# FOMA script for ku-latn and ku-arab conversions
# Generates .att files when run from this directory using:
#   foma -f ku.foma
# Then generate executable .pfst files with:
#   ../tools/build-langconv-fst.js -l ku-latn ku-arab
#   ../tools/build-langconv-fst.js -l ku-arab ku-latn
echo EXPECT 12

source brackets.foma
source safety.foma

define CONVERT'KU'LATN NO'BRACKETS .o. [
  {ب} @-> {b} ,, {ج} @-> {c} ,, {چ} @-> {ç} ,, {د} @-> {d} ,, {ف} @-> {f} ,, {گ} @-> {g} ,, {ھ} @-> {h} ,,
  {ہ} @-> {h} ,, {ه} @-> {h} ,, {ح} @-> {h} ,, {ژ} @-> {j} ,, {ك} @-> {k} ,, {ک} @-> {k} ,, {ل} @-> {l} ,,
  {م} @-> {m} ,, {ن} @-> {n} ,, {پ} @-> {p} ,, {ق} @-> {q} ,, {ر} @-> {r} ,, {س} @-> {s} ,, {ش} @-> {ş} ,,
  {ت} @-> {t} ,, {ڤ} @-> {v} ,, {خ} @-> {x} ,, {غ} @-> {x} ,, {ز} @-> {z} ,,

# ک و => ku -- ist richtig
#  و ك=> ku -- ist auch richtig

  # Doppel- und Halbvokale
  {ڵ} @-> {ll} ,, # ll
  {ڕ} @-> {rr} ,, # rr
  {ا} @-> {a} ,,
  # {ئێ} @-> {ê} ,, # initial e
  {ە} @-> {e} ,,
  {ه‌} @-> {e} ,, # with one non-joiner
  {ه‌‌} @-> {e} ,, # with two non-joiner
  {ة} @-> {e} ,,
  {ێ} @-> {ê} ,,
  {ي} @-> {î} ,,
  {ی} @-> {î} ,, # U+06CC  db 8c  ARABIC LETTER FARSI YEH
  {ى} @-> {î} ,, # U+0649  d9 89  ARABIC LETTER ALEF MAKSURA
  {ۆ} @-> {o} ,,
  {و} @-> {w} ,,
  {ئ} @-> 0 ,, # initial hemze should not be shown
  {،} @-> {,} ,,
  {ع} @-> {'} ,, # ayn '
  {؟} @-> {?} ,,

  # digits
  {٠} @-> {0} ,, # ٠
  {١} @-> {1} ,, # ١
  {٢} @-> {2} ,, # ٢
  {٣} @-> {3} ,, # ٣
  {٤} @-> {4} ,, # ٤
  {٥} @-> {5} ,, # ٥
  {٦} @-> {6} ,, # ٦
  {٧} @-> {7} ,, # ٧
  {٨} @-> {8} ,, # ٨
  {٩} @-> {9}    # ٩
];

# XXX This table imported from PHP has duplicate entries for H and K :(
define CONVERT'KU'ARAB NO'BRACKETS .o. [
  {b} @-> {ب} ,, {c} @-> {ج} ,, {ç} @-> {چ} ,, {d} @-> {د} ,, {f} @-> {ف} ,, {g} @-> {گ} ,,
  {h} @-> {ه} ,, {j} @-> {ژ} ,, {k} @-> {ک} ,, {l} @-> {ل} ,,
  {m} @-> {م} ,, {n} @-> {ن} ,, {p} @-> {پ} ,, {q} @-> {ق} ,, {r} @-> {ر} ,, {s} @-> {س} ,, {ş} @-> {ش} ,,
  {t} @-> {ت} ,, {v} @-> {ڤ} ,,
  {x} @-> {خ} ,, {y} @-> {ی} ,, {z} @-> {ز} ,,

  {B} @-> {ب} ,, {C} @-> {ج} ,, {Ç} @-> {چ} ,, {D} @-> {د} ,, {F} @-> {ف} ,, {G} @-> {گ} ,,
  #{H} @-> {ھ} ,, # AMBIGUOUS IN PHP
  #{H} @-> {ہ} ,, # AMBIGUOUS IN PHP
  #{H} @-> {ه} ,, # AMBIGUOUS IN PHP
  {H} @-> {ح} ,,
  {J} @-> {ژ} ,,
  # {K} @-> {ك} ,, # AMBIGUOUS IN PHP
  {K} @-> {ک} ,,
  {L} @-> {ل} ,,
  {M} @-> {م} ,, {N} @-> {ن} ,, {P} @-> {پ} ,, {Q} @-> {ق} ,, {R} @-> {ر} ,, {S} @-> {س} ,, {Ş} @-> {ش} ,,
  {T} @-> {ت} ,, {V} @-> {ڤ} ,, {W} @-> {و} ,, {X} @-> {خ} ,,
  {Y} @-> {ی} ,, {Z} @-> {ز} ,,

  ## Doppelkonsonanten
  # {ll} @-> {ڵ} ,, # wenn es geht, doppel-l und l getrennt zu behandeln
  # {rr} @-> {ڕ} ,, # selbiges für doppel-r

