diff --git a/haskell-postal.cabal b/haskell-postal.cabal --- a/haskell-postal.cabal +++ b/haskell-postal.cabal @@ -1,5 +1,5 @@ name: haskell-postal -version: 0.2.0.1 +version: 0.2.1.0 synopsis: Haskell binding for the libpostal library description: Provides an interface for the libpostal NLP library homepage: https://github.com/netom/haskell-postal#readme @@ -25,9 +25,12 @@ , text >= 1.2 && < 1.3 , inline-c >= 0.6 && < 0.7 -executable haskell-postal - hs-source-dirs: src - main-is: main.hs +test-suite spec + type: exitcode-stdio-1.0 + hs-source-dirs: + test + , src + main-is: Spec.hs default-language: Haskell2010 pkgconfig-depends: libpostal ghc-options: -Wall @@ -36,8 +39,10 @@ , bytestring >= 0.10 && < 0.11 , text >= 1.2 && < 1.3 , inline-c >= 0.6 && < 0.7 + , hspec >= 2.5 && < 2.6 other-modules: NLP.Postal + , NLP.Postal.PostalSpec source-repository head type: git diff --git a/src/NLP/Postal.hs b/src/NLP/Postal.hs --- a/src/NLP/Postal.hs +++ b/src/NLP/Postal.hs @@ -31,6 +31,7 @@ import Data.Monoid ((<>)) import Foreign.C.Types import qualified Language.C.Inline as C +import qualified Language.C.Inline.Unsafe as CU import Language.C.Inline.Context as CX import Foreign.Ptr @@ -59,22 +60,22 @@ -- |Calls libpostal_setup() to set up the library -- Make sure you call this before anything else setup :: IO Int -setup = fromIntegral <$> [C.exp| int { libpostal_setup() } |] +setup = fromIntegral <$> [CU.exp| int { libpostal_setup() } |] -- |Calls libpostal_setup_parser() to set up the address parser -- Call this before trying to parse addresses setupParser :: IO Int -setupParser = fromIntegral <$> [C.exp| int { libpostal_setup_parser() } |] +setupParser = fromIntegral <$> [CU.exp| int { libpostal_setup_parser() } |] -- |Calls libpostal_setup_language_classifier() to set up the classifier -- Call this before doing address normaliation setupLanguageClassifier :: IO Int -setupLanguageClassifier = fromIntegral <$> [C.exp| int { libpostal_setup_language_classifier() } |] +setupLanguageClassifier = fromIntegral <$> [CU.exp| int { libpostal_setup_language_classifier() } |] -- |Returns default parser options -- returned by libpostal_get_address_parser_default_options() getAddressParserDefaultOptions :: IO (Ptr AddressParserOptions) -getAddressParserDefaultOptions = castPtr <$> [C.block| +getAddressParserDefaultOptions = castPtr <$> [CU.block| void * { libpostal_address_parser_options_t options = libpostal_get_address_parser_default_options(); void * hoptions = malloc(sizeof(options)); @@ -85,7 +86,7 @@ -- |Returns default normalizer options -- returned by libpostal_get_default_options() getDefaultNormalizeOptions :: IO (Ptr NormalizeOptions) -getDefaultNormalizeOptions = castPtr <$> [C.block| +getDefaultNormalizeOptions = castPtr <$> [CU.block| void * { libpostal_normalize_options_t options = libpostal_get_default_options(); void * hoptions = malloc(sizeof(options)); @@ -98,7 +99,7 @@ -- parts as a list of key-value tuples (an association list). parseAddress :: Ptr AddressParserOptions -> T.Text -> IO [(T.Text, T.Text)] parseAddress options address = do - response <- [C.exp| void * { libpostal_parse_address($bs-ptr:bsAddress, * (libpostal_address_parser_options_t *) $(void * castedOptions)) } |] + response <- [CU.exp| void * { libpostal_parse_address($bs-ptr:bsAddress, * (libpostal_address_parser_options_t *) $(void * castedOptions)) } |] when (response == nullPtr) $ fail "libpostal_parse_address returned NULL a pointer" @@ -133,7 +134,7 @@ , TE.decodeUtf8 $ fromForeignPtr (castForeignPtr compFPtr) 0 (fromIntegral compLen) ) - [C.exp| void { libpostal_address_parser_response_destroy((libpostal_address_parser_response_t *) $(void * response)) } |] + [CU.exp| void { libpostal_address_parser_response_destroy((libpostal_address_parser_response_t *) $(void * response)) } |] return result @@ -147,7 +148,7 @@ expandAddress :: Ptr NormalizeOptions -> T.Text -> IO [T.Text] expandAddress options address = do (numExpansions, expansions) <- C.withPtr $ \numExpansions -> - [C.exp| char * * { libpostal_expand_address($bs-ptr:bsAddress, * (libpostal_normalize_options_t *) $(void * cOptions), $(size_t * numExpansions) ) } |] + [CU.exp| char * * { libpostal_expand_address($bs-ptr:bsAddress, * (libpostal_normalize_options_t *) $(void * cOptions), $(size_t * numExpansions) ) } |] -- TODO: this also turns an array of (char *) to a list of Texts. result <- forM [0..numExpansions-1] $ \i -> do @@ -163,7 +164,7 @@ return $ TE.decodeUtf8 $ fromForeignPtr (castForeignPtr xpFPtr) 0 (fromIntegral xpLen) - [C.exp| void { libpostal_expansion_array_destroy($(char * * expansions), $(size_t numExpansions)) } |] + [CU.exp| void { libpostal_expansion_array_destroy($(char * * expansions), $(size_t numExpansions)) } |] return result @@ -173,12 +174,12 @@ -- |Calls libpostal_teardown_parser() tearDownParser :: IO () -tearDownParser = [C.exp| void { libpostal_teardown_parser() } |] +tearDownParser = [CU.exp| void { libpostal_teardown_parser() } |] -- |Calls libpostal_teardown_language_classifier() tearDownLanguageClassifier :: IO () -tearDownLanguageClassifier = [C.exp| void { libpostal_teardown_language_classifier() } |] +tearDownLanguageClassifier = [CU.exp| void { libpostal_teardown_language_classifier() } |] -- |Calls libpostal_teardown() tearDown :: IO () -tearDown = [C.exp| void { libpostal_teardown() } |] +tearDown = [CU.exp| void { libpostal_teardown() } |] diff --git a/src/main.hs b/src/main.hs deleted file mode 100644 --- a/src/main.hs +++ /dev/null @@ -1,37 +0,0 @@ -{-# LANGUAGE OverloadedStrings #-} - -module Main where - -import Control.Monad -import System.Environment -import System.Exit -import NLP.Postal - -import qualified Data.Text as T - -main :: IO () -main = do - args <- getArgs - - when (length args /= 1) $ die "Usage: haskell-postal
" - - let address = T.pack $ head args - - putStrLn "Setting up libpostal..." - setup >>= \x -> when (x == 0) $ die "FAIL" - - putStrLn "Setting up the parser..." - setupParser >>= \x -> when (x == 0) $ die "FAIL" - - putStrLn "Setting up the language classifier..." - setupLanguageClassifier >>= \x -> when (x == 0) $ die "FAIL" - - defParserOpts <- getAddressParserDefaultOptions - defNormalizeOptions <- getDefaultNormalizeOptions - - parseAddress defParserOpts address >>= print - expandAddress defNormalizeOptions address >>= print - - tearDownLanguageClassifier - tearDownParser - tearDown diff --git a/test/NLP/Postal/PostalSpec.hs b/test/NLP/Postal/PostalSpec.hs new file mode 100644 --- /dev/null +++ b/test/NLP/Postal/PostalSpec.hs @@ -0,0 +1,49 @@ +{-# LANGUAGE OverloadedStrings #-} + +module NLP.Postal.PostalSpec (spec) where + +import Test.Hspec +import NLP.Postal +import Control.Monad (forM_) +import Control.Monad.IO.Class (liftIO) +import Data.List (lookup) +import qualified Data.Text as T + +spec :: Spec +spec = do + describe "setup" $ do + it "general setup runs successfully" $ do + r <- liftIO setup + r `shouldBe` 1 + it "parser setup runs successfully" $ do + r <- liftIO setupParser + r `shouldBe` 1 + it "language classifier setup runs successfully" $ do + r <- liftIO setupLanguageClassifier + r `shouldBe` 1 + + describe "expansion" $ do + it "expands an address correctly" $ do + defNormalizeOptions <- liftIO getDefaultNormalizeOptions + r <- liftIO $ expandAddress defNormalizeOptions "11 Wall Street New York, NY" + forM_ r $ \xp -> T.take 14 xp `shouldBe` "11 wall street" + + describe "parser" $ do + it "parses an address correctly" $ do + defParserOpts <- liftIO getAddressParserDefaultOptions + r <- liftIO $ parseAddress defParserOpts "11 Wall Street New York, NY" + lookup "house_number" r `shouldBe` Just "11" + lookup "road" r `shouldBe` Just "wall street" + lookup "city" r `shouldBe` Just "new york" + lookