f5078d76c7541ab15387ab62fef22a01 186e631cb3b09ac33e5c124a79e20915 bratseth /vespa-users/msg/4498622c3f131e1067c4e7bc18ac96db This was done to make it simpler for people to get started with jars loaded through OSGi on the assumption that most wouldn't need to import anything I'm not sure it actually gives a net saving though for the reason you point out We'll probably make the existence of global packages optional somehow later Jon Den 21 mai 2011 kl 03.21 skrev LG Just out of curiosity why motivated the choice of exposing them as global packages instead of exported packages I had personally a hard time figuring out which packages I had to import and to not import javax com.yahoo.vespa etc and I'm wondering if it means I need to repackage some 3rdparty component if they decide to import some of these global packages like javax.foo lg geoinformatics software engineer mail@host.com direct y!im 701 first avenue sunnyvale ca 94089-0703 us phone fax Le 5/10/11 12:10 AM Jon S Bratseth a écrit The packages in the Vespa public api is always available global That's the ones included here http://vespa/javadoc/5.0/all In addition the packages belonging to the Java SE API javax etc are global The exported list is not quite as well defined though You can always just try to import when some package is not in the global category above if it cannot be resolved then we do not export it Or check the source as Francois mentions 8bce4afebf9dca1aab84651db62d4269 How to load JSON library in vespa 5.0.8 This was done to make it simpler for people to get started with jars loaded through OSGi on the assumption that most wouldn't need to import anything I'm not sure it actually gives a net saving though for the reason you point out We'll probably make the existence of global packages optional somehow later Jon Den 21 mai 2011 kl 03.21 skrev LG Just out of curiosity why motivated the choice of exposing them as global packages instead of exported packages I had personally a hard time figuring out which packages I had to import and to not import javax com.yahoo.vespa etc and I'm wondering if it means I need to repackage some 3rdparty component if they decide to import some of these global packages like javax.foo laurent goujon geoinformatics software engineer mail@host.com direct y!im 701 first avenue sunnyvale ca 94089-0703 us phone fax Le 5/10/11 12:10 AM Jon S Bratseth a écrit The packages in the Vespa public api is always available global That's the ones included here http://vespa/javadoc/5.0/all In addition the packages belonging to the Java SE API javax etc are global The exported list is not quite as well defined though You can always just try to import when some package is not in the global category above if it cannot be resolved then we do not export it Or check the source as Francois mentions 1306231886 0 1306228138 email (NULL) (NULL) 1 How to load JSON library in vespa 5.0.8 4498622c3f131e1067c4e7bc18ac96db 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 /2011/05/09/how_to_load_json_library_in_vespa_5_0_8 (NULL) f4467ca341409984b9c3ab4c431f0a7c c0d00a07cc1486404f5e23a2cc3cc5ed bratseth /vespa-users/msg/2538973d3c06d9d5163bf9bd069d0bd6 The unit test and doc I pointed to had some examples were you looking for something else The unit tests are using processors from this set http://vespa/view/vespa/trunk/container/processing/src/main/java/com/yahoo/processing/test/ProcessorLibrary.java?view=markup From this you might be interested in Federator forks the execution to multiple chains executed in parallel FutureDataSource creates and returns response containing future data BlockingSplitter An example of waiting for some future data to complete before using them for some processing of course don t do this if you want full async for some reason AsyncDataProcessingInitiator An example of registering future processing when a list of data is completed i.e process the data without blocking StreamProcessingInitiator An example of registering future processing on every additional piece of data entering a list None of this shows how to create a real async data source which gets its data from the network the FutureDataSource above is of course just a mock I have an example of that somewhere which I can dig up if you need it Jon On 6 feb 2014 at 02:40 GOs wrote Is there an example which uses the futures We have async working but it's not sending all of the data until the request is marked as completed Best GGO Tech Yahoo Software Systems Development Engineer M 701 First Avenue Sunnyvale CA 94089 http://forgood.zenfs.com/logos/yahoo.png On Wednesday February 5 2014 2:02 AM Jon Bratseth wrote See the processing framework in particular http://vespa/5/documentation/jdisc/processing.html#asynchronous-processing and the example AsyncDataProducer towards the end of the page In short the Processing framework supports this use case out of the box just return a Response with data futures and the renderer will render as much as possible at once and then the future data whenever it becomes available See http://vespa/view/vespa/trunk/container/core/src/test/java/com/yahoo/processing/handler/ProcessingHandlerTestCase.java?view=markup for some complete examples the tests named something with async Also note that by default the renderer will preserve the order of the future placeholders in the response such that if a Response containing future data list A and B it will wait for A before rendering further even if B is available If you want to return data as soon as possible in any order have your DataList instances implement com.yahoo.processing.response.Ordered and return false from isOrdered There s an example of that as well in the test above Jon On 5 feb 2014 at 10:43 Kristian Aune wrote vespa-users is the best list K On 5 feb 2014 at 01:24 GO wrote Hey Kristian Not sure what the ilist is for JDisc user support so if you know just reply and I'll forward this mail that way Basically here's the scenario I'm looking at We get a request from client and receive it in JDisc Then we do a bunch of processing and serve responses as they come this way the client can immediately render finished responses while still waiting for pending responses What is the easiest way to do this in a JDisc container I was looking at websockets but not sure how much they are supported Best GE desde mi iPhone Vespa Information http://vespa 1c809fe04d3807dc72be71cf1aa559d6 JDisc streaming / websockets The unit test and doc I pointed to had some examples were you looking for something else The unit tests are using processors from this set http://vespa/view/vespa/trunk/container/processing/src/main/java/com/yahoo/processing/test/ProcessorLibrary.java?view=markup From this you might be interested in Federator forks the execution to multiple chains executed in parallel FutureDataSource creates and returns response containing future data BlockingSplitter An example of waiting for some future data to complete before using them for some processing of course don t do this if you want full async for some reason AsyncDataProcessingInitiator An example of registering future processing when a list of data is completed i.e process the data without blocking StreamProcessingInitiator An example of registering future processing on every additional piece of data entering a list None of this shows how to create a real async data source which gets its data from the network the FutureDataSource above is of course just a mock I have an example of that somewhere which I can dig up if you need it Jon On 6 feb 2014 at 02:40 GO wrote Is there an example which uses the futures We have async working but it's not sending all of the data until the request is marked as completed Best GGO Tech Yahoo Software Systems Development Engineer MFirst Avenue Sunnyvale CA 94089 http://forgood.zenfs.com/logos/yahoo.png On Wednesday February 5 2014 2:02 AM Jon Bratseth wrote See the processing framework in particular http://vespa/5/documentation/jdisc/processing.html#asynchronous-processing and the example AsyncDataProducer towards the end of the page In short the Processing framework supports this use case out of the box just return a Response with data futures and the renderer will render as much as possible at once and then the future data whenever it becomes available See http://vespa/view/vespa/trunk/container/core/src/test/java/com/yahoo/processing/handler/ProcessingHandlerTestCase.java?view=markup for some complete examples the tests named something with async Also note that by default the renderer will preserve the order of the future placeholders in the response such that if a Response containing future data list A and B it will wait for A before rendering further even if B is available If you want to return data as soon as possible in any order have your DataList instances implement com.yahoo.processing.response.Ordered and return false from isOrdered There s an example of that as well in the test above Jon On 5 feb 2014 at 10:43 Kristian Aune wrote vespa-users is the best list K On 5 feb 2014 at 01:24 GO wrote Hey Kristian Not sure what the ilist is