rere 0.0.1

data/example/solr/collection1/conf/lang/stopwords_nl.txt

@@ -0,0 +1,117 @@
+ | From svn.tartarus.org/snowball/trunk/website/algorithms/dutch/stop.txt
+ | This file is distributed under the BSD License.
+ | See http://snowball.tartarus.org/license.php
+ | Also see http://www.opensource.org/licenses/bsd-license.html
+ |  - Encoding was converted to UTF-8.
+ |  - This notice was added.
+
+ | A Dutch stop word list. Comments begin with vertical bar. Each stop
+ | word is at the start of a line.
+
+ | This is a ranked list (commonest to rarest) of stopwords derived from
+ | a large sample of Dutch text.
+
+ | Dutch stop words frequently exhibit homonym clashes. These are indicated
+ | clearly below.
+
+de             |  the
+en             |  and
+van            |  of, from
+ik             |  I, the ego
+te             |  (1) chez, at etc, (2) to, (3) too
+dat            |  that, which
+die            |  that, those, who, which
+in             |  in, inside
+een            |  a, an, one
+hij            |  he
+het            |  the, it
+niet           |  not, nothing, naught
+zijn           |  (1) to be, being, (2) his, one's, its
+is             |  is
+was            |  (1) was, past tense of all persons sing. of 'zijn' (to be) (2) wax, (3) the washing, (4) rise of river
+op             |  on, upon, at, in, up, used up
+aan            |  on, upon, to (as dative)
+met            |  with, by
+als            |  like, such as, when
+voor           |  (1) before, in front of, (2) furrow
+had            |  had, past tense all persons sing. of 'hebben' (have)
+er             |  there
+maar           |  but, only
+om             |  round, about, for etc
+hem            |  him
+dan            |  then
+zou            |  should/would, past tense all persons sing. of 'zullen'
+of             |  or, whether, if
+wat            |  what, something, anything
+mijn           |  possessive and noun 'mine'
+men            |  people, 'one'
+dit            |  this
+zo             |  so, thus, in this way
+door           |  through by
+over           |  over, across
+ze             |  she, her, they, them
+zich           |  oneself
+bij            |  (1) a bee, (2) by, near, at
+ook            |  also, too
+tot            |  till, until
+je             |  you
+mij            |  me
+uit            |  out of, from
+der            |  Old Dutch form of 'van der' still found in surnames
+daar           |  (1) there, (2) because
+haar           |  (1) her, their, them, (2) hair
+naar           |  (1) unpleasant, unwell etc, (2) towards, (3) as
+heb            |  present first person sing. of 'to have'
+hoe            |  how, why
+heeft          |  present third person sing. of 'to have'
+hebben         |  'to have' and various parts thereof
+deze           |  this
+u              |  you
+want           |  (1) for, (2) mitten, (3) rigging
+nog            |  yet, still
+zal            |  'shall', first and third person sing. of verb 'zullen' (will)
+me             |  me
+zij            |  she, they
+nu             |  now
+ge             |  'thou', still used in Belgium and south Netherlands
+geen           |  none
+omdat          |  because
+iets           |  something, somewhat
+worden         |  to become, grow, get
+toch           |  yet, still
+al             |  all, every, each
+waren          |  (1) 'were' (2) to wander, (3) wares, (3)
+veel           |  much, many
+meer           |  (1) more, (2) lake
+doen           |  to do, to make
+toen           |  then, when
+moet           |  noun 'spot/mote' and present form of 'to must'
+ben            |  (1) am, (2) 'are' in interrogative second person singular of 'to be'
+zonder         |  without
+kan            |  noun 'can' and present form of 'to be able'
+hun            |  their, them
+dus            |  so, consequently
+alles          |  all, everything, anything
+onder          |  under, beneath
+ja             |  yes, of course
+eens           |  once, one day
+hier           |  here
+wie            |  who
+werd           |  imperfect third person sing. of 'become'
+altijd         |  always
+doch           |  yet, but etc
+wordt          |  present third person sing. of 'become'
+wezen          |  (1) to be, (2) 'been' as in 'been fishing', (3) orphans
+kunnen         |  to be able
+ons            |  us/our
+zelf           |  self
+tegen          |  against, towards, at
+na             |  after, near
+reeds          |  already
+wil            |  (1) present tense of 'want', (2) 'will', noun, (3) fender
+kon            |  could; past tense of 'to be able'
+niets          |  nothing
+uw             |  your
+iemand         |  somebody
+geweest        |  been; past participle of 'be'
+andere         |  other

data/example/solr/collection1/conf/lang/stopwords_no.txt

@@ -0,0 +1,192 @@
+ | From svn.tartarus.org/snowball/trunk/website/algorithms/norwegian/stop.txt
+ | This file is distributed under the BSD License.
+ | See http://snowball.tartarus.org/license.php
+ | Also see http://www.opensource.org/licenses/bsd-license.html
+ |  - Encoding was converted to UTF-8.
+ |  - This notice was added.
+
+ | 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/example/solr/collection1/conf/lang/stopwords_pt.txt

@@ -0,0 +1,251 @@
+ | From svn.tartarus.org/snowball/trunk/website/algorithms/portuguese/stop.txt
+ | This file is distributed under the BSD License.
+ | See http://snowball.tartarus.org/license.php
+ | Also see http://www.opensource.org/licenses/bsd-license.html
+ |  - Encoding was converted to UTF-8.
+ |  - This notice was added.
+
+ | 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