September 2016 - Pharo-users

Re: [Pharo-users] turning a collection (interval for example) into a stream

by Denis Kudriashov

2016-09-29 12:39 GMT+02:00 Denis Kudriashov <dionisiydk(a)gmail.com>: > and collect: could be rewritten to use a generator. >> >> Now does anybody play with this already? >> >> Does xtream provide a way to do something in the same vein? >> > > Yes. XStream supports it: > > (1 to: 1000) reading transforming: [:in :out | | each | each := in get. > out put: each -1; put: each; put: each + 1] > But you probably wants different: (1 to: 1000) reading collecting: [:each | {each -1. each. each + 1} ]

Sept. 29, 2016

Re: [Pharo-users] turning a collection (interval for example) into a stream

by Denis Kudriashov

2016-09-29 6:54 GMT+02:00 stepharo <stepharo(a)free.fr>: > and now imagine that we have > > (1 to: 1000) collect: [:each | {each -1 . each . each + 1}]. > > I was wondering how we could turn it into a stream > > (1 to: 1000) asStream collect: [:each | {each -1 . each . each + 1}]. > > and what would be the asStream method. > > for index based I can see that next > > next is something like > > next > > ^ self at: (current := current + 1) > > and collect: could be rewritten to use a generator. > > Now does anybody play with this already? > > Does xtream provide a way to do something in the same vein? > Yes. XStream supports it: (1 to: 1000) reading transforming: [:in :out | | each | each := in get. out put: each -1; put: each; put: each + 1]

Sept. 29, 2016

Re: [Pharo-users] Speeding-up >>#instVarNamed: in Pharo-5.0 and beyond?

by Denis Kudriashov

Cool. What the percentage of speedup? 2016-09-28 23:41 GMT+02:00 Holger Freyther <holger(a)freyther.de>: > Hi, > > Magritte and my TagLengthValue (TLV) both use >>#instVarNamed:/>>#instVarNamed:put: > to read and write from an object. I was just running >>#bench on my SMPP > library and noticed that besides Spur Pharo5 is slower than Pharo3. > > > I added this to PointerLayout: > > >># instVarIndexFor: aString ifAbsent: aBlockClosure > | idx | > idx := 1. > slotScope do: [:each | > each isVisible ifTrue: [ > each name = aString ifTrue: [^idx]. > idx := idx +1]]. > ^aBlockClosure value > > and modified ClassDescription/TClassDescription to use it: > > "protocol: instance variables" > instVarIndexFor: instVarName ifAbsent: aBlock > "Answer the index of the named instance variable." > > | index | > index := self classLayout instVarIndexFor: instVarName ifAbsent: > [0]. > index = 0 ifTrue: > [^self superclass == nil > ifTrue: [aBlock value] > ifFalse: [self superclass instVarIndexFor: > instVarName ifAbsent: aBlock]]. > ^self superclass == nil > ifTrue: [index] > ifFalse: [index + self superclass instSize] > > > The speed-up comes from filtering allSlots to allVisibleSlots (and > creating an Array), then collecting the slot names and finally searching > the name. Does it make sense to integrate such speed-ups? > > > cheers > holger > > >

Sept. 29, 2016

Re: [Pharo-users] turning a collection (interval for example) into a stream

by Guille Polito

Hi Stef, I have some experiments I did last year in here http://www.smalltalkhub.com/#!/~Guille/ReactiveExtensions They allow you to do something like you want. It is based on the ideas in https://gist.github.com/staltz/868e7e9bc2a7b8c1f754 http://www.introtorx.com/ These guys model a stream with different transformations (such as select:, collect: flatCollect: and so on...) as a chain of observers where each transformation observes the one before, transforms the result and notifies its observers with the modified value. - The funny side is that with these they can manage polymorphically both collections as streams, and normal streams (that may never end). Also they do nice things with concurrency and retry operations fairly easily. - The odd part is DEBUGGING this. * You may have tons of elements into your collection => Then you may block all incoming events as soon as you debug one. * You have to debug chains of observer/observed pairs, with the boilerplate code that comes to it. Maybe a domain specific debugger would be need for this... Guille Guille -------- Original Message -------- > Hi > > I was looking at how we can turn a collection into a stream. > > I played with > > g := Generator on: [ :gen | > | current | > current := 0. > [ gen yield: (current := current + 1) ] repeat ]. > g next. > > and now imagine that we have > > (1 to: 1000) collect: [:each | {each -1 . each . each + 1}]. > > I was wondering how we could turn it into a stream > > (1 to: 1000) asStream collect: [:each | {each -1 . each . each + 1}]. > > and what would be the asStream method. > > for index based I can see that next > > next is something like > > next > > ^ self at: (current := current + 1) > > and collect: could be rewritten to use a generator. > > Now does anybody play with this already? > > Does xtream provide a way to do something in the same vein? > > We can have a stream of blocks. > > Stef > > > > > > > > >

Sept. 29, 2016

Re: [Pharo-users] Speeding-up >>#instVarNamed: in Pharo-5.0 and beyond?

