smartmachine 0.6.0 → 0.7.0

data/lib/smart_machine/grids/solr/config/configsets/_default/conf/lang/stopwords_ja.txt

@@ -0,0 +1,127 @@
+#
+# This file defines a stopword set for Japanese.
+#
+# This set is made up of hand-picked frequent terms from segmented Japanese Wikipedia.
+# Punctuation characters and frequent kanji have mostly been left out.  See LUCENE-3745
+# for frequency lists, etc. that can be useful for making your own set (if desired)
+#
+# Note that there is an overlap between these stopwords and the terms stopped when used
+# in combination with the JapanesePartOfSpeechStopFilter.  When editing this file, note
+# that comments are not allowed on the same line as stopwords.
+#
+# Also note that stopping is done in a case-insensitive manner.  Change your StopFilter
+# configuration if you need case-sensitive stopping.  Lastly, note that stopping is done
+# using the same character width as the entries in this file.  Since this StopFilter is
+# normally done after a CJKWidthFilter in your chain, you would usually want your romaji
+# entries to be in half-width and your kana entries to be in full-width.
+#
+の
+に
+は
+を
+た
+が
+で
+て
+と
+し
+れ
+さ
+ある
+いる
+も
+する
+から
+な
+こと
+として
+い
+や
+れる
+など
+なっ
+ない
+この
+ため
+その
+あっ
+よう
+また
+もの
+という
+あり
+まで
+られ
+なる
+へ
+か
+だ
+これ
+によって
+により
+おり
+より
+による
+ず
+なり
+られる
+において
+ば
+なかっ
+なく
+しかし
+について
+せ
+だっ
+その後
+できる
+それ
+う
+ので
+なお
+のみ
+でき
+き
+つ
+における
+および
+いう
+さらに
+でも
+ら
+たり
+その他
+に関する
+たち
+ます
+ん
+なら
+に対して
+特に
+せる
+及び
+これら
+とき
+では
+にて
+ほか
+ながら
+うち
+そして
+とともに
+ただし
+かつて
+それぞれ
+または
+お
+ほど
+ものの
+に対する
+ほとんど
+と共に
+といった
+です
+とも
+ところ
+ここ
+##### End of file

data/lib/smart_machine/grids/solr/config/configsets/_default/conf/lang/stopwords_lv.txt

@@ -0,0 +1,172 @@
+# Set of Latvian stopwords from A Stemming Algorithm for Latvian, Karlis Kreslins
+# the original list of over 800 forms was refined: 
+#   pronouns, adverbs, interjections were removed
+# 
+# prepositions
+aiz
+ap
+ar
+apakš
+ārpus
+augšpus
+bez
+caur
+dēļ
+gar
+iekš
+iz
+kopš
+labad
+lejpus
+līdz
+no
+otrpus
+pa
+par
+pār
+pēc
+pie
+pirms
+pret
+priekš
+starp
+šaipus
+uz
+viņpus
+virs
+virspus
+zem
+apakšpus
+# Conjunctions
+un
+bet
+jo
+ja
+ka
+lai
+tomēr
+tikko
+turpretī
+arī
+kaut
+gan
+tādēļ
+tā
+ne
+tikvien
+vien
+kā
+ir
+te
+vai
+kamēr
+# Particles
+ar
+diezin
+droši
+diemžēl
+nebūt
+ik
+it
+taču
+nu
+pat
+tiklab
+iekšpus
+nedz
+tik
+nevis
+turpretim
+jeb
+iekam
+iekām
+iekāms
+kolīdz
+līdzko
+tiklīdz
+jebšu
+tālab
+tāpēc
+nekā
+itin
+jā
+jau
+jel
+nē
+nezin
+tad
+tikai
+vis
+tak
+iekams
+vien
+# modal verbs
+būt  
+biju 
+biji
+bija
+bijām
+bijāt
+esmu
+esi
+esam
+esat 
+būšu     
+būsi
+būs
+būsim
+būsiet
+tikt
+tiku
+tiki
+tika
+tikām
+tikāt
+tieku
+tiec
+tiek
+tiekam
+tiekat
+tikšu
+tiks
+tiksim
+tiksiet
+tapt
+tapi
+tapāt
+topat
+tapšu
+tapsi
+taps
+tapsim
+tapsiet
+kļūt
+kļuvu
+kļuvi
+kļuva
+kļuvām
+kļuvāt
+kļūstu
+kļūsti
+kļūst
+kļūstam
+kļūstat
+kļūšu
+kļūsi
+kļūs
+kļūsim
+kļūsiet
+# verbs
+varēt
+varēju
+varējām
+varēšu
+varēsim
+var
+varēji
+varējāt
+varēsi
+varēsiet
+varat
+varēja
+varēs

data/lib/smart_machine/grids/solr/config/configsets/_default/conf/lang/stopwords_nl.txt

@@ -0,0 +1,119 @@
+ | 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.
+ |
+ | NOTE: To use this file with StopFilterFactory, you must specify format="snowball"
+
+ | 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/lib/smart_machine/grids/solr/config/configsets/_default/conf/lang/stopwords_no.txt

@@ -0,0 +1,194 @@
+ | 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.
+ |
+ | NOTE: To use this file with StopFilterFactory, you must specify format="snowball"
+
+ | 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 *
+