u06BE"
+
+    "dh\u0323": "\u062F\u062D"
+    "Dh": "\u062F\u06BE"
+    "dh": "\u062F\u06BE"
+
+    "d\u0323h\u0323": "\u0688\u062D"
+    "D\u0323h": "\u0688\u06BE"
+    "d\u0323h": "\u0688\u06BE"
+
+    "r\u0323h\u0323": "\u0691\u062D"
+    "R\u0323h": "\u0691\u06BE"
+    "r\u0323h": "\u0691\u06BE"
+
+    "kh\u0323": "\u06A9\u062D"
+    "Kh": "\u06A9\u06BE"
+    "kh": "\u06A9\u06BE"
+
+    "gh\u0323": "\u06AF\u062D"
+    "Gh": "\u06AF\u06BE"
+    "gh": "\u06AF\u06BE"
+
+    # prime = ZWNJ"
+    "\u02B9A\u0304": "\u200C\u0622"
+    "\u02B9a\u0304": "\u200C\u0622"
+    "a\u0304\u02BC\u02B9": "\u0627\u0621\u200C"
+    "i\u0304\u02B9": "\u0649\u200C"
+    "\u02B9": "\u200C"
+
+    # Izafah here
+    "%a\u0304-yi": "\u0627\u0626\u06D2"
+    "%u\u0304-yi": "\u0648\u0626\u06D2"
+    "%o-yi": "\u0648\u0626\u06D2"
+    "%e-yi": "\u06D2"
+    "%i\u0304-yi": "\u0649"
+    "%h-yi": "\u06C0"
+    "%-yi": "\u06C0"
+    "%al-i": "\u0644"
+    "%ul-i": "\u0644"
+    "%-i": ""
+
+    # Hyphenated prefixes:
+    "bi-": "\u0628"
+    "al-a\u0304%": "\u0627\u0644\u0627"
+    "ul-a\u0304%": "\u0627\u0644\u0627"
+    "al-": "\u0627\u0644"
+    "ul-": "\u0627\u0644"
+    "lil-i": "\u0644\u0644"
+    "lil-": "\u0644\u0644"
+
+    # al-/ul- plus sun letters
+    "ar-r": "\u0627\u0644\u0631"
+    "ur-r": "\u0627\u0644\u0631"
+    "ar-R": "\u0627\u0644\u0631"
+    "ur-R": "\u0627\u0644\u0631"
+    "az\u0332-z\u0332": "\u0627\u0644\u0630"
+    "uz\u0332-z\u0332": "\u0627\u0644\u0630"
+    "az\u0332-Z\u0332": "\u0627\u0644\u0630"
+    "uz\u0332-Z\u0332": "\u0627\u0644\u0630"
+    "ad-d": "\u0627\u0644\u0627"
+    "ud-d": "\u0627\u0644\u0627"
+    "ad-D": "\u0627\u0644\u0627"
+    "ud-D": "\u0627\u0644\u0627"
+    "as\u0332-s\u0332": "\u0627\u0644\u062B"
+    "us\u0332-s\u0332": "\u0627\u0644\u062B"
+    "as\u0332-S\u0332": "\u0627\u0644\u062B"
+    "us\u0332-S\u0332": "\u0627\u0644\u062B"
+    "at-t": "\u0627\u0644\u062A"
+    "ut-t": "\u0627\u0644\u062A"
+    "at-T": "\u0627\u0644\u062A"
+    "ut-T": "\u0627\u0644\u062A"
+    "an-n": "\u0627\u0644\u0646"
+    "un-n": "\u0627\u0644\u0646"
+    "an-N": "\u0627\u0644\u0646"
+    "un-N": "\u0627\u0644\u0646"
+    "al-l": "\u0627\u0644\u0644"
+    "ul-l": "\u0627\u0644\u0644"
+    "al-L": "\u0627\u0644\u0644"
+    "ul-L": "\u0627\u0644\u0644"
+    "az\u0324-z\u0324": "\u0627\u0644\u0638"
+    "uz\u0324-z\u0324": "\u0627\u0644\u0638"
+    "az\u0324-Z\u0324": "\u0627\u0644\u0638"
+    "uz\u0324-Z\u0324": "\u0627\u0644\u0638"
+    "at\u0324-t\u0324": "\u0627\u0644\u0637"
+    "ut\u0324-t\u0324": "\u0627\u0644\u0637"