by Holger Freyther

> On 29 Sep 2016, at 10:25, Marcus Denker <marcus.denker(a)inria.fr> wrote: > > > > Slice committed: > > https://pharo.fogbugz.com/f/cases/19155/speedup-instVarNamed > > Marcus that was quick! I had locally modified to use 1 to: specLayout size do: [] and then use specLayout at: directly. It seemed to make a small difference as well So I was "lucky" to pick PointerLayout as class to put the method in? classLayout will always be an instance of PointerLayout (or its subclasses)? thank you holger PS: It luckily speeds up instVarNamed:put: as well :)

Sept. 29, 2016

Re: [Pharo-users] Speeding-up >>#instVarNamed: in Pharo-5.0 and beyond?

by Marcus Denker

> On 29 Sep 2016, at 10:08, Marcus Denker <marcus.denker(a)inria.fr> wrote: > > >> On 28 Sep 2016, at 23:41, Holger Freyther <holger(a)freyther.de> wrote: >> >> Hi, >> >> Magritte and my TagLengthValue (TLV) both use >>#instVarNamed:/>>#instVarNamed:put: to read and write from an object. I was just running >>#bench on my SMPP library and noticed that besides Spur Pharo5 is slower than Pharo3. >> > > Yes! I will add it. > Slice committed: https://pharo.fogbugz.com/f/cases/19155/speedup-instVarNamed <https://pharo.fogbugz.com/f/cases/19155/speedup-instVarNamed> Marcus

Sept. 29, 2016

Re: [Pharo-users] PhyloclassTalk Demo Video

by Serge Stinckwich

On Thu, Sep 29, 2016 at 10:02 AM, HernÃ¡n Morales Durand <hernan.morales(a)gmail.com> wrote: > Hi Serge, > > 2016-09-29 4:05 GMT-03:00 Serge Stinckwich <serge.stinckwich(a)gmail.com>: >> >> Hi HernÃ¡n, >> >> very impressive demo ! >> > > Thanks! > >> >> We are working on similar tools in the context of epidemiological >> modelling: >> https://github.com/UMMISCO/kendrick >> >> I see the nice table views. >> Did you build something specific or reuse existing stuff ? Can you >> edit tables from UI ? > > > I built several support packages. From helpers like SpecUIAddOns and > StringExtensions to a Query Builder and a Project Framework (the Project UI > is based on the work of Peter Uhnak IIRC). The GenBank Browser could be > generalized to a XML node editor. > > For editing tables it was never a requirement for the app. Ok I will have a look later. >> Territories looks interesting. We might be interesting to use such >> information for building spatial epidemiology simulation evolving many >> countries. Looks at Ebola modelling we have done here: >> >> https://dl.dropboxusercontent.com/u/31543901/AgileVisualization/Epidemiolog… >> > > It looks really interesting! > > For Territorial I am analyzing now what we could get from Bioregions. > Integration of repositories is not so straightforward, so I plan also to > finish my Code Generator to automate generation of classes and methods when > importing CSV's and XML's, a good test case could be the Unicode's CLDR. > In PhyloclasTalk for most analysis we start always from the reported DNA > sequences, for example if we would have to analyze Ebola we would start from > here > https://www.ncbi.nlm.nih.gov/genomes/VirusVariation/Database/nph-select.cgi > and then apply NER to curate the locations or other features of isolates. If > you execute the query for Zaire ebolavirus, the PhyloclassTalk GenBank > Browser could be adapted to load the CSV (which seems pretty curated) or > accession list, and then use the Classifier directly. > I see you define routes in your code, good starting point to add the feature > :) Routes ? yes, because usually when you are doing an epidemiological modeling, you want to know how the countries are connected and what the rates of propagation between these countries. Some people are using flights routes also (they are available freely on Internet). We are not working at the level of virus DNA sequences, we are just modelling propagation of diseases. But integrating both would be amazing in the long term. I think that Territorial is not MIT, all Kendrick code is MIT and I'm bit reluctant to mix licences. This is the only drawback for me at the moment. Regards, -- Serge Stinckwich UCBN & UMI UMMISCO 209 (IRD/UPMC) Every DSL ends up being Smalltalk http://www.doesnotunderstand.org/

Sept. 29, 2016

Re: [Pharo-users] Speeding-up >>#instVarNamed: in Pharo-5.0 and beyond?

by Marcus Denker

> On 28 Sep 2016, at 23:41, Holger Freyther <holger(a)freyther.de> wrote: > > Hi, > > Magritte and my TagLengthValue (TLV) both use >>#instVarNamed:/>>#instVarNamed:put: to read and write from an object. I was just running >>#bench on my SMPP library and noticed that besides Spur Pharo5 is slower than Pharo3. > > > I added this to PointerLayout: > >>> # instVarIndexFor: aString ifAbsent: aBlockClosure > | idx | > idx := 1. > slotScope do: [:each | > each isVisible ifTrue: [ > each name = aString ifTrue: [^idx]. > idx := idx +1]]. > ^aBlockClosure value > > and modified ClassDescription/TClassDescription to use it: > > "protocol: instance variables" > instVarIndexFor: instVarName ifAbsent: aBlock > "Answer the index of the named instance variable." > > | index | > index := self classLayout instVarIndexFor: instVarName ifAbsent: [0]. > index = 0 ifTrue: > [^self superclass == nil > ifTrue: [aBlock value] > ifFalse: [self superclass instVarIndexFor: instVarName ifAbsent: aBlock]]. > ^self superclass == nil > ifTrue: [index] > ifFalse: [index + self superclass instSize] > > > The speed-up comes from filtering allSlots to allVisibleSlots (and creating an Array), then collecting the slot names and finally searching the name. Does it make sense to integrate such speed-ups? > > Yes! I will add it. Marcus