  ## Einzelne Großbuchstaben
  # { C} @-> {ج} ,,

  ## Vowels
  {a} @-> {ا} ,,
  {e} @-> {ە} ,,
  {ê} @-> {ێ} ,,
  {i} @-> 0,
  {î} @-> {ی} ,,
  {o} @-> {ۆ} ,,
  {u} @-> {و} ,,
  {û} @-> {وو} ,,
  {w} @-> {و} ,,
  {,} @-> {،} ,,
  {?} @-> {؟} ,,

  ## Try to replace the leading vowel
  { a} @-> {ئا } ,,
  { e} @-> {ئە } ,,
  { ê} @-> {ئێ } ,,
  { î} @-> {ئی } ,,
  { o} @-> {ئۆ } ,,
  { u} @-> {ئو } ,,
  { û} @-> {ئوو } ,,
  {A} @-> {ئا} ,,
  {E} @-> {ئە} ,,
  {Ê} @-> {ئێ} ,,
  {Î} @-> {ئی} ,,
  {O} @-> {ئۆ} ,,
  {U} @-> {ئو} ,,
  {Û} @-> {ئوو} ,,
  { A} @-> {ئا } ,,
  { E} @-> {ئە } ,,
  { Ê} @-> {ئێ } ,,
  { Î} @-> {ئی } ,,
  { O} @-> {ئۆ } ,,
  { U} @-> {ئو } ,,
  { Û} @-> {ئوو }
  # eyn erstmal deaktivieren, einfache Anführungsstriche sind einfach zu
  # häufig, um sie als eyn zu interpretieren.
  # {'} @-> {ع} ,, # '

  ### deactivated for now, breaks links i.e. in header of Special:Recentchanges :-(
  ## digits
  #{0} @-> {٠} ,, # ٠
  #{1} @-> {١} ,, # ١
  #{2} @-> {٢} ,, # ٢
  #{3} @-> {٣} ,, # ٣
  #{4} @-> {٤} ,, # ٤
  #{5} @-> {٥} ,, # ٥
  #{6} @-> {٦} ,, # ٦
  #{7} @-> {٧} ,, # ٧
  #{8} @-> {٨} ,, # ٨
  #{9} @-> {٩} ,, # ٩
];

# Generate and save the transducers
regex TEST'UPPER'UNIVERSAL(CONVERT'KU'LATN);
echo EXPECT[trans-ku-latn universal]: 1 (1 = TRUE, 0 = FALSE)
test null
pop stack

regex CONVERT'KU'LATN;
minimize net
echo EXPECT[trans-ku-latn functional]: 1 (1 = TRUE, 0 = FALSE)
test functional
byte machine
write att trans-ku-latn.att
pop stack

regex LONGSAFE'NOT'OK(CONVERT'KU'LATN .o. CONVERT'KU'ARAB);
echo EXPECT[brack-ku-latn-ku-arab safe]: 1 (1 = TRUE, 0 = FALSE)
test null
print shortest-string
pop stack

regex LONGSAFE(CONVERT'KU'LATN .o. CONVERT'KU'ARAB);
minimize net
echo EXPECT[brack-ku-latn-ku-arab functional]: 1 (1 = TRUE, 0 = FALSE)
test functional
byte machine
write att brack-ku-latn-ku-arab.att
pop stack

regex LONGSAFE'NOT'OK(CONVERT'KU'LATN .o. IDENTITY);
echo EXPECT[brack-ku-latn-noop safe]: 1 (1 = TRUE, 0 = FALSE)
test null
print shortest-string
pop stack

regex LONGSAFE(CONVERT'KU'LATN .o. IDENTITY);
minimize net
echo EXPECT[brack-ku-latn-noop functional]: 1 (1 = TRUE, 0 = FALSE)
test functional
byte machine
write att brack-ku-latn-noop.att
pop stack


regex TEST'UPPER'UNIVERSAL(CONVERT'KU'ARAB);
echo EXPECT[trans-ku-arab universal]: 1 (1 = TRUE, 0 = FALSE)
test null
pop stack

regex CONVERT'KU'ARAB;
minimize net
echo EXPECT[trans-ku-arab functional]: 1 (1 = TRUE, 0 = FALSE)
test functional
byte machine
write att trans-ku-arab.att
pop stack

regex LONGSAFE'NOT'OK(CONVERT'KU'ARAB .o. CONVERT'KU'LATN);
echo EXPECT[brack-ku-arab-ku-latn safe]: 1 (1 = TRUE, 0 = FALSE)
test null
print shortest-string
pop stack

regex LONGSAFE(CONVERT'KU'ARAB .o. CONVERT'KU'LATN);
minimize net
echo EXPECT[brack-ku-arab-ku-latn functional]: 1 (1 = TRUE, 0 = FALSE)
test functional
byte machine
write att brack-ku-arab-ku-latn.att
pop stack

regex LONGSAFE'NOT'OK(CONVERT'KU'ARAB .o. IDENTITY);
echo EXPECT[brack-ku-arab-noop safe]: 1 (1 = TRUE, 0 = FALSE)
test null
print shortest-string
pop stack

regex LONGSAFE(CONVERT'KU'ARAB .o. IDENTITY);
minimize net
echo EXPECT[brack-ku-arab-noop functional]: 1 (1 = TRUE, 0 = FALSE)
test functional
byte machine
write att brack-ku-arab-noop.att
pop stack