xapian-fu 1.7.0 → 1.8.0

data/lib/xapian_fu/stopwords/italian.txt

@@ -1,295 +0,0 @@
-
- | A French stop word list. Comments begin with vertical bar. Each stop
- | word is at the start of a line.
-
-ad             |  a (to) before vowel
-al             |  a + il
-allo           |  a + lo
-ai             |  a + i
-agli           |  a + gli
-all            |  a + l'
-agl            |  a + gl'
-alla           |  a + la
-alle           |  a + le
-con            |  with
-col            |  con + il
-coi            |  con + i (forms collo, cogli etc are now very rare)
-da             |  from
-dal            |  da + il
-dallo          |  da + lo
-dai            |  da + i
-dagli          |  da + gli
-dall           |  da + l'
-dagl           |  da + gll'
-dalla          |  da + la
-dalle          |  da + le
-di             |  of
-del            |  di + il
-dello          |  di + lo
-dei            |  di + i
-degli          |  di + gli
-dell           |  di + l'
-degl           |  di + gl'
-della          |  di + la
-delle          |  di + le
-in             |  in
-nel            |  in + el
-nello          |  in + lo
-nei            |  in + i
-negli          |  in + gli
-nell           |  in + l'
-negl           |  in + gl'
-nella          |  in + la
-nelle          |  in + le
-su             |  on
-sul            |  su + il
-sullo          |  su + lo
-sui            |  su + i
-sugli          |  su + gli
-sull           |  su + l'
-sugl           |  su + gl'
-sulla          |  su + la
-sulle          |  su + le
-per            |  through, by
-tra            |  among
-contro         |  against
-io             |  I
-tu             |  thou
-lui            |  he
-lei            |  she
-noi            |  we
-voi            |  you
-loro           |  they
-mio            |  my
-mia            |
-miei           |
-mie            |
-tuo            |
-tua            |
-tuoi           |  thy
-tue            |
-suo            |
-sua            |
-suoi           |  his, her
-sue            |
-nostro         |  our
-nostra         |
-nostri         |
-nostre         |
-vostro         |  your
-vostra         |
-vostri         |
-vostre         |
-mi             |  me
-ti             |  thee
-ci             |  us, there
-vi             |  you, there
-lo             |  him, the
-la             |  her, the
-li             |  them
-le             |  them, the
-gli            |  to him, the
-ne             |  from there etc
-il             |  the
-un             |  a
-uno            |  a
-una            |  a
-ma             |  but
-ed             |  and
-se             |  if
-perché         |  why, because
-anche          |  also
-come           |  how
-dov            |  where (as dov')
-dove           |  where
-che            |  who, that
-chi            |  who
-cui            |  whom
-non            |  not
-più            |  more
-quale          |  who, that
-quanto         |  how much
-quanti         |
-quanta         |
-quante         |
-quello         |  that
-quelli         |
-quella         |
-quelle         |
-questo         |  this
-questi         |
-questa         |
-queste         |
-si             |  yes
-tutto          |  all
-tutti          |  all
-
-               |  single letter forms:
-
-a              |  at
-c              |  as c' for ce or ci
-e              |  and
-i              |  the
-l              |  as l'
-o              |  or
-
-               | forms of avere, to have (not including the infinitive):
-
-ho
-hai
-ha
-abbiamo
-avete
-hanno
-abbia
-abbiate
-abbiano
-avrò
-avrai
-avrà
-avremo
-avrete
-avranno
-avrei
-avresti
-avrebbe
-avremmo
-avreste
-avrebbero
-avevo
-avevi
-aveva
-avevamo
-avevate
-avevano
-ebbi
-avesti
-ebbe
-avemmo
-aveste
-ebbero
-avessi
-avesse
-avessimo
-avessero
-avendo
-avuto
-avuta
-avuti
-avute
-
-               | forms of essere, to be (not including the infinitive):
-sono
-sei
-è
-siamo
-siete
-sia
-siate
-siano
-sarò
-sarai
-sarà
-saremo
-sarete
-saranno
-sarei
-saresti
-sarebbe
-saremmo
-sareste
-sarebbero
-ero
-eri
-era
-eravamo
-eravate
-erano
-fui
-fosti
-fu
-fummo
-foste
-furono
-fossi
-fosse
-fossimo
-fossero
-essendo
-
-               | forms of fare, to do (not including the infinitive, fa, fat-):
-faccio
-fai
-facciamo
-fanno
-faccia
-facciate
-facciano
-farò
-farai
-farà
-faremo
-farete
-faranno
-farei
-faresti
-farebbe
-faremmo
-fareste
-farebbero
-facevo
-facevi
-faceva
-facevamo
-facevate
-facevano
-feci
-facesti
-fece
-facemmo
-faceste
-fecero
-facessi
-facesse
-facessimo
-facessero
-facendo
-
-               | forms of stare, to be (not including the infinitive):
-sto
-stai
-sta
-stiamo
-stanno
-stia
-stiate
-stiano
-starò
-starai
-starà
-staremo
-starete
-staranno
-starei
-staresti
-starebbe
-staremmo
-stareste
-starebbero
-stavo
-stavi
-stava
-stavamo
-stavate
-stavano
-stetti
-stesti
-stette
-stemmo
-steste
-stettero
-stessi
-stesse
-stessimo
-stessero
-stando