+    "at\u0324-T\u0324": "\u0627\u0644\u0637"
+    "ut\u0324-T\u0324": "\u0627\u0644\u0637"
+    "az\u0323-z\u0323": "\u0627\u0644\u0636"
+    "uz\u0323-z\u0323": "\u0627\u0644\u0636"
+    "az\u0323-Z\u0323": "\u0627\u0644\u0636"
+    "uz\u0323-Z\u0323": "\u0627\u0644\u0636"
+    "as\u0323-s\u0323": "\u0627\u0644\u0635"
+    "us\u0323-s\u0323": "\u0627\u0644\u0635"
+    "as\u0323-S\u0323": "\u0627\u0644\u0635"
+    "us\u0323-S\u0323": "\u0627\u0644\u0635"
+    "ash-sh": "\u0627\u0644\u0634"
+    "ush-sh": "\u0627\u0644\u0634"
+    "ash-Sh": "\u0627\u0644\u0634"
+    "ush-Sh": "\u0627\u0644\u0634"
+    "as-s": "\u0627\u0644\u0633"
+    "us-s": "\u0627\u0644\u0633"
+    "as-S": "\u0627\u0644\u0633"
+    "us-S": "\u0627\u0644\u0633"
+    "az-z": "\u0627\u0644\u0632"
+    "uz-z": "\u0627\u0644\u0632"
+    "az-Z": "\u0627\u0644\u0632"
+    "uz-Z": "\u0627\u0644\u0632"
+
+    # Diphthongs here
+    "Ae": "\u0627\u06D2"
+    "%ai": "\u06D2"
+    "Ai": "\u0627\u064A"
+    "ai%": "\u0627\u064A"
+    "ai": "\u064A"
+    "\u02BBAu": "\u0639\u0648"
+    "\u02BBau": "\u0639\u0648"
+    "Au": "\u0627\u0648"
+    "au": "\u0648"
+
+    # ayn-alif combo
+    "%\u02BBa\u0304\u02BE": "\u0639\u0627\u0621"
+    "%\u02BBa\u0304\u02BC": "\u0639\u0627\u0621"
+    "%\u02BBa\u0304%": "\u0639\u0627"
+
+    # hamza and vowel combo
+    # [in final position]
+    "%u\u0304\u02BEi\u0304": "\u0648\u0626\u0649"
+    "%u\u0304\u02BCi\u0304": "\u0648\u0626\u0649"
+    "%\u02BEi\u0304": "\u0626\u0649"
+    "%\u02BCi\u0304": "\u0626\u0649"
+    "%\u02BEe": "\u0626\u06D2"
+    "%\u02BCe": "\u0626\u06D2"
+
+    "%\u02BEu\u0304": "\u0624"
+    "%\u02BCu\u0304": "\u0624"
+    "%\u02BEo": "\u0624"
+    "%\u02BCo": "\u0624"
+
+    # [in medial position]
+    "a\u02BEa": "\u0623"
+    "a\u02BCa": "\u0623"
+    "a\u0304\u02BEa": "\u0627\u0621"
+    "a\u0304\u02BCa": "\u0627\u0621"
+    "a\u02BEa\u0304": "\u0622"
+    "a\u02BCa\u0304": "\u0622"
+    "o\u02BEi\u0304": "\u0648\u0626\u064A"
+    "o\u02BCi\u0304": "\u0648\u0626\u064A"
+    "o\u02BEi": "\u0648\u0626"
+    "o\u02BCi": "\u0648\u0626"
+    "\u02BEi\u0304": "\u0626\u064A"
+    "\u02BCi\u0304": "\u0626\u064A"
+    "\u02BEi": "\u0626"
+    "\u02BCi": "\u0626"
+    "\u02BEe": "\u0626\u064A"
+    "\u02BCe": "\u0626\u064A"
+
+    "\u02BEu\u0304": "\u0624"
+    "\u02BCu\u0304": "\u0624"
+    "u\u0304\u02BE": "\u0624"
+    "u\u0304\u02BC": "\u0624"
+    "\u02BEo": "\u0624"
+    "\u02BCo": "\u0624"
+    "o\u02BE": "\u0624"
+    "o\u02BC": "\u0624"
+    "au\u02BE": "\u0624"
+    "au\u02BC": "\u0624"
+
+    "\u02BEa": "\u0626"
+    "\u02BCa": "\u0626"
