diff --git a/Tests.hs b/Tests.hs deleted file mode 100644 --- a/Tests.hs +++ /dev/null @@ -1,10 +0,0 @@ -import Test.Framework -import Test.Framework.Providers.HUnit -import Test.Framework.Providers.QuickCheck2 - -import Tests.NLP.FullStop ( suite ) - -main :: IO () -main = defaultMain - [ Tests.NLP.FullStop.suite - ] diff --git a/Tests/NLP/FullStop.hs b/Tests/NLP/FullStop.hs deleted file mode 100644 --- a/Tests/NLP/FullStop.hs +++ /dev/null @@ -1,50 +0,0 @@ -module Tests.NLP.FullStop ( suite ) where - -import Data.Char ( isSpace ) - -import Test.HUnit -import Test.Framework -import Test.Framework.Providers.HUnit -import Test.Framework.Providers.QuickCheck2 - -import NLP.FullStop - --- ------------------------------------------------------------ --- --- ------------------------------------------------------------ - -suite :: Test.Framework.Test -suite = - testGroup "NLP.FullStop" - [ testGroup "basic sanity checking" - [ testProperty "concat (segment s) == id s, modulo whitespace" prop_segment_concat - ] - , testGroup "segmentation" - [ testCaseSegments "simple" ["Foo.", "Bar."] "Foo. Bar." - , testCaseSegments "condense" ["What?!", "Yeah"] "What?! Yeah" - , testCaseSegments "URLs" ["Check out http://www.example.com.", "OK?"] - "Check out http://www.example.com. OK?" - , testCaseNoSplit "titles" "Mr. Doe, Mrs. Durand, St. Orolo and Dr. Singh" - , testCaseNoSplit "abbreviations" "e.g., or eg., i.e. or ie. should not be split" - , testCaseSegments "abbreviations 2" ["No lie.", "Honestly"] "No lie. Honestly" - , testCaseNoSplit "initials" "E. Y. Kow" - , testCaseNoSplit "initials 2" "Hello, E. Y. Kow" - , testCaseSegments "initials counter" [ "E. Y. Kow.", "Hello" ] "E. Y. Kow. Hello" - , testCaseNoSplit "numbers" "version 2.3.99.2" - -- TODO: what's a good way of dealing with ellipsis (...)? - -- TODO: He said "Foo." Bar (tricky because Foo. "Bar" is legit) - -- TODO: Very likely to be cases where it's just plain ambiguous - ] - ] - -testCaseNoSplit d x = testCaseSegments d [x] x -testCaseSegments d xs x = testCase d $ assertEqual "" xs (segment x) - --- TODO: perhaps create a newtype that skews the random generation of tests --- towards things that look more like text (but not too much, because we still --- want to make sure we're covering edge-cases) - -prop_segment_concat s = - noWhite s == concatMap noWhite (segment s) - where - noWhite = filter (not . isSpace) diff --git a/fullstop.cabal b/fullstop.cabal --- a/fullstop.cabal +++ b/fullstop.cabal @@ -1,5 +1,5 @@ name: fullstop -version: 0.1.2 +version: 0.1.3.1 synopsis: Simple sentence segmenter description: FullStop splits texts into sentences, using some orthographical conventions (used in English and hopefully other languages). @@ -17,35 +17,40 @@ license: BSD3 license-file: LICENSE author: Eric Kow -maintainer: Eric Kow -homepage: http://patch-tag.com/r/kowey/fullstop -cabal-version: >= 1.6 +maintainer: Eric Kow +homepage: http://hub.darcs.net/kowey/fullstop +cabal-version: >= 1.10 build-type: Custom -flag test - Description: Build the test suite - Default: False +source-repository head + type: darcs + location: http://hub.darcs.net/kowey/fullstop library exposed-modules: NLP.FullStop -- NB: there actualy isn't any particular need for GHC 6.10 - build-Depends: base >= 3 && < 4.3 + build-Depends: base >= 3 && < 5 , split == 0.1.