Re: TSearch2: Problems with compound words and stop words

On Wed, 17 Nov 2004, Timo Haberkern wrote:

> sorry for the late answer, i was on holyday,
>
> see my remarks below
>
>
> Oleg Bartunov wrote:
>
>> On Fri, 5 Nov 2004, Timo Haberkern wrote:
>>
>>> Oleg,
>>>
>>> i use TSearch2 with PostgreSQL 7.4.6 and i applied the compoundword patch
>>> yesterday. The configuration changed a little bit but the result is the
>>> same. I get no compound words. I'm using the locale de_DE with encoding
>>> ISO8859-1 for the database.
>>>
>>> I think i spell is working correctly except the compound words. If i try
>>>
>>> SELECT lexize('de_ispell', 'springt')
>>>
>>> i get
>>>
>>> lexize
>>> {springen,springen}
>>>
>>> which seems correct.
>>>
>>>
>>> But a SELECT lexize('de_ispell', 'Autobahn')
>>>
>>> results in
>>>
>>> lexize
>>> {autobahn}
>>>
>>> i would expect {auto,bahn, autobahn}
>>
>>
>> Hmm, have you checked 'Autobahn' in ispell dictionary ? Does dictionary you
>> used supports 'Z' flag for compound words ?
>
> Autobahn is in the ispell dictionary. What does a ispell dictionary need to
> support the Z flag???
>

Try ispell -C Autobahn
search 'compound' in 'man ispell' for details.
the problem exists only if ispell *does* splits word correctly while tsearch2
doesn't. You should find correct ispell dictionary for german or create it
yourself. You may consult monzilla.net
http://staff.science.uva.nl/~christof/monzilla/research/project-dr.html

>
> Timo
>
>
>
>
>
>>
>>
>>>
>>> The new configuration after the compound word patch:
>>>
>>
>> Seems you overestimate my capabilities :)
>>
>>
>>>
>>> Actions dict_name
>>> <http://www.rotex-service.com/phppgadmin/display.php?database=selina_rotex&schema=public&table=pg_ts_dict&return_url=tblproperties.php%3Fdatabase%3Dselina_rotex%26amp%3Bschema%3Dpublic%26table%3Dpg_ts_dict&return_desc=Back&sortkey=2&sortdir=asc&strings=expanded&page=1>
>>> dict_init
>>> <http://www.rotex-service.com/phppgadmin/display.php?database=selina_rotex&schema=public&table=pg_ts_dict&return_url=tblproperties.php%3Fdatabase%3Dselina_rotex%26amp%3Bschema%3Dpublic%26table%3Dpg_ts_dict&return_desc=Back&sortkey=3&sortdir=asc&strings=expanded&page=1>
>>> dict_initoption
>>> <http://www.rotex-service.com/phppgadmin/display.php?database=selina_rotex&schema=public&table=pg_ts_dict&return_url=tblproperties.php%3Fdatabase%3Dselina_rotex%26amp%3Bschema%3Dpublic%26table%3Dpg_ts_dict&return_desc=Back&sortkey=4&sortdir=asc&strings=expanded&page=1>
>>> dict_lexize
>>> <http://www.rotex-service.com/phppgadmin/display.php?database=selina_rotex&schema=public&table=pg_ts_dict&return_url=tblproperties.php%3Fdatabase%3Dselina_rotex%26amp%3Bschema%3Dpublic%26table%3Dpg_ts_dict&return_desc=Back&sortkey=5&sortdir=asc&strings=expanded&page=1>
>>> dict_comment
>>> <http://www.rotex-service.com/phppgadmin/display.php?database=selina_rotex&schema=public&table=pg_ts_dict&return_url=tblproperties.php%3Fdatabase%3Dselina_rotex%26amp%3Bschema%3Dpublic%26table%3Dpg_ts_dict&return_desc=Back&sortkey=6&sortdir=asc&strings=expanded&page=1>
>>> Edit
>>> <http://www.rotex-service.com/phppgadmin/display.php?action=confeditrow&strings=expanded&page=1&key%5Bdict_name%5D=simple&database=selina_rotex&schema=public&table=pg_ts_dict&return_url=tblproperties.php%3Fdatabase%3Dselina_rotex%26amp%3Bschema%3Dpublic%26table%3Dpg_ts_dict&return_desc=Back&sortkey=&sortdir=>
>>> Delete
>>> <http://www.rotex-service.com/phppgadmin/display.php?action=confdelrow&strings=expanded&page=1&key%5Bdict_name%5D=simple&database=selina_rotex&schema=public&table=pg_ts_dict&return_url=tblproperties.php%3Fdatabase%3Dselina_rotex%26amp%3Bschema%3Dpublic%26table%3Dpg_ts_dict&return_desc=Back&sortkey=&sortdir=>
>>> simple dex_init(text) /NULL/
>>> dex_lexize(internal,internal,integer) Simple example of dictionary.
>>> Edit
>>> <http://www.rotex-service.com/phppgadmin/display.php?action=confeditrow&strings=expanded&page=1&key%5Bdict_name%5D=en_stem&database=selina_rotex&schema=public&table=pg_ts_dict&return_url=tblproperties.php%3Fdatabase%3Dselina_rotex%26amp%3Bschema%3Dpublic%26table%3Dpg_ts_dict&return_desc=Back&sortkey=&sortdir=>
>>> Delete
>>> <http://www.rotex-service.com/phppgadmin/display.php?action=confdelrow&strings=expanded&page=1&key%5Bdict_name%5D=en_stem&database=selina_rotex&schema=public&table=pg_ts_dict&return_url=tblproperties.php%3Fdatabase%3Dselina_rotex%26amp%3Bschema%3Dpublic%26table%3Dpg_ts_dict&return_desc=Back&sortkey=&sortdir=>
>>> en_stem snb_en_init(text) /usr/local/pgsql/share/contrib/english.stop
>>> snb_lexize(internal,internal,integer) English Stemmer. Snowball.
>>> Edit
>>> <http://www.rotex-service.com/phppgadmin/display.php?action=confeditrow&strings=expanded&page=1&key%5Bdict_name%5D=ru_stem&database=selina_rotex&schema=public&table=pg_ts_dict&return_url=tblproperties.php%3Fdatabase%3Dselina_rotex%26amp%3Bschema%3Dpublic%26table%3Dpg_ts_dict&return_desc=Back&sortkey=&sortdir=>
>>> Delete
>>> <http://www.rotex-service.com/phppgadmin/display.php?action=confdelrow&strings=expanded&page=1&key%5Bdict_name%5D=ru_stem&database=selina_rotex&schema=public&table=pg_ts_dict&return_url=tblproperties.php%3Fdatabase%3Dselina_rotex%26amp%3Bschema%3Dpublic%26table%3Dpg_ts_dict&return_desc=Back&sortkey=&sortdir=>
>>> ru_stem snb_ru_init(text) /usr/local/pgsql/share/contrib/russian.stop
>>> snb_lexize(internal,internal,integer) Russian Stemmer. Snowball.
>>> Edit
>>> <http://www.rotex-service.com/phppgadmin/display.php?action=confeditrow&strings=expanded&page=1&key%5Bdict_name%5D=ispell_template&database=selina_rotex&schema=public&table=pg_ts_dict&return_url=tblproperties.php%3Fdatabase%3Dselina_rotex%26amp%3Bschema%3Dpublic%26table%3Dpg_ts_dict&return_desc=Back&sortkey=&sortdir=>
>>> Delete
>>> <http://www.rotex-service.com/phppgadmin/display.php?action=confdelrow&strings=expanded&page=1&key%5Bdict_name%5D=ispell_template&database=selina_rotex&schema=public&table=pg_ts_dict&return_url=tblproperties.php%3Fdatabase%3Dselina_rotex%26amp%3Bschema%3Dpublic%26table%3Dpg_ts_dict&return_desc=Back&sortkey=&sortdir=>
>>> ispell_template spell_init(text) /NULL/
>>> spell_lexize(internal,internal,integer) ISpell interface. Must have
>>> .dict and .aff files
>>> Edit
>>> <http://www.rotex-service.com/phppgadmin/display.php?action=confeditrow&strings=expanded&page=1&key%5Bdict_name%5D=synonym&database=selina_rotex&schema=public&table=pg_ts_dict&return_url=tblproperties.php%3Fdatabase%3Dselina_rotex%26amp%3Bschema%3Dpublic%26table%3Dpg_ts_dict&return_desc=Back&sortkey=&sortdir=>
>>> Delete
>>> <http://www.rotex-service.com/phppgadmin/display.php?action=confdelrow&strings=expanded&page=1&key%5Bdict_name%5D=synonym&database=selina_rotex&schema=public&table=pg_ts_dict&return_url=tblproperties.php%3Fdatabase%3Dselina_rotex%26amp%3Bschema%3Dpublic%26table%3Dpg_ts_dict&return_desc=Back&sortkey=&sortdir=>
>>> synonym syn_init(text) /NULL/
>>> syn_lexize(internal,internal,integer) Example of synonym dictionary
>>> Edit
>>> <http://www.rotex-service.com/phppgadmin/display.php?action=confeditrow&strings=expanded&page=1&key%5Bdict_name%5D=de_ispell&database=selina_rotex&schema=public&table=pg_ts_dict&return_url=tblproperties.php%3Fdatabase%3Dselina_rotex%26amp%3Bschema%3Dpublic%26table%3Dpg_ts_dict&return_desc=Back&sortkey=&sortdir=>
>>> Delete
>>> <http://www.rotex-service.com/phppgadmin/display.php?action=confdelrow&strings=expanded&page=1&key%5Bdict_name%5D=de_ispell&database=selina_rotex&schema=public&table=pg_ts_dict&return_url=tblproperties.php%3Fdatabase%3Dselina_rotex%26amp%3Bschema%3Dpublic%26table%3Dpg_ts_dict&return_desc=Back&sortkey=&sortdir=>
>>> de_ispell spell_init(text)
>>> DictFile="/usr/local/pgsql/share/contrib/dictonary/german_comb.dict",
>>> AffFile="/usr/local/pgsql/share/contrib/dictonary/german_comb.aff",
>>> StopFile="/usr/local/pgsql/share/contrib/dictonary/german.stop"
>>> spell_lexize(internal,internal,integer) /NULL/
>>>
>>>
>>>
>>> Timo
>>>
>>>
>>> Oleg Bartunov wrote:
>>>
>>>> Timo,
>>>>
>>>> please, check you apply patch for compound word support.
>>>> What is version of postgresql ?
>>>> Does ispell dict works for non-compound words ?
>>>>
>>>> Oleg
>>>>
>>>> On Fri, 5 Nov 2004, Timo Haberkern wrote:
>>>>
>>>>> Hi there,
>>>>>
>>>>> i have some troubles with my TSearch2 Installation. I have done this
>>>>> installation as described in
>>>>> http://www.sai.msu.su/~megera/oddmuse/index.cgi/Tsearch_V2_compound_words
>>>>> <http://www.sai.msu.su/%7Emegera/oddmuse/index.cgi/Tsearch_V2_compound_words>
>>>>> I used the german myspell dictionary from
>>>>> http://lingucomponent.openoffice.org/spell_dic.html and converted it
>>>>> with
>>>>> my2ispell
>>>>>
>>>>> Nearly everything is working fine so far, except two problems:
>>>>>
>>>>> 1.) The stopword-file seems to be ignored: If i try it with SELECT
>>>>> to_tsvector("default_german", "ein Haus") i get "ein":1 "haus":2
>>>>>
>>>>> ein should be a Stopword for german (and is defined the german.stop file
>>>>> as
>>>>> well)
>>>>>
>>>>> 2.) The compound words feature doesn"t work too. I have tried a lot of
>>>>> words,
>>>>> i.e. "Fehlermeldung" with SELECT to_tsvector("default_german",
>>>>> "Fehlermeldung")
>>>>> i only get
>>>>> "fehlermeldung":1 but i would expect "fehler" and "meldung" as seperated
>>>>> entries. Is there anything wrong with the dictonary or my configuration?
>>>>>
>>>>>
>>>>> My current configuration:
>>>>>
>>>>> pg_ts_cfg:
>>>>>
>>>>> default default C
>>>>> default_russian default ru_RU.KOI8-R
>>>>> simple default NULL
>>>>> default_german default de_DE.ISO8859-1
>>>>> pg_ts_cfgmap:
>>>>>
>>>>> default_german host {simple}
>>>>> default_german hword {simple}
>>>>> default_german int {simple}
>>>>> default_german nlhword {simple}
>>>>> default_german nlpart_hword {simple}
>>>>> default_german nlword {simple}
>>>>> default_german part_hword {simple}
>>>>> default_german sfloat {simple}
>>>>> default_german uint {simple}
>>>>> default_german uri {simple}
>>>>> default_german url {simple}
>>>>> default_german version {simple}
>>>>> default_german word {simple}
>>>>> default_german lpart_hword {de_ispell,german_snowball}
>>>>> default_german lword {de_ispell,german_snowball}
>>>>> default_german lhword {de_ispell,german_snowball}
>>>>>
>>>>>
>>>>> pg_ts_dict:
>>>>>
>>>>> de_ispell | 17166 |
>>>>> DictFile="/usr/local/pgsql/share/contrib/dictonary/german.dict",
>>>>> AffFile="/usr/local/pgsql/share/contrib/dictonary/german.aff",
>>>>> StopFile="/usr/local/pgsql/share/contrib/dictonary/german.stop" |
>>>>> 17167 | NULL
>>>>> german_snowball | 17357 | NULL | 17162 | Snowball stemmer for
>>>>> german
>>>>>
>>>>>
>>>>>
>>>>> Can anyone help me?
>>>>>
>>>>> regards
>>>>>
>>>>> Timo
>>>>>
>>>>>
>>>>> ---------------------------(end of broadcast)---------------------------
>>>>> TIP 4: Don't 'kill -9' the postmaster
>>>>>
>>>>
>>>> Regards,
>>>> Oleg
>>>> _____________________________________________________________
>>>> Oleg Bartunov, sci.researcher, hostmaster of AstroNet,
>>>> Sternberg Astronomical Institute, Moscow University (Russia)
>>>> Internet: oleg(at)sai(dot)msu(dot)su, http://www.sai.msu.su/~megera/
>>>> phone: +007(095)939-16-83, +007(095)939-23-83
>>>>
>>>> ---------------------------(end of broadcast)---------------------------
>>>> TIP 2: you can get off all lists at once with the unregister command
>>>> (send "unregister YourEmailAddressHere" to majordomo(at)postgresql(dot)org)
>>>>
>>>>
>>>
>>
>> Regards,
>> Oleg
>> _____________________________________________________________
>> Oleg Bartunov, sci.researcher, hostmaster of AstroNet,
>> Sternberg Astronomical Institute, Moscow University (Russia)
>> Internet: oleg(at)sai(dot)msu(dot)su, http://www.sai.msu.su/~megera/
>> phone: +007(095)939-16-83, +007(095)939-23-83
>>
>> ---------------------------(end of broadcast)---------------------------
>> TIP 2: you can get off all lists at once with the unregister command
>> (send "unregister YourEmailAddressHere" to majordomo(at)postgresql(dot)org)
>>
>>
>

Regards,
Oleg
_____________________________________________________________
Oleg Bartunov, sci.researcher, hostmaster of AstroNet,
Sternberg Astronomical Institute, Moscow University (Russia)
Internet: oleg(at)sai(dot)msu(dot)su, http://www.sai.msu.su/~megera/
phone: +007(095)939-16-83, +007(095)939-23-83