data/lib/xapian_fu/stopwords/norwegian.txt

@@ -1,186 +0,0 @@
-
- | A Norwegian stop word list. Comments begin with vertical bar. Each stop
- | word is at the start of a line.
-
- | This stop word list is for the dominant bokmål dialect. Words unique
- | to nynorsk are marked *.
-
- | Revised by Jan Bruusgaard <Jan.Bruusgaard@ssb.no>, Jan 2005
-
-og             | and
-i              | in
-jeg            | I
-det            | it/this/that
-at             | to (w. inf.)
-en             | a/an
-et             | a/an
-den            | it/this/that
-til            | to
-er             | is/am/are
-som            | who/that
-på             | on
-de             | they / you(formal)
-med            | with
-han            | he
-av             | of
-ikke           | not
-ikkje          | not *
-der            | there
-så             | so
-var            | was/were
-meg            | me
-seg            | you
-men            | but
-ett            | one
-har            | have
-om             | about
-vi             | we
-min            | my
-mitt           | my
-ha             | have
-hadde          | had
-hun            | she
-nå             | now
-over           | over
-da             | when/as
-ved            | by/know
-fra            | from
-du             | you
-ut             | out
-sin            | your
-dem            | them
-oss            | us
-opp            | up
-man            | you/one
-kan            | can
-hans           | his
-hvor           | where
-eller          | or
-hva            | what
-skal           | shall/must
-selv           | self (reflective)
-sjøl           | self (reflective)
-her            | here
-alle           | all
-vil            | will
-bli            | become
-ble            | became
-blei           | became *
-blitt          | have become
-kunne          | could
-inn            | in
-når            | when
-være           | be
-kom            | come
-noen           | some
-noe            | some
-ville          | would
-dere           | you
-som            | who/which/that
-deres          | their/theirs
-kun            | only/just
-ja             | yes
-etter          | after
-ned            | down
-skulle         | should
-denne          | this
-for            | for/because
-deg            | you
-si             | hers/his
-sine           | hers/his
-sitt           | hers/his
-mot            | against
-å              | to
-meget          | much
-hvorfor        | why
-dette          | this
-disse          | these/those
-uten           | without
-hvordan        | how
-ingen          | none
-din            | your
-ditt           | your
-blir           | become
-samme          | same
-hvilken        | which
-hvilke         | which (plural)
-sånn           | such a
-inni           | inside/within
-mellom         | between
-vår            | our
-hver           | each
-hvem           | who
-vors           | us/ours
-hvis           | whose
-både           | both
-bare           | only/just
-enn            | than
-fordi          | as/because
-før            | before
-mange          | many
-også           | also
-slik           | just
-vært           | been
-være           | to be
-båe            | both *
-begge          | both
-siden          | since
-dykk           | your *
-dykkar         | yours *
-dei            | they *
-deira          | them *
-deires         | theirs *
-deim           | them *
-di             | your (fem.) *
-då             | as/when *
-eg             | I *
-ein            | a/an *
-eit            | a/an *
-eitt           | a/an *
-elles          | or *
-honom          | he *
-hjå            | at *
-ho             | she *
-hoe            | she *
-henne          | her
-hennar         | her/hers
-hennes         | hers
-hoss           | how *
-hossen         | how *
-ikkje          | not *
-ingi           | noone *
-inkje          | noone *
-korleis        | how *
-korso          | how *
-kva            | what/which *
-kvar           | where *
-kvarhelst      | where *
-kven           | who/whom *
-kvi            | why *
-kvifor         | why *
-me             | we *
-medan          | while *
-mi             | my *
-mine           | my *
-mykje          | much *
-no             | now *
-nokon          | some (masc./neut.) *
-noka           | some (fem.) *
-nokor          | some *
-noko           | some *
-nokre          | some *
-si             | his/hers *
-sia            | since *
-sidan          | since *
-so             | so *
-somt           | some *
-somme          | some *
-um             | about*
-upp            | up *
-vere           | be *
-vore           | was *
-verte          | become *
-vort           | become *
-varte          | became *
-vart           | became *
-