Sept. 29, 2016

Re: [Pharo-users] PhyloclassTalk Demo Video

by HernÃ¡n Morales Durand

Hi Serge, 2016-09-29 4:05 GMT-03:00 Serge Stinckwich <serge.stinckwich(a)gmail.com>: > Hi HernÃ¡n, > > very impressive demo ! > > Thanks! > We are working on similar tools in the context of epidemiological > modelling: > https://github.com/UMMISCO/kendrick > > I see the nice table views. > Did you build something specific or reuse existing stuff ? Can you > edit tables from UI ? > I built several support packages. From helpers like SpecUIAddOns and StringExtensions to a Query Builder and a Project Framework (the Project UI is based on the work of Peter Uhnak IIRC). The GenBank Browser could be generalized to a XML node editor. For editing tables it was never a requirement for the app. > Territories looks interesting. We might be interesting to use such > information for building spatial epidemiology simulation evolving many > countries. Looks at Ebola modelling we have done here: > https://dl.dropboxusercontent.com/u/31543901/AgileVisualization/ > EpidemiologicalModels/0301-EpidemiologicalModels.html > > It looks really interesting! For Territorial I am analyzing now what we could get from Bioregions. Integration of repositories is not so straightforward, so I plan also to finish my Code Generator to automate generation of classes and methods when importing CSV's and XML's, a good test case could be the Unicode's CLDR. In PhyloclasTalk for most analysis we start always from the reported DNA sequences, for example if we would have to analyze Ebola we would start from here https://www.ncbi.nlm.nih.gov/genomes/VirusVariation/Database/nph-select.cgi and then apply NER to curate the locations or other features of isolates. If you execute the query for Zaire ebolavirus, the PhyloclassTalk GenBank Browser could be adapted to load the CSV (which seems pretty curated) or accession list, and then use the Classifier directly. I see you define routes in your code, good starting point to add the feature :) Best regards, HernÃ¡n Regards, > > On Thu, Sep 29, 2016 at 7:38 AM, HernÃ¡n Morales Durand > <hernan.morales(a)gmail.com> wrote: > > > > I want to share with you a video I did just yesterday (no audio yet) of > the > > PhyloclassTalk application: > > > > https://youtu.be/1iPbDHbjbLU > > > > As some of you may know, I work in a research institute, so the workflow > is > > very specific to phylogeography and bioinformatics. But I hope you may > find > > it useful as a showcase for Pharo, or to grab some ideas. It took some > years > > to get here. > > > > It also uses code many of you wrote, so you can say Hi! if you recognize > > your source there :) > > > > Here is the web site for documentation and details: > > http://phyloclasstalk.github.io/ > > > > I would love to read your recommendations, or comments. > > > > Cheers, > > > > HernÃ¡n > > > > > > > > -- > Serge Stinckwich > UCBN & UMI UMMISCO 209 (IRD/UPMC) > Every DSL ends up being Smalltalk > http://www.doesnotunderstand.org/ > >

Sept. 29, 2016

Re: [Pharo-users] PhyloclassTalk Demo Video

by Serge Stinckwich

Hi HernÃ¡n, very impressive demo ! We are working on similar tools in the context of epidemiological modelling: https://github.com/UMMISCO/kendrick I see the nice table views. Did you build something specific or reuse existing stuff ? Can you edit tables from UI ? Territories looks interesting. We might be interesting to use such information for building spatial epidemiology simulation evolving many countries. Looks at Ebola modelling we have done here: https://dl.dropboxusercontent.com/u/31543901/AgileVisualization/Epidemiolog… Regards, On Thu, Sep 29, 2016 at 7:38 AM, HernÃ¡n Morales Durand <hernan.morales(a)gmail.com> wrote: > > I want to share with you a video I did just yesterday (no audio yet) of the > PhyloclassTalk application: > > https://youtu.be/1iPbDHbjbLU > > As some of you may know, I work in a research institute, so the workflow is > very specific to phylogeography and bioinformatics. But I hope you may find > it useful as a showcase for Pharo, or to grab some ideas. It took some years > to get here. > > It also uses code many of you wrote, so you can say Hi! if you recognize > your source there :) > > Here is the web site for documentation and details: > http://phyloclasstalk.github.io/ > > I would love to read your recommendations, or comments. > > Cheers, > > HernÃ¡n > > -- Serge Stinckwich UCBN & UMI UMMISCO 209 (IRD/UPMC) Every DSL ends up being Smalltalk http://www.doesnotunderstand.org/

Sept. 29, 2016