+
+    "%i\u0304": "\u0649"
+    "%a\u0301": "\u0649\u0670"
+
+    # A
+    "\u02BBA\u0304": "\u0639\u0627"
+    "\u02BBa\u0304": "\u0639\u0627"
+    "\u02BBA%": "\u0639"
+    "\u02BBa": "\u0639"
+    "A\u02BB": "\u0627\u0639"
+    "a\u02BB%": "\u0627\u0639"
+    "a\u02BB": "\u0639"
+    "A\u0304%": "\u0622"
+    "a\u0304%": "\u0622"
+    "a\u0304": "\u0627"
+    "a\u0301": "\u0649"
+    "ayy": "\u064A\u0651"
+    "A%": "\u0627"
+    "a%": "\u0627"
+    "A": ""
+    "a": ""
+
+    # E
+    "%e": "\u06D2"
+    "E%": "\u0627\u064A"
+    "e%": "\u0627\u064A"
+    "e": "\u064A"
+
+    # I
+    "\u02BBI\u0304": "\u0639\u064A"
+    "\u02BBi\u0304": "\u0639\u064A"
+    "I\u02BB": "\u0627\u0639"
+    "i\u02BB": "\u0639"
+    "\u02BBI": "\u0639"
+    "I\u0304%": "\u0627\u064A"
+    "i\u0304%": "\u0627\u064A"
+    "i\u0304y": "\u064A"
+    "i\u0304": "\u064A"
+    "iyy": "\u064A\u0651"
+    "I%": "\u0627"
+    "i%": "\u0627"
+    "I": "\u0627"
+    "i": ""
+
+    # O
+    "O%": "\u0627\u0648"
+    "o": "\u0648"
+
+    # U
+    "\u02BBu\u0304": "\u0639\u0648"
+    "\u02BBU": "\u0639"
+    "\u02BBu": "\u0639"
+    "U\u0304%": "\u0627\u0648"
+    "u\u0304%": "\u0627\u0648"
+    "u\u0304": "\u0648"
+    "U%": "\u0627"
+    "u%": "\u0627"
+    "U": ""
+    "u": ""
+
+    # Consonants:
+    "B": "\u0628"
+    "bb": "\u0628\u0651"
+    "b": "\u0628"
+    "P": "\u067E"
+    "pp": "\u067E\u0651"
+    "p": "\u067E"
+    "T\u0323": "\u0679"
+    "t\u0323t\u0323": "\u0679\u0651"
+    "t\u0323": "\u0679"
+    "T\u0324": "\u0637"
+    "t\u0324t\u0324": "\u0637\u0651"
+    "t\u0324": "\u0637"
+    "T": "\u062A"
+    "tt": "\u062A\u0651"
+    "t": "\u062A"
+    "Sh": "\u0634"
+    "shsh": "\u0634\u0651"
+    "sh": "\u0634"
+    "S\u0323": "\u0635"
+    "s\u0323s\u0323": "\u0635\u0651"
+    "s\u0323": "\u0635"
+    "S\u0332": "\u062B"
+    "s\u0332s\u0332": "\u062B\u0651"
+    "s\u0332": "\u062B"
+    "S": "\u0633"
+    "ss": "\u0633\u0651"
+    "s": "\u0633"
+    "J": "\u062C"
+    "jj": "\u062C\u0651"
+    "j": "\u062C"
+    "C": "\u0686"
+    "cc": "\u0686\u0651"
+    "c": "\u0686"
+    "H\u0323": "\u062D"
+    "h\u0323h\u0323": "\u062D\u0651"
+    "h\u0323": "\u062D"
+    "H": "\u0647"
+    "hh": "\u0647\u0651"
+    "h": "\u0647"
+    "K\u0332h\u0332": "\u062E"
+    "k\u0332h\u0332k\u0332h\u0332": "\u062E\u0651"
+    "k\u0332h\u0332": "\u062E"
+    "K": "\u06A9"
+    "kk": "\u06A9\u0651"
+    "k": "\u06A9"
+    "D\u0323": "\u0688"
+    "d\u0323d\u0323": "\u0688\u0651"
+    "d\u0323": "\u0688"
+    "D": "\u062F"
+    "dd": "\u062F\u0651"
+    "d": "\u062F"