data/lib/xapian_fu/stopwords/portuguese.txt

@@ -1,245 +0,0 @@
-
- | A Portuguese stop word list. Comments begin with vertical bar. Each stop
- | word is at the start of a line.
-
-
- | The following is a ranked list (commonest to rarest) of stopwords
- | deriving from a large sample of text.
-
- | Extra words have been added at the end.
-
-de             |  of, from
-a              |  the; to, at; her
-o              |  the; him
-que            |  who, that
-e              |  and
-do             |  de + o
-da             |  de + a
-em             |  in
-um             |  a
-para           |  for
-  | é          from SER
-com            |  with
-não            |  not, no
-uma            |  a
-os             |  the; them
-no             |  em + o
-se             |  himself etc
-na             |  em + a
-por            |  for
-mais           |  more
-as             |  the; them
-dos            |  de + os
-como           |  as, like
-mas            |  but
-  | foi        from SER
-ao             |  a + o
-ele            |  he
-das            |  de + as
-  | tem        from TER
-à              |  a + a
-seu            |  his
-sua            |  her
-ou             |  or
-  | ser        from SER
-quando         |  when
-muito          |  much
-  | há         from HAV
-nos            |  em + os; us
-já             |  already, now
-  | está       from EST
-eu             |  I
-também         |  also
-só             |  only, just
-pelo           |  per + o
-pela           |  per + a
-até            |  up to
-isso           |  that
-ela            |  he
-entre          |  between
-  | era        from SER
-depois         |  after
-sem            |  without
-mesmo          |  same
-aos            |  a + os
-  | ter        from TER
-seus           |  his
-quem           |  whom
-nas            |  em + as
-me             |  me
-esse           |  that
-eles           |  they
-  | estão      from EST
-você           |  you
-  | tinha      from TER
-  | foram      from SER
-essa           |  that
-num            |  em + um
-nem            |  nor
-suas           |  her
-meu            |  my
-às             |  a + as
-minha          |  my
-  | têm        from TER
-numa           |  em + uma
-pelos          |  per + os
-elas           |  they
-  | havia      from HAV
-  | seja       from SER
-qual           |  which
-  | será       from SER
-nós            |  we
-  | tenho      from TER
-lhe            |  to him, her
-deles          |  of them
-essas          |  those
-esses          |  those
-pelas          |  per + as
-este           |  this
-  | fosse      from SER
-dele           |  of him
-
- | other words. There are many contractions such as naquele = em+aquele,
- | mo = me+o, but they are rare.
- | Indefinite article plural forms are also rare.
-
-tu             |  thou
-te             |  thee
-vocês          |  you (plural)
-vos            |  you
-lhes           |  to them
-meus           |  my
-minhas
-teu            |  thy
-tua
-teus
-tuas
-nosso          | our
-nossa
-nossos
-nossas
-
-dela           |  of her
-delas          |  of them
-
-esta           |  this
-estes          |  these
-estas          |  these
-aquele         |  that
-aquela         |  that
-aqueles        |  those
-aquelas        |  those
-isto           |  this
-aquilo         |  that
-
-               | forms of estar, to be (not including the infinitive):
-estou
-está
-estamos
-estão
-estive
-esteve
-estivemos
-estiveram
-estava
-estávamos
-estavam
-estivera
-estivéramos
-esteja
-estejamos
-estejam
-estivesse
-estivéssemos
-estivessem
-estiver
-estivermos
-estiverem
-
-               | forms of haver, to have (not including the infinitive):
-hei
-há
-havemos
-hão
-houve
-houvemos
-houveram
-houvera
-houvéramos
-haja
-hajamos
-hajam
-houvesse
-houvéssemos
-houvessem
-houver
-houvermos
-houverem
-houverei
-houverá
-houveremos
-houverão
-houveria
-houveríamos
-houveriam
-
-               | forms of ser, to be (not including the infinitive):
-sou
-somos
-são
-era
-éramos
-eram
-fui
-foi
-fomos
-foram
-fora
-fôramos
-seja
-sejamos
-sejam
-fosse
-fôssemos
-fossem
-for
-formos
-forem
-serei
-será
-seremos
-serão
-seria
-seríamos
-seriam
-
-               | forms of ter, to have (not including the infinitive):
-tenho
-tem
-temos
-tém
-tinha
-tínhamos
-tinham
-tive
-teve
-tivemos
-tiveram
-tivera
-tivéramos
-tenha
-tenhamos
-tenham
-tivesse
-tivéssemos
-tivessem
-tiver
-tivermos
-tiverem
-terei
-terá
-teremos
-terão
-teria
-teríamos
-teriam