# =============================================================
# CONTENT: Finite-State Tokenizer (no multi-words)
# AUTHOR: Original version by Anne Schiller, copied from the
# Karttunen/Beesley book on finite state grammar
# Modified and extended by Trond Trosterud, 2001-2003.
# CREATED: 12-Jun-1997
# UPDATED: 05-Sep-2001 - 15-Feb-2003
# =============================================================
# Usage: xfst -f [ThisFile]
# =============================================================
# Thomas testing... delete
clear stack
echo >>> define white space
define SP " ";
define TAB "\t";
define NL "\n";

define WS [SP|NL|TAB];

# =============================================================
echo >>> define single character symbols
define SINGLE [ %" | %« | %» | %. | %, | %; | %: | %! | %?
| %( | %) | %[ | %] | %{ | %} | %/ | %% 
| %' ];
# Added %' to include 'dat' pro "dat", this may crash due to the grade 3 mark.
# The parser errouneously #stops# for a % mark, awaits the next one
# (cf. trying 987% in xfst lookup mode).

# Added the string ' | %/' above, for ja/dahje cases.
define PUNCT [ %. %. (%.) | %' %' | %' %' | %, %, ] ;
define Char \[ WS | SINGLE ] ;

# =============================================================
echo >>> define SYMBOL
#define SYMBOL [ SINGLE | PUNCT ] ;
define SYMBOL [ SINGLE | PUNCT | EXTRAPERIOD ] ;

echo >>> define WORD
define WORD [ Char ]+ ;

# =============================================================
echo >>> regular abbreviations

define Capital [
Č|Đ|Ŋ|Š|Ŧ|Ž|
A|B|C|D|E|F|G|H|I|J|K|L|M|N|O|P|Q|R|S|T|U|V|W|X|Y|Z|
Å|Ä|Ö|Ø|Æ|Á|É|Ó|Ú|Í|À|È|Ò|Ù|Ì|Ë|Ü|Ï|Â|Ê|Ô|Û|Î|Ã|Ý|þ|Ñ|Ð ] ;

define Small [ č|đ|ŋ|š|ŧ|ž|
a|b|c|d|e|f|g|h|i|j|k|l|m|n|o|p|q|r|s|t|u|v|w|x|y|z|
å|ä|ö|ø|æ|á|é|ó|ú|í|à|è|ò|ù|ì|ë|ü|ï|â|ê|ô|û|î|ã|ý|þ|ñ|ð|ç|ß];

# =============================================================
echo >>> numeric expressions
define Digit [ %0 | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9];
define NumOp [ %- | %+ | %* | %/ | %= | %: ];
define NumSep [ %. | %, | %: ]; # %: added to original, for 1993:18 etc.
!define NUM [ Digit | NumOp | NumSep]+ & $[Digit] ;
define NUM [ Digit | NumOp | NumSep]+ Digit+ ;
define NUMOrd NUM %. ;
define ROM [ I|V|X ]+ ; # does not work :-(
define NUMROM [ NUM | ROM ] ;
define NUMCASED [ NUM+ %: Small+ ] ;
# Does this mean a string of at least one of Digit or NumOp or NumSep, 
# followed by at least one Digit?
# It seems that the preproc errouneously leaves NumOp and NumSep on the
# numbers, instead of separating (it splits 987" but not 987: or 987.
# Thus, it errouneously tells that a number cannot close a sentence.

# =============================================================
define Letter [ Small | Capital ] ;
#define INIT [ Capital %. ]+ ;
define INIT [ Letter %. ]+ ; # Thus, INIT not only for J. Smith
                             # but also for j. Smith and s. 43.
define CAPNUM [ Capital | NUM ] ;

# =============================================================
echo >>> list of abbreviations
#<cn> There are four groups of abbreviations:
#<++> INTRANSABBR - ends a snt when foll. by capital letter or number.
#<-+> INTRANSNUMABBR - ends a snt when foll. by number.
#<+-> INTRANSCAPABBR - ends a snt when foll. by capital letter.
#<--> TRANSABBR - does not end a snt.

# Many of the abbreviations are commented out, awaiting
# a more thorough evaluation, and, while waiting, in
# order to shorten compilation time.
# They should be re-introduced on an individual
# basis.