+    "Z\u0324": "\u0638"
+    "z\u0324z\u0324": "\u0638\u0651"
+    "z\u0324": "\u0638"
+    "Z\u0323": "\u0636"
+    "z\u0323z\u0323": "\u0636\u0651"
+    "z\u0323": "\u0636"
+    "Z\u0332": "\u0630"
+    "z\u0332z\u0332": "\u0630\u0651"
+    "z\u0332": "\u0630"
+    "zz": "\u0632\u0651"
+    "Zh": "\u0698"
+    "zhzh": "\u0698\u0651"
+    "zh": "\u0698"
+    "Z": "\u0632"
+    "z": "\u0632"
+    "R\u0323": "\u0691"
+    "r\u0323r\u0323": "\u0691\u0651"
+    "r\u0323": "\u0691"
+    "R": "\u0631"
+    "rr": "\u0631\u0651"
+    "r": "\u0631"
+    "G\u0332h\u0332": "\u063A"
+    "g\u0332h\u0332g\u0332h\u0332": "\u063A\u0651"
+    "g\u0332h\u0332": "\u063A"
+    "G": "\u06AF"
+    "gg": "\u06AF\u0651"
+    "g": "\u06AF"
+    "F": "\u0641"
+    "ff": "\u0641\u0651"
+    "f": "\u0641"
+    "Q": "\u0642"
+    "qq": "\u0642\u0651"
+    "q": "\u0642"
+    "L": "\u0644"
+    "ll": "\u0644\u0651"
+    "l": "\u0644"
+    "M": "\u0645"
+    "mm": "\u0645\u0651"
+    "m": "\u0645"
+    "N\u0332": "\u06BA"
+    "n\u0332n\u0332": "\u06BA\u0651"
+    "n\u0332": "\u06BA"
+    "N": "\u0646"
+    "nn": "\u0646\u0651"
+    "n": "\u0646"
+    "V": "\u0648"
+    "vv": "\u0648\u0651"
+    "v": "\u0648"
+    "Y": "\u064A"
+    "yy": "\u064A\u0651"
+    "y": "\u064A"
+
+    # ain (alone)
+    "\u02BB": "\u0639"
+
+    # hamza (alone in final position)
+    "%\u02BE": "\u0621"
+    "%\u02BC": "\u0621"
diff --git a/tests/data/script_samples/tibetan.csv b/tests/data/script_samples/tibetan.csv
new file mode 100644
index 0000000..d976a0e
--- /dev/null
+++ b/tests/data/script_samples/tibetan.csv
@@ -0,0 +1,5 @@
+tibetan,བྱང་ཕྱོགས་བསྟན་འགྲོའི་སྐྱབས་མགོན་ཐམས་ཅད་མཁྱེན་པ་ཁལ་ཁ་ཨེར་ཏེ་ནེ་ཁུ་ཐག་ཐུ་བློ་བཟང་བསྟན་འཛིན་རྒྱལ་མཚན་གྱིའི་གསུང་འབུམ།,Byang phyogs bstan ‘gro’i skyabs mgon Thams-cad-mkhyen-pa Khal-kha Er-te-ne Khu-thag-thu Blo-bzang-bstan-‘dzin-rgyal-mtshan gyi’i gsung ʼbum,{"capitalize": "first"},
+tibetan,རྗེ་བཙུན་དམ་པ་སྐུ་ཕྲེང་བརྒྱད་པའི་གསུང་འབུམ,Rje-btsun-dam-pa sku phreng brgyad paʼi gsung ʼbum,{"capitalize": "first"},
+tibetan,རྗེ་བཙུན་ཐམས་ཅད་མཁྱེན་པ་དགེ་འདུན་རྒྱ་མཚོའི་གསུང་འབུམ་བཞུགས་སོ་,Rje-btsun Thams-cad-mkhyen-pa Dge-ʼdun-rgya-mtshoʼi gsung ʼbum bzhugs so,{"capitalize": "first"},
+tibetan,སྒྲུབ་ཐབས་འདོད་འཇོའི་བུམ་བཟང་གི་བརྒྱུད་པའི་རིམ་པ་ཕྱོགས་གཅིག་ཏུ་བསྡེབས་པ་བཞུགས་སོ།,Sgrub thabs ʼdod ʼjoʼi bum bzang gi brgyud paʼi rim pa phyogs gcig tu bsdebs pa bzhugs so,{"capitalize": "first"},
+tibetan,བཀའ་གདམས་ཀྱི་སྐྱེས་བུ་དམ་པ་རྣམས་ཀྱི་གསུང་བགྲོས་ཐོར་བུ་རྣམས་བཞུགས་སོ།,Bkaʼ gdams kyi skyes bu dam pa rnams kyi gsung bgros thor bu rnams bzhugs so,{"capitalize": "first"},