for JDisc user support so if you know just reply and I'll forward this mail that way Basically here's the scenario I'm looking at We get a request from client and receive it in JDisc Then we do a bunch of processing and serve responses as they come this way the client can immediately render finished responses while still waiting for pending responses What is the easiest way to do this in a JDisc container I was looking at websockets but not sure how much they are supported Best GE desde mi iPhone Vespa Information http://vespa 1391683805 0 1391680103 email (NULL) (NULL) 1 JDisc streaming / websockets 2538973d3c06d9d5163bf9bd069d0bd6 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 /2014/02/05/jdisc_streaming_websockets (NULL) 39e8f15ad422399d1c72d3f178f956c2 d9d95e6364aaac73b3bfb77f377236fc /vespa-users/msg/4f12250dae5ee19458973a0afaf6967e I don't get any error while deployment the expression using created_at directly I have filed a support BZ 3455702 for this Thanks Yi On Mar 7 2010 at 10:23 AM Jo Kristian Bergum wrote On Mar 5 2010 at 11:47 PM YZ wrote Hi folks I'm trying to confirm the range of value for nativeRank is between 0 and 1 With this rank profile rank-profile native first-phase expression nativeRank I get the following result Question is relevancy showing the ranking score If so why it's 0 Depends on your query what fields were searched and your search definition I did another test this time using rank-profile created_at first-phase expression created_at This should have failed during deploy correct is expression attribute(created_at You probably have quite a few failed blueprint compilation errors in the vespa.log on the search nodes The result is as follows relevancy is still 0 and I would expect the value of created_at which is 1267569885 1267569885 Thanks Y df5b35d538401b06cfbe803b55856a3e Value of nativeRank I don't get any error while deployment the expression using created_at directly I have filed a support BZ 3455702 for this Thanks Yi On Mar 7 2010 at 10:23 AM Jo Kristian Bergum wrote On Mar 5 2010 at 11:47 PM YZ wrote Hi folks I'm trying to confirm the range of value for nativeRank is between 0 and 1 With this rank profile rank-profile native first-phase expression nativeRank I get the following result Question is relevancy showing the ranking score If so why it's 0 Depends on your query what fields were searched and your search definition I did another test this time using rank-profile created_at first-phase expression created_at This should have failed during deploy correct is expression attribute(created_at You probably have quite a few failed blueprint compilation errors in the vespa.log on the search nodes The result is as follows relevancy is still 0 and I would expect the value of created_at which is 1267569885 1267569885 Thanks Y 1267989987 0 1267986312 email (NULL) (NULL) 1 Value of nativeRank 4f12250dae5ee19458973a0afaf6967e SSBkb24ndCBnZXQgYW55IGVycm9yIHdoaWxlIGRlcGxveW1lbnQgdGhlIGV4cHJlc3Npb24gdXNpbmcgY3JlYXRlZF9hdCBkaXJlY3RseS4KSSBoYXZlIGZpbGVkIGEgc3VwcG9ydCBCWiAzNDU1NzAyIGZvciB0aGlzLgoKVGhhbmtzLAotWWkKCk9uIE1hciA3LCAyMDEwLCBhdCAxMDoyMyBBTSwgSm8gS3Jpc3RpYW4gQmVyZ3VtIHdyb3RlOgoKCk9uIE1hciA1LCAyMDEwLCBhdCAxMTo0NyBQTSwgWWkgWmhhbmcgd3JvdGU6Cgo+IEhpIGZvbGtzLAo+Cj4gICBJJ20gdHJ5aW5nIHRvIGNvbmZpcm0gdGhlIHJhbmdlIG9mIHZhbHVlIGZvciBuYXRpdmVSYW5rIGlzIGJldHdlZW4KPiAwIGFuZCAxLgo+Cj4gICBXaXRoIHRoaXMgcmFuayBwcm9maWxlOgo+ICAgcmFuay1wcm9maWxlIG5hdGl2ZXsKPiAgICAgZmlyc3QtcGhhc2UgeyBleHByZXNzaW9uOiBuYXRpdmVSYW5rIH0KPiAgIH0KPiAgIEkgZ2V0IHRoZSBmb2xsb3dpbmcgcmVzdWx0Ogo+Cj4gPHJlc3VsdCB0b3RhbC1oaXQtY291bnQ9IjIiPgo+IDxoaXQgcmVsZXZhbmN5PSIwIiBzb3VyY2U9InNjMC5udW0wIj4uLi4KPgo+ICAgUXVlc3Rpb246IGlzICJyZWxldmFuY3kiIHNob3dpbmcgdGhlIHJhbmtpbmcgc2NvcmU/IElmIHNvLCB3aHkgaXQncwo+IDA/CgpEZXBlbmRzIG9uIHlvdXIgcXVlcnkgKHdoYXQgZmllbGRzIHdlcmUgc2VhcmNoZWQpIGFuZCB5b3VyIHNlYXJjaApkZWZpbml0aW9uLgoKCj4gICBJIGRpZCBhbm90aGVyIHRlc3QsIHRoaXMgdGltZSB1c2luZwo+ICAgcmFuay1wcm9maWxlIGNyZWF0ZWRfYXR7Cj4gICAgIGZpcnN0LXBoYXNlIHsgIGV4cHJlc3Npb246IGNyZWF0ZWRfYXQgfQo+ICAgfQoKVGhpcyBzaG91bGQgaGF2ZSBmYWlsZWQgZHVyaW5nIGRlcGxveSwgY29ycmVjdCBpcwoKZXhwcmVzc2lvbjogYXR0cmlidXRlKGNyZWF0ZWRfYXQpCgpZb3UgcHJvYmFibHkgaGF2ZSBxdWl0ZSBhIGZldyAiZmFpbGVkIGJsdWVwcmludCBjb21waWxhdGlvbiIgZXJyb3JzIGluCnRoZSB2ZXNwYS5sb2cgb24gdGhlIHNlYXJjaCBub2Rlcz8KCgo+ICAgVGhlIHJlc3VsdCBpcyBhcyBmb2xsb3dzLCByZWxldmFuY3kgaXMgc3RpbGwgMCwgYW5kIEkgd291bGQgZXhwZWN0Cj4gdGhlIHZhbHVlIG9mIGNyZWF0ZWRfYXQsIHdoaWNoIGlzIDEyNjc1Njk4ODUuCj4gPHJlc3VsdCB0b3RhbC1oaXQtY291bnQ9IjIiPgo+IDxoaXQgcmVsZXZhbmN5PSIwIiBzb3VyY2U9InNjMC5udW0wIj4KPiA8ZmllbGQgbmFtZT0iY3JlYXRlZF9hdCI+MTI2NzU2OTg4NTwvZmllbGQ+Cj4KPiBUaGFua3MsCj4gLVlpCj4gPG1lc3NhZ2UtZm9vdGVyLnR4dD4= /2010/03/05/value_of_nativerank (NULL) 1a60037c3896299b0a90dd5d62034529 7d08e2ecdc2a7d243a748059db0e09ee bratseth /vespa-users/msg/3eac9311090c9cf7c34f9d8c897f2c4b search=incr&restrict=abc RJ skrev Hi Jo We have multiple SDs deployed in one cluster and we normally use search= in our queries How do we use it for differentiating both the cluster name and the sd name For example We have two clusters namely incr and realtime each having SDs abc xyz We need a way to query from incr cluster in abc sd Thanks RJ Jo Kristian Bergum wrote On Mon 2009-08-24 at 16:55 +0530 AS wrote On the serving side how do we indicate the cluster to get results from search= /JKB 00b89ea33262247e728f7534c9eafd0d Need a real time index along with an Incremental index search=incr&restrict=abc RJ skrev Hi Jo We have multiple SDs deployed in one cluster and we normally use search= in our queries How do we use it for differentiating both the cluster name and the sd name For example We have two clusters namely incr and realtime each having SDs abc xyz We need a way to query from incr cluster in abc sd Thanks RJ Jo Kristian Bergum wrote On Mon 2009-08-24 at 16:55 +0530 AS wrote On the serving side how do we indicate the cluster to get results from search= /JKB 1265031079 0 1251115178 email (NULL) (NULL) 1 Need a real time index along with an Incremental index 3eac9311090c9cf7c34f9d8c897f2c4b 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 /2009/08/20/partial_update_of_the_entire_document (NULL) 5b3a4526d33f45b09f63d43f476f824a 3700f8d3da3a8963f34f11816ffda2cc bergum /vespa-users/msg/a0c55f0c7669d097e470182bd7d411cf On Mon 2009-05-18 at 15:44 +0800 JLB wrote Hi All Am looking for a documentation on how to setup Vespa 1.1.5 I found this link http://vespa/documentation/setup/howtorun.html and it's not available anymore Am just wondering if you have it somewhere If so can you send it to me Thank you Why do you want to use Vespa 1.1.5 It's legacy and not longer supported Please consider using latest stable vespa release http://vespa Best Jo Kristian J plain text document attachment message-footer.txt Vespa Information http://vespa acca854c8876330b43143c61a1e1f32c Vespa 1.1.5 Documentation On Mon 2009-05-18 at 15:44 +0800 JLB wrote Hi All Am looking for a documentation on how to setup Vespa 1.1.5 I found this link http://vespa/documentation/setup/howtorun.html and it's not available anymore Am just wondering if you have it somewhere If so can you send it to me Thank you Why do you want to use Vespa 1.1.5 It's legacy and not longer supported Please consider using latest stable vespa release http://vespa Best Jo Kristian J plain text document attachment message-footer.txt Vespa Information http://vesp 1265031079 0 1242632220 email (NULL) (NULL) 1 Vespa 1.1.5 Documentation a0c55f0c7669d097e470182bd7d411cf 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 /2009/05/18/vespa_1_1_5_documentation (NULL) cb34c66221f7188a09e6151062c14e16 fba771ebc63a57f191cf783b0a59298f peng /vespa-users/msg/079588cac2dfd1117bef476409da724b There is no rules to write When you enable libyell_poststemmer if a word is not known by dictionary libyell will try to do plural singular stemming according to its builtin rules Peng Original Message From BK mailto:mail@yhost.com Sent Tuesday July 03 2007 3:10 AM To mail@host.com Subject rule based stemming in vespa Hi I came across this document http://VespaStemming on plural singular stemming I want to try rule based plural singular stemming libyell_poststemmer in vespa Can someone point me to the relevant documents on how to write these rules and use them Thanks B 3be4bf063c46c7e3d336dbfb4c58f6e7 Rule based stemming in vespa There is no rules to write When you enable libyell_poststemmer if a word is not known by dictionary libyell will try to do plural singular stemming according to its builtin rules Peng Original Message From BK mailto:mail@yhost.com Sent Tuesday July 03 2007 3:10 AM To mail@host.com Subject rule based stemming in vespa Hi I came across this document http://LocalVespaStemming on plural singular stemming I want to try rule based plural singular stemming libyell_poststemmer in vespa Can someone point me to the relevant documents on how to write these rules and use them Thanks B 1265031079 0 1184187769 email (NULL) (NULL) 1 Rule based stemming in vespa 079588cac2dfd1117bef476409da724b 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 /2007/07/03/rule_based_stemming_in_vespa (NULL)