# <+c,+n> (+c,+n means clb if followed by capital,number)
# Abbreviations that end a sentence if 
# followed by letter in upper case or a number.
# These are typically nouns.

define INTRANSABBR [
#{A.L.} | 
{A.S.} | 
{Eftf.} |
#{Fr.P.} | 
{Fr.p.} | 
{Fuom.} | 
#{Inc.} |
#{K.S.} | 
{Kbh.} | 
#{Kr.F.} | 
{Kr.f.} |
{Krf.} | 
{Ltd.} | 
#{N.A.F.} | 
{NB.} | 
{O.S.} | 
{Sth.} | 
{adj.} | 
#{adm.} | 
{adv.} | 
{arkeol.} | 
{avd.} | 
#{bakt.} | 
{bea.} |
{bear.} |
{bearj.} |
#{bm.} | 
{bot.} | 
{čál.} | 
{dis.} |
#{disp.} | 
#{do.} | 
{duo.} |
{duor.} |
#{e.e} | # wrong, TA
#{ee.} | # wrong
#{eg.} | # wrong
#{eng.} | 
{eŋg.} |
#{est.} | 
{etc.} | 
#{ev.luth.} | 
{fem.} | 
{ff.} |
{fi.} | 
{fil.} | 
#{fl.} | 
#{fon.} | 
{fuom.} | 
#{fut.} | 
{gas.} |
{gask.} |
{gen.} | 
{gram.} | 
#{hell.} | 
{hist.} | 
#{hoh.} | 
{holl.} | 
{háldd.} | 
{ibid.} | 
{id.} | 
{imp.} | 
#{imperat.} | 
#{imperf.} | 
#{inc.} | 
{iness.} | 
{inf.} | 
#{innb.} | 
{instr.} | 
{int.} | 
{interj.} | 
{intr.} | 
{intrans.} | 
{inv.} | 
{jna.} | 
{jnv.} | 
{joatkkask.} | 
{jr.} | 
{kem.} |
{kgl.res.} | 
{km.} | 
{lat.} | 
#{lg.} | 
#{lim.} | 
#{ln.} | 
{lst.} | 
{lstr.} | 
#{ltd.} | 
{láv.} |
{lávv.} | 
{m.Kr.} | 
{ma.} | 
{mat.} | 
{med.} | 
#{medl.} | 
{miel.} |
#{mil.} | 
{mill.} | 
{milj.} | 
{mrd.} | 
{min.} | 
#{mod.} | 
{mus.} | 
{mán.} |
{mánn.} |
{ndf.} | 
{ned.} | 
{nl.} | 
{nn.} | 
#{nord.} | 
{nuor.} |
{o.Kr.} | 
{oKr.} | 
{oss.} | 
{p.s.} | 
{pass.} | 
{pol.} | 
#{port.} | 
{pres.} | 
{pret.} | 
{pst.} | 
#{q.e.d.} | 
{refl.} | 
{reg.} | 
{repr.} | 
{rom.} | 
{ru.} | 
{sek.} | 
{sen.} | 
{sg.} | 
{sing.} |
{sotn.} |
{ss.} | 
#{strl.} | 
#{tyrk.} | 
{ung.} | 
{v.v.} | 
{vuoss.} |
{vgs.} ] ;

# <-c,+n>
# Abbreviations that are only CLB 
# if followed by a number.
# Mainly titles.

define INTRANSNUMABBR [
{1.aman.} | 
{Avd.dir.} |
{Bj.} | 
{Chr.} | 
{Dep.} | 
{Edv.} | 
#{Fr.} |
{H.K.H.} | 
{H.M.} | 
{Johs.} |
{Kr.} | 
#{Mme.} | 
{Mr.} | 
{Mrs.} | 
{Ms.} | 
{Seb.} | 
#{Siv.ark.} |
{aman.} | 
{ass.} | 
{avd.dir.} | 
#{avs.} | 
#{bet.} | 
{cand.mag.} | 
#{cand.oecon.} | 
{cand.philol.} | 
{cand.real.} | 
{cand.scient.} | 
{cand.theol.} | 
{cand.} | 
{dipl.ing.} | 
{dipl.ins.} | 
{dir.} | 
{dr.art.} | 
{dr.med.} | 
#{dr.philos.} | 
{dr.theol.} | 
{dr.} | 
{eksp.hoavd.} |
#{eksp.sj.} | 
{fil.tri.} |
#{fenr.} | 
#{gen.lt.} | 
#{genr.lt.} | 
#{genr.} | 
#{h.r.adv.} | 
{hr.} | 
#{kapt.} | 
#{korp.} | 
#{korpr.} | 
#{lekt.} | 
#{lic.} | 
#{ltn.} | 
#{mag.art.} | 
#{mag.} | 
#{merc.} | 
#{mr.} | 
#{mrs.} | 
{ms.} | 
#{obl.} | 
#{oblt.} | 
#{odont.} | 
{oecon.} | 
#{ordf.} | 
{ossod.dir.} |
#{overs.} | 
{pharm.} | 
{philol.} | 
{polit.} | 
{prof.} | 
#{psychol.} | 
{psyk.} | 
{res.kap.} | 
#{scient.} | 
#{sekr.} | 
#{sgt.} | 
#{siv.ark.} | 
#{sjt.} | 
#{stud.phil.} | 
#{stud.polit.} | 
#{stud.psych.} | 
#{stud.real.} | 
#{stud.scient.} | 
#{stud.theol.} | 
#{stud.} | 
#{u.dir.} | 
{varaordf.} | 
{vit.ass.} ] ;


# <+c,-n>
# Abbreviations that only end a sentence
# if followed by an upper case letter,
# not when followed by a number.

define INTRANSCAPABBR [
#{1.Sam.} | 
#{2.Sam.} | 
{Akersgt.} | 
#{B.innst.S.nr.} | 
#{Besl.L.nr.} | 
#{Besl.O.nr.} | 
{Co.} | 
{Dan.} | 
{Dok.nr.} | 
#{Dronningensgt.} | 
#{Eidsvollsgt.} | 
#{Elvegt.} | 
#{Hausmannsgt.} | 
#{Holmegt.} | 
#{Industrigt.} | 
{Innst.O.nr.} | 
{Innst.S.nr.} | 
#{Johansgt.} | 
{Josv.} | 
#{Kirkegt.} | 
{Kong.} | 
#{Kongensgt.} | 
#{Langgt.} | 
{Mark.} | 
{Matt.} | 
{Mos.} | 
#{Musegt.} | 
{Nkr.} | 
{Ob.} | 
#{Olavsgt.} | 
#{Oslov.} | 
{Ot.prp.nr.} | 
{Paul.} | 
#{Prinsensgt.} | 
{Rom.} | 
#{Rådhusgt.} | 
{Sam.} | 
#{Slottsgt.} | 
{St.dieđ.nr.} | 
{St.meld.nr.} | 
{St.prp.nr.} | 
{Storgt.} | 
{Tel.} | 
{Tlf.nr.} | 
{Tlf.} | 
#{Strandgt.} | 
#{Trondheimsv.} | 
#{a.D.} | 
{bd.} |
{borg.} |
{čakč.} |
{cuoŋ.} |
#{da.} | 
{geas.} |
{gnr.} | 
{golg.} |
{guov.} |
#{it.} | 
{jnr.} | 
{juov.} |
{kap.} | 
{kr.} | 
{lnr.} | 
{ltd.} | 
{ltr.} | 
{maks.} | 
{mar.} | 
{mask.} | 
{mies.} |
{mobiltlf.} | 
{njuk.} |
{no.} | 
{nr.} | 
{ođđj.} |
{p.b.} | 
{pb.} | 
{pgf.} | 
{ru.} |
{rv.} | 
{skáb.} |
{suoi.} ] ;


# dot% noStb.db
# <-c,-n>
# Abbreviations that never end a sentence, also if followed
# by a capital letter or a number.

define TRANSABBR [
{Alm.} |
{Dom.} |
#{Eftf.} | 
#{Eks.} | 
#{Esek.} | 
#{I.% N.% D.} | 
#{I.N.D.} | 
#{IL.} | 
{Innst.} | 
#{Jer.} | 
#{Joh.} | 
#{Lim.} | 
#{Lt.} | 
#{Neh.} | 
{Od.} | 
#{Ot.% prp.} | 
{Ot.prp.} | 
{Ot.} | 
#{Pf.} | 
#{Pfg.} | 
{Rt.} | 
{St.% meld.} | 
{St.% prp.} | 
{St.meld.} | 
{Sd.% dieđ.} |
{Sd.% prp.} |
{Sd.} | 
{St.} | 
#{Su.} | 
#{Sv.} | 
{Univ.} | 
{adr.} | 
#{afr.} | 
{akk.} | 
#{alb.} | 
{alm.} | 
{am.} | 
{amer.} | 
#{arab.} | 
{art.} | 
#{atomnr.} | 
#{att.} | 
#{austr.} | 
{aut.} | 
#{balt.} | 
#{bet.} | 
{bib.} | 
{biol.} | 
{bnr.} | 
#{brig.} | 
#{brit.-am.} | 
#{brit.-austr.} | 
{brit.} | 
{bto.} | 
{buo.} | 
{čuo.} |
{ca.} | 
{cea.} |
{cf.} | 
{cit.} |
{co.} | 
#{d.% c.} | 
#{d.% d.} | 
#{d.% m.} | 
#{d.c.} | 
#{d.d.} | 
#{d.m.} | 
#{d.s.} | 
{dat.} | 
{dbm.} |
{dea.} | 
{dept.} | 
{dieđ.} | 
{dim.} | 
{div.} | 
{dkr.} | 
{dok.} |
{ds.} | 
{e.g.} | 
{ea.} |
{ea.ea.} |
{ee.} | 
{eftf.} | 
#{eks.} | 
#{ekskl.} | 
#{et.} | 
{etc.} | 
{eur.} | 
{ev.} | 
#{evt.} | 
#{fec.} | 
{fig.} | 
#{fol.} | 
{fr.} | 
{gč.} | 
#{geom.} | 
{ggl.} |
{gr.} | 
#{hebr.} | 
#{i.% e.} | 
{i.e.} | 
{ib.} | 
{ibid.} |
{ill.} | 
{ind.} | 
#{indoeur.} | 
{inkl.} | 
#{j.d.} | 
#{j.e.} | 
#{j.s.} | 
#{jap.} | 
{jur.} | 
#{jvf.} | 
{kat.} | 
{kgl.} | 
{kl.} | 
{km/t.} | 
{konj.} | 
{konst.} | 
{kst.} | 
#{kurd.} | 
{kvm.} | 
#{l.c.} | 
#{l.r.d.} | 
#{lab.} | 
#{lat.-gr} | 
#{log.} | 
#{m.v.} | 
{m/sek.} | 
{mek.} | 
#{meld.} | 
#{mil.-tekn.} | 
{mv.} | 
{mva.} | 
{myt.} | 
{nkr.} | 
#{no.-da.} | 
{nom.} | 
{nr.} |
{num.} | 
{o.m.a.} | 
{o.m.d.} | 
{o.m.} | 
{Omd.} | 
{omd.} | 
{obj.} | 
{obs.} | 
{off.} | 
#{op.% cit.} | 
{op.cit.} | 
{op.} | 
#{ot.% prp.} | 
{ot.prp.} | 
#{p.% a.} | 
#{p.% m.} | 
#{p.% p.% m.} | 
#{p.% p.} | 
#{p.% r.} | 
#{p.a.} | 
#{p.m.} | 
#{p.p.m.} | 
#{p.p.} | 
#{p.r.} | 
#{p.t.} | 
#{pa.} | 
#{pf.} | 
{pkt.} | 
{pl.} | 
{pop.} | 
#{pr.pr.} | 
{pr.} | 
{prep.} | 
{pron.} | 
{prp.} | 
{q.s.} | 
{red.} | 
{rek.} | 
{rev.} | 
#{rum.-am.} | 
#{sc.} | 
#{sd.} | 
#{serbokr.} | 
{sign.} | 
#{sovj.} | 
#{spes.} | 
#{sr.} | 
{st.} | 
{subj.} | 
{subst.} | 
#{sveits.} | 
#{tekn.} | 
{tel.} | 
{tlf.nr.} | 
{tlf.} | 
{vol.} | 
{vrd.} | 
{vs.} | 
{v.r.d.} ];

echo >>> collecting all abbreviations into one set
define ABBR [ INTRANSABBR | INTRANSCAPABBR | INTRANSNUMABBR | TRANSABBR ] ;

# =============================================================
echo >>> some multi-words
define MWE [
{Bassi Vuoigŋa} |
{Bassi Vuoiŋŋa} |
{Helse Midt} |
{Helse Nord} |
{Helse Vest} |
{dan dihte} | 
{dan láhkái} |
{dan oktavuođas} |
{danin go} |
{das go} | 
{earret eará} | 
{earret go} | 
{eará go} | 
{ele ge} | 
{feara gii} |
{feara mii} |
{gii nu} |
{guhkit áiggi}
{hui amas} | 
{makkaráš nu} |
{makkár nu} |
{man láhkai} | 
{maŋŋel go} | 
{mii nu} |
{muhtun muddui} |
{máná vuostái} |
{ovdal go}
];

# Marker for multi-words
define M1 "<<" ;
define M2 ">>" ;
define MWE1 [M1 MWE M2] ;

# =============================================================
echo >>> define web addresses
define URL [ [{http://www.}|{www.}] [ Char | SINGLE ]+ ];


# =============================================================
echo >>> define tokens
define Token [ WORD | SYMBOL | ABBR | INIT | 
               NUMOrd | NUMCASED | NUM | MWE1 | URL ];

# some non-functioning versions left for evaluation here.
# define Token [ WORD | SYMBOL | TRANSABBR | INIT | NUM ];
# define Token [ WORD | SYMBOL | ABBR | INIT | NUM | NUMCASED ];
# define Token [ 
# WORD | SYMBOL | TRANSCAPABBR | TRANSNUMABBR | TRANSABBR  | INIT | NUM ];

# ==============================================================
echo >>> finding titles etc.
#define TOK3 [ NL -> NL %. || _ NL ] ; # 2209 out!
# I take two newlines, and insert a period in between them.
# The result I call TOK3, and I insert it in the regex line at the
# bottom of this file.


# =============================================================
# handling the abbreviations

echo >>> newline and period copying after numeral
define TOKNUMOrd [ NUMOrd @-> ... NL %. NL || _ WS+ Capital ] ;

echo >>> newline and period copying before abbr. that are intr. wrt. capitals
define TOKINTRANSCAP [ INTRANSCAPABBR @-> ... NL %. NL || _ WS+ Capital ] ; 

echo >>> newline and period copying before abbr. that are intr. wrt. numbers
define TOKINTRANSNUM [ INTRANSNUMABBR @-> ... NL %. NL || _ WS+ NUMROM ] ; 

echo >>> newline and period copying before abbr. that are always intr.
define TOKINTRANS [ INTRANSABBR @-> ... NL %. NL || _ WS+ CAPNUM ] ; 

#echo >>> fix the 34.
#define TOK01 [ NUMOrd @-> ... NL ] ;

echo >>> longest match--insert a newline after each token

# multi start
define Bound [ SINGLE | WS | .#. ]; 
define TOK1 [ MWE @-> M1 ... M2 || Bound _ Bound .o.
              Token @-> ... NL .o. [M1|M2] -> 0 ] ;

echo >>> normalize space
define TOK0 [ WS+ @-> SP ];

echo >>> remove spaces
define WŠ [WS]+ & $[NL] ;



#define TOK1 [ Token @-> ... NL ] ;

echo >>> map spaces to a newline
define TOK2 [ [WŠ]+ @-> NL ];

#define TOK2 [ [WS]+ @-> NL ];
# multi end, remove this note when working

# =============================================================
echo >>> compose the above
read regex [TOKNUMOrd .o. TOKINTRANSCAP .o. TOKINTRANSNUM .o. TOKINTRANS 
#            .o. TOK1 .o. TOK2 .o. TOK3 ]; # 2209 out!
#             .o. TOK1 .o. TOK2 ]; # this one works!
             .o. TOK0 .o. TOK1 .o. TOK2 ];
invert net