* + default-language: Haskell98 executable fullstop main-is: fullstop.hs + build-Depends: base + , fullstop + hs-source-dirs: fullstop + default-language: Haskell98 -executable hstest-fullstop +test-suite hstest-fullstop + type: exitcode-stdio-1.0 + hs-source-dirs: test main-is: Tests.hs other-modules: Tests.NLP.FullStop - if flag(test) - build-depends: base >= 3 && < 4.3 - , HUnit == 1.2.* - , QuickCheck == 2.4.* - , test-framework == 0.3.* - , test-framework-hunit == 0.2.* - , test-framework-quickcheck2 == 0.2.* - else - Buildable: False + build-depends: base + , fullstop + , HUnit == 1.2.* + , QuickCheck == 2.4.* + , test-framework + , test-framework-hunit + , test-framework-quickcheck2 diff --git a/fullstop.hs b/fullstop.hs deleted file mode 100644 --- a/fullstop.hs +++ /dev/null @@ -1,3 +0,0 @@ -import NLP.FullStop - -main = interact (unlines . segment) diff --git a/fullstop/fullstop.hs b/fullstop/fullstop.hs new file mode 100644 --- /dev/null +++ b/fullstop/fullstop.hs @@ -0,0 +1,3 @@ +import NLP.FullStop + +main = interact (unlines . segment) diff --git a/test/Tests.hs b/test/Tests.hs new file mode 100644 --- /dev/null +++ b/test/Tests.hs @@ -0,0 +1,10 @@ +import Test.Framework +import Test.Framework.Providers.HUnit +import Test.Framework.Providers.QuickCheck2 + +import Tests.NLP.FullStop ( suite ) + +main :: IO () +main = defaultMain + [ Tests.NLP.FullStop.suite + ] diff --git a/test/Tests/NLP/FullStop.hs b/test/Tests/NLP/FullStop.hs new file mode 100644 --- /dev/null +++ b/test/Tests/NLP/FullStop.hs @@ -0,0 +1,50 @@ +module Tests.NLP.FullStop ( suite ) where + +import Data.Char ( isSpace ) + +import Test.HUnit +import Test.Framework +import Test.Framework.Providers.HUnit +import Test.Framework.Providers.QuickCheck2 + +import NLP.FullStop + +-- ------------------------------------------------------------ +-- +-- ------------------------------------------------------------ + +suite :: Test.Framework.Test +suite = + testGroup "NLP.FullStop" + [ testGroup "basic sanity checking" + [ testProperty "concat (segment s) == id s, modulo whitespace" prop_segment_concat + ] + , testGroup "segmentation" + [ testCaseSegments "simple" ["Foo.", "Bar."] "Foo. Bar." + , testCaseSegments "condense" ["What?!", "Yeah"] "What?! Yeah" + , testCaseSegments "URLs" ["Check out http://www.example.com.", "OK?"] + "Check out http://www.example.com. OK?" + , testCaseNoSplit "titles" "Mr. Doe, Mrs. Durand, St. Orolo and Dr. Singh" + , testCaseNoSplit "abbreviations" "e.g., or eg., i.e. or ie. should not be split" + , testCaseSegments "abbreviations 2" ["No lie.", "Honestly"] "No lie. Honestly" + , testCaseNoSplit "initials" "E. Y. Kow" + , testCaseNoSplit "initials 2" "Hello, E. Y. Kow" + , testCaseSegments "initials counter" [ "E. Y. Kow.", "Hello" ] "E. Y. Kow. Hello" + , testCaseNoSplit "numbers" "version 2.3.99.2" + -- TODO: what's a good way of dealing with ellipsis (...)? + -- TODO: He said "Foo." Bar (tricky because Foo. "Bar" is legit) + -- TODO: Very likely to be cases where it's just plain ambiguous + ] + ] + +testCaseNoSplit d x = testCaseSegments d [x] x +testCaseSegments d xs x = testCase d $ assertEqual "" xs (segment x) + +-- TODO: perhaps create a newtype that skews the random generation of tests +-- towards things that look more like text (but not too much, because we still +-- want to make sure we're covering edge-cases) + +prop_segment_concat s = + noWhite s == concatMap noWhite (segment s) + where + noWhite = filter (not . isSpace)