IRC logs of #tryton for Wednesday, 2011-05-18

chat.freenode.net #tryton log beginning Wed May 18 00:00:01 CEST 2011
-!- trifon_(~trifon@p548C43D1.dip.t-dialin.net) has joined #tryton00:38
-!- gremly(~gremly@200.106.202.91) has joined #tryton03:11
-!- effenber1(~chlu@pD9E9C34B.dip.t-dialin.net) has joined #tryton04:15
-!- yangoon(~mathiasb@p549F3A83.dip.t-dialin.net) has joined #tryton05:18
-!- gremly(~gremly@200.106.202.91) has joined #tryton05:50
-!- haso(~haso@17.79-161-33.customer.lyse.net) has joined #tryton06:12
-!- vladimirek(~vladimire@bband-dyn166.178-41-220.t-com.sk) has joined #tryton06:29
-!- predatell(~predatell@85.198.132.141) has joined #tryton06:55
-!- okko(~okko@dhcp-077-251-140-095.chello.nl) has joined #tryton07:06
-!- okko(~okko@62.58.29.41) has joined #tryton08:04
-!- mhi1(~mhi@pd95b104a.dip0.t-ipconnect.de) has joined #tryton08:18
-!- predatell(~predatell@85.198.132.141) has joined #tryton08:35
-!- pjstevns(~pjstevns@a83-163-46-103.adsl.xs4all.nl) has joined #tryton08:57
-!- predatell(~predatell@85.198.132.141) has joined #tryton09:33
-!- cedk(~ced@gentoo/developer/cedk) has joined #tryton09:41
-!- bechamel(~user@cismwks02-virtual1.cism.ucl.ac.be) has joined #tryton09:49
-!- udono(~udono@ip-95-222-129-145.unitymediagroup.de) has joined #tryton10:06
-!- helmor(~helmor@46.115.4.226) has joined #tryton10:33
-!- paepke(~paepke@gate.paepke.net) has joined #tryton11:14
-!- nicoe(~nicoe@ced.homedns.org) has joined #tryton11:22
-!- trifon(~trifon@p548C3CF9.dip.t-dialin.net) has joined #tryton14:01
-!- enlightx(~enlightx@static-217-133-61-144.clienti.tiscali.it) has joined #tryton15:00
-!- predatell(~predatell@85.198.132.141) has joined #tryton15:10
-!- sharoon(~sharoon@173-9-190-185-miami.txt.hfc.comcastbusiness.net) has joined #tryton15:27
sharoonwhat is pound15:27
sharoonsorry wrong IRC15:27
-!- trifon_(~trifon@p548C7C99.dip.t-dialin.net) has joined #tryton15:31
-!- pjstevns(~pjstevns@a83-163-46-103.adsl.xs4all.nl) has left #tryton15:53
-!- ecarreras(~under@81.184.3.75.static.user.ono.com) has joined #tryton16:11
-!- ecarreras(~under@unaffiliated/ecarreras) has joined #tryton16:11
-!- trifon__(~trifon@p548C5663.dip.t-dialin.net) has joined #tryton16:16
-!- okko(~okko@62.58.29.41) has joined #tryton16:22
-!- cedk_(~ced@ced.homedns.org) has joined #tryton16:36
-!- cedk_(~ced@ced.homedns.org) has joined #tryton16:37
-!- zodman(~andres-va@foresight/developer/zodman) has joined #tryton16:43
-!- okko(~okko@62.58.29.41) has joined #tryton16:53
-!- cedk_(~ced@ced.homedns.org) has joined #tryton16:56
-!- sharoon(~sharoon@173-9-190-185-miami.txt.hfc.comcastbusiness.net) has left #tryton17:06
-!- cedk(~ced@gentoo/developer/cedk) has joined #tryton17:13
-!- bechamel(~user@host-85-201-144-79.brutele.be) has joined #tryton17:36
-!- trifon(~trifon@p548C5663.dip.t-dialin.net) has joined #tryton17:59
-!- FWiesing(~franz@mail.tryton.at) has joined #tryton19:14
-!- gremly(~gremly@200.106.202.91) has joined #tryton19:37
-!- plantian(~ian@c-67-169-72-36.hsd1.ca.comcast.net) has joined #tryton19:42
-!- cedk(~ced@gentoo/developer/cedk) has joined #tryton21:03
dfamoratocedk: ping21:04
cedkdfamorato: pong21:05
dfamoratocedk: Hey there. I have some questions about my GSOC project on full-text search21:05
dfamoratocedk: I've been doing some research on the sphinx search server21:06
dfamoratocedk: and we have mainly 2 ways to index data: Real-Time and Delta(batch) indexing at a constant time21:06
dfamoratocedk: So, since there are major differences on implementation details between these two21:07
dfamoratocedk: I would like your opinion on how the data should be index for tryton... Real-Time or Batch21:08
cedkdfamorato: of course Real-Time is better but what will be the overload of this?21:09
dfamoratocedk: regarding index database ... it says we should expect something from 40% to 70% of the size of the database being indexed21:10
dfamoratocedk; regaring memory, so far I haven't found much information expect the fact that we can declare how much memory can be used by the indexer and the indexer will consume UP TO that limit21:11
dfamoratocedk:  then will start writing on the disk21:11
dfamoratocedk; It is also possible with Sphinx to have multiple indexes....21:12
cedkdfamorato: I think you need to have one index per language21:12
dfamoratocedk: example: one index for products.... 1 for invoices, 1 for party and so on21:12
dfamoratocedk:  and when we want to query the index, they are queried in parallel21:13
cedkdfamorato: yes, but I think this should be configured21:13
cedkdfamorato: how do you see how the update of the index will be done in RT?21:13
dfamoratocedk:  well... last time I spoke with bertrand... we discussed that ideally we would make a tryton "driver"   to be used by sphinx indexer21:14
dfamoratothat means, any changes it happens on tryton should be PUSHED to the indexer in order for it to be indexed21:15
dfamoratothat's the main difference between the batch indexing and RT21:15
dfamoratocedk: batch processing  = we can query for the data we want and then we index it21:16
cedkdfamorato: ok but what will be the cost of the push? Could it be done in separate thread?21:16
dfamoratocedk: RT = the tryton server pushes the data to us21:16
dfamoratocedk: yes, i believe we can do it asynchronously... that way the tryton server does not need to way for saving/pushing any info21:17
cedkdfamorato: so if async is possible, I will prefer the RT way21:18
-!- sharoon(~sharoon@173-9-190-185-miami.txt.hfc.comcastbusiness.net) has joined #tryton21:18
dfamoratocedk: ok... so we use RT indexing the... point 1 decided..21:19
dfamoratocedk: now... in order to implement asynchronous, I think i should use queue ... what do you think21:20
cedkdfamorato: why not21:21
cedkdfamorato: but try to keep simple21:22
-!- gremly(~gremly@200.106.202.91) has joined #tryton21:23
dfamoratook... I will do some research for what we can use to glue tryton server messages and sphinx indexer....21:23
dfamoratocedk: now... regarding multiple languages.....21:23
dfamoratocedk: sphinx uses a process call stemming21:24
dfamoratocedk:  which reduces the words to their "root"  form21:25
dfamoratocedk: example:  happinness = happy       lawfull = law  and so on21:25
-!- okko(~okko@dhcp-077-251-140-095.chello.nl) has joined #tryton21:26
dfamoratocedk: using the default sphinx installation packages... it only supports english, czech and russian by default21:27
cedkdfamorato: no other language?21:27
dfamoratocedk: if we need to support additional languages on the stemming process, we need to compile sphinx from source and use the snowball libstemmer library21:28
dfamoratocedk: which then adds support for 13 languages: Danish, Dutch, Finish, French, German, Hungarian, Italia, Norwegian, Portuguese, Romanina, Spanish, Swedish and Turkisg21:28
dfamoratocedk: so I guess i should base my development using the libstemmer then... so it supports the additional languages right ?21:29
cedkdfamorato: yes21:29
dfamoratocedk: ok, now regarding the "tryton-driver" for sphinx... which will "expose"  the objects used by the tryton model to be indexed21:31
dfamoratocedk: As it was suggested to me, I was think about using the same concept that the current modules use.. searchable = 1 or 2, so we make the searchable = 1 to be full-text indexes  and searchable =2 to be attributes. Attributes can be used for sorting21:33
dfamoratocedk: Attributes are not used as full-text index.21:33
dfamoratocedk: example:  let's say we sell books.....  author name and book title should be full  text... as well as the id/code  ... all other data is attribute...21:34
cedkdfamorato: I don't understand the searchable stuff?21:35
dfamoratocedk: so, we can group and sort based on the publication date, cost, and so on21:35
cedkdfamorato: what do you mean by grouping/sorting ?21:36
dfamoratocedk: sorry.. i meant select.... select = 1 creates an index on the current backend.....21:36
dfamoratocedk: i believe we can extend the view of the search fields in order to query only a specific range of dates for example21:37
cedkdfamorato: there is a new GUI interface by bechamel to get just a textbox21:39
dfamoratocedk: on 2.0 ?21:40
dfamoratocedk: speaking about bechamel, when he will be back ?21:40
cedkdfamorato: don't know, he is there most of the time during office time in Europe21:41
dfamoratocedk: must be the timezon difference that i haven't seen him21:43
dfamoratocedk: so I will do more research then about my idea of the tryton-driver for sphix and then I get back to you guys again21:44
cedkdfamorato: yes, you can also send emails21:44
cedkdfamorato: ask bechamel about the new search view in progress21:44
dfamoratocedk:  I will21:44
dfamoratocedk: Also... I shoudl focus this development on version 2.0 right21:45
-!- cedk(~ced@gentoo/developer/cedk) has joined #tryton21:53
-!- sharoon(~sharoon@173-9-190-185-miami.txt.hfc.comcastbusiness.net) has left #tryton22:16
-!- sharoon(~sharoon@173-9-190-185-miami.txt.hfc.comcastbusiness.net) has joined #tryton22:16
-!- vladimirek(~vladimire@bband-dyn166.178-41-220.t-com.sk) has joined #tryton22:16
-!- hoRn(~jan@dslb-188-106-171-149.pools.arcor-ip.net) has joined #tryton22:30
hoRnhi22:30
hoRnI'm searching for an example to define a tree of a One2Many in the form - i don't need all the fields and i have seen such a example in an other model. Anybody has a hint?22:33
udonohoRn: hi, take a look at party/party.xml. There is the included address tree defined.22:41
udonohoRn: <tree string="Addresses" sequence="sequence" ...22:41
hoRnthank you22:53
hoRnok - nice22:54
udonohoRn: welcome22:59
hoRnudono: and nice to have you in my contacts at xing ;)23:00
udonohoRn: yes23:07
hoRnYesterday a client asked me, if it is possible to cut the seconds of datetime fields. So I was looking arround and find the code  where it is defined in the code of the client. changing this ends in errors, because there are other places where the datetimefields are accesed in tuples  with the length of the field with seconds. now I think about a solution for this ....23:13
-!- trifon_(~trifon@p548C6E73.dip.t-dialin.net) has joined #tryton23:44

Generated by irclog2html.py 2.11.0 by Marius Gedminas - find it at mg.pov.lt!