July 2011 - Haskell-Cafe

NLP libraries and tools?
by Dmitri O.Kondratiev 10 Jul '11

10 Jul '11

Hi, Please advise on NLP libraries similar to Natural Language Toolkit ( www.nltk.org) First of all I need: - tools to construct 'bag of words' ( http://en.wikipedia.org/wiki/Bag_of_words_model) which is a list of words in the article. - tools to prune common words, such as prepositions and conjunctions, as well as extremely rare words, such as the ones with typos. - stemming tools - Naive Bayes classifier - SVM classifier - k-means clustering Thanks!

4 8

unbuffered raw keyboard input on Windows
by José Romildo Malaquias 10 Jul '11

10 Jul '11

Hello. I want to write a Haskell console application (a game) in ghc where the user will interact using the keyboard. I need to read the keys as soon as they are typed and without echoing to the console. Although Haskell provides this capability in the standard libraries (using hSetEcho and hSetBuffering from System.IO), it does not work on Windows due to a bug (http://hackage.haskell.org/trac/ghc/ticket/2189) I have tried installing packages like vty, hscurses, and ncurses, but they did not install on Windows. Which workarounds are available for that? Romildo

2 4

Re: [Haskell-cafe] error with nanocurses-1.5.2
by José Romildo Malaquias 10 Jul '11

10 Jul '11

On Sat, Jul 09, 2011 at 01:59:12PM -0300, José Romildo Malaquias wrote: > Hello. > > When trying to install nanocurses-1.5.2 on my Fedora Linux system, I am > getting the following error: > > $ cabal install nanocurses > Resolving dependencies... > Configuring nanocurses-1.5.2... > Preprocessing library nanocurses-1.5.2... > In file included from Curses.hsc:42:0: > cbits/utils.h:7:20: fatal error: config.h: No such file or directory > compilation terminated. > compiling dist/build/UI/Nanocurses/Curses_hsc_make.c failed (exit code 1) > command was: /usr/bin/gcc -c dist/build/UI/Nanocurses/Curses_hsc_make.c > -o dist/build/UI/Nanocurses/Curses_hsc_make.o -fno-stack-protector > -fno-stack-protector -D__GLASGOW_HASKELL__=700 -Dlinux_BUILD_OS > -Dlinux_HOST_OS -Dx86_64_BUILD_ARCH -Dx86_64_HOST_ARCH -Icbits > -I/usr/lib64/ghc-7.0.2/unix-2.4.2.0/include > -I/usr/lib64/ghc-7.0.2/bytestring-0.9.1.10/include > -I/usr/lib64/ghc-7.0.2/base-4.3.1.0/include > -I/usr/lib64/ghc-7.0.2/include -I/usr/lib64/ghc-7.0.2/include/ > cabal: Error: some packages failed to install: > nanocurses-1.5.2 failed during the building phase. The exception was: > ExitFailure 1 > > It would be nice to have this buf fixed. The attached patch fixes the compilation issues. Basically it changes the build type from simple to configure in the cabal file, and it renames the function getline in cbits/utils.{h,c} to mygetline. If appropriate, please release a new version with the fixes. Romildo

1 0

JSON Querry examples
by Dmitri O.Kondratiev 10 Jul '11

10 Jul '11

Hi, I am trying to search known-in-advance fields in JSON object with unknown structure. It looks like Data.JSON2.Query is most appropriate for this task. Yet, I can not figure out how to use filters this package provides to filter JSON arrays and objects. Are there any end-to-end examples of making queries of unknown JSON objects? Thanks!

1 0

class and instance
by Patrick Browne 10 Jul '11

10 Jul '11

Hi, I am trying to understand the following code. I have written my current (mis-)understanding and questions below. I do not wish to improve the code, it is from a research paper[1] that I am trying to understand. Pat [1] ftp://ftp.geoinfo.tuwien.ac.at/medak/phdmedak.pdf -- A specification. The class Points takes two type variables. -- The functions take variables of type p, a; return a value of type a. class Points p a where getX :: p a -> a getY :: p a -> a -- A parameterized representation -- One constructor takes two elements of type a and produces a Point. data Point a = Pt a a -- An implementation of the class Points on the data type (Point a) -- In Pt b c constructor the variables b and c areboth of type a. -- The type a is left undefined until evaluation instance Points Point a where getX (Pt b c) = b getY (Pt b c) = c -- This runs with any type e.g. Integers -- getX(Pt 1 2) -- :t getX(Pt 1 2) -- getX(Pt 1 2) :: forall t. (Num t) => t My main question is in understanding the relationship between the arguments of the functions getX and getY in the class and in the instance. It seems to me that the constructor Pt 1 2 produces one element of type Point which has two components. How does this square with the class definition of getX which has two arguments? Is there a difference between: getX :: p a -> a and getX :: p -> a -> a This message has been scanned for content and viruses by the DIT Information Services E-Mail Scanning Service, and is believed to be clean. http://www.dit.ie

4 3

Haddock and the type inference
by Yves Parès 10 Jul '11

10 Jul '11

Hello Café, As far as I know, there is currently no way to tell haddock to run the type inferer to document functions without explicit type signature. Is that done on purpose (e.g. to force the developpers to write at least the type signatures of the public API, which is a fair reason) or is it a limitation? Is it planned to add that support to haddock?

1 0

Re: [Haskell-cafe] NLP libraries and tools?
by wren ng thornton 10 Jul '11

10 Jul '11

(Psst, the nlp list is <nlp(a)projects.haskell.org> :) On 7/9/11 3:10 AM, Jack Henahan wrote: > On Jul 7, 2011, at 10:53 PM, wren ng thornton wrote: >> I can't help but be a (meta)theorist. But then, I'm of the firm opinion >> that theory must be grounded in actual practice, else it belongs more to >> the realm of theology than science. > > Oof, you're liable to wound my (pure) mathematician's pride with remarks like that, wren. :P How's that now? Pure mathematics is perfectly grounded in the practice of mathematics :) I've no qualms with pure maths. Afterall, mathematics isn't trying to model anything (except itself). The problems I have are when the theory branch of a field loses touch with what the field is trying to do in the first place, and consequently ends up arguing over details which can be neither proven nor disproven. It is this which makes them non-scientific and not particularly helpful for practicing scientists. Linguistics is one of the fields where this has happened, but it's by no means the only one (AI, declarative databases, postmodernism,...) There's nothing wrong with not being science. I'm a big fan of the humanities, mathematics, and philosophy. It's only a problem when non-science is pretending to be science: it derails the scientists and it does a disservice to the non-science itself. Non-science is fine; pseudo-science is the problem. For the same reason, I despise math envy and all the pseudo-math that gets bandied about in social sciences wishing they were economics (or economics wishing it were statistics, or statistics wishing it were mathematics). -- Live well, ~wren

3 2

[ANN] PAM library, Iptables web-interface
by Evgeny Tarasov 10 Jul '11

10 Jul '11

Hello! I'd like to announce my project called Iptadmin. It's a web interface for linux Iptables firewall [1]. It supports just few Iptables options by now. The main feature of Iptadmin is user friendly help messages. Iptadmin works in a daemon mode. It doesn't keep any auxiliary information in a system, firewall configuration is taken from the «iptables-save» script. All changes is made by Iptables calling. Firewall rules are saved by «/etc/init.d/iptables save» script. Several linux distributions such as Debian require additional Iptables configuration. Iptadmin uses PAM [2] authentication. Access is granted for root and for any other user with membership in «iptadmin» system group. Iptadmin is written using Happstack, so it doesn't support https yet. To prevent sending system password by network in plain text, the web interface is listening only on loopback network interface. In order to access Iptadmin remotely user should setup https proxying using external web server. There is also dedicated library for parsing and printing Iptables rules [3]. While working on Iptadmin I've implemented library with subset of PAM api bindings[4]. A code of my project is licensed under BSD3. Sources are uploaded to hackage [3,4,5]. More information can be found on a project homepage[6]. [1] http://www.netfilter.org/ [2] http://www.kernel.org/pub/linux/libs/pam/ [3] http://hackage.haskell.org/package/iptables-helpers-0.3 [4] http://hackage.haskell.org/package/pam-0.1 [5] http://hackage.haskell.org/package/iptadmin-1.0.0 [6] http://iptadmin.confmgr.org -- Best regards, Evgeny Tarasov

1 0

Cabal uninstall
by Andrew Pennebaker 09 Jul '11

09 Jul '11

Please add an automated uninstall option for Cabal packages. It's a pain to remove them manually, and the user expectation based on other package managers (Gem, Aptitude, MacPorts, Homebrew, Yum, Emerge) is that "cabal uninstall"/"cabal remove" does the intuitive thing: remove packages and their dependencies. Cheers, Andrew Pennebaker www.yellosoft.us

3 2

Re: [Haskell-cafe] Call for GUI examples - Functional Reactive Programming
by John Lato 09 Jul '11

09 Jul '11

From: Chris Smith <cdsmith(a)gmail.com> > > On Fri, 2011-07-08 at 08:08 +0200, Heinrich Apfelmus wrote: > > Do you know any *small GUI programs* that you would *like* to see > > *implemented with Functional Reactive Programming?* > > This isn't really a specific application, but what I'd like to see most > from FRP is an example of something that involves moving between windows > and dialog boxes. All of the GUI-based FRP examples I've seen so far > involve interactions in a specific GUI layout. It's unclear to me how > FRP extends into a situation like: > > - There's a starting window showing several options for what to do > - When the user chooses an option, that window closes, and a new one > opens with that activity. > - In response to some actions, dialog boxes appear with their own > interactions. > > It's not clear to me if anyone in FRP has an idea of how stuff like this > fits in. Is there some FRP trick to handle this declaratively? Or > would you just say each move to a new window ends or pauses one network > of events and behaviors, and starts a new one? I can show you how I've been handling this with reactive-banana and gtk2hs. As a small example, the "File->Open" command is meant to open a new project. The codez: > maybeEvent0 :: Typeable a => Action -> IO (Maybe a) -> NetworkDescription (Event a) > maybeEvent0 act ops = do > (addHandler, runHandlers) <- liftIO newAddHandler > liftIO $ on act actionActivated $ ops >>= maybe (return ()) runHandlers > fromAddHandler addHandler > -- | Load a saved project file > openHandler :: ActionGroup -> Window -> NetworkDescription (Event (String, HTree)) > openHandler actGrp _win = do > act <- liftIO openAction > liftIO $ actionGroupAddActionWithAccel actGrp act Nothing > maybeEvent0 act openProjectDialog > openProjectDialog :: IO (Maybe (String, HTree)) > openProjectDialog = do > fc <- fileChooserDialogNew (Just "Select a project file") > Nothing > FileChooserActionOpen > [] > fileChooserSetSelectMultiple fc False > dialogAddButton fc stockCancel ResponseCancel > dialogAddButton fc stockOk ResponseOk > widgetShowAll fc > resp <- dialogRun fc > case resp of > ResponseOk -> runMaybeT $ do > fp <- MaybeT $ fileChooserGetFilename fc > liftIO $ widgetDestroy fc > (fp, ) <$> liftIO (readProject fp) > _ -> widgetDestroy fc >> return Nothing The `openHandler` function connects the Gtk signal to the reactive-banana framework. This uses the helper `maybeEvent0`, which filters an IO action and triggers a reactive-banana Event when that IO action actually returns a value. I've been using this pattern for most of my modal dialogs, and I think it works very well. The one dialog I haven't used it for is the starting window, which is straight-line imperative code. I don't yet have any non-modal dialogs or alternate windows, but there shouldn't be a problem. There's nothing special about which window a widget is in; it can be connected to the FRP framework just like a widget in the main window. John L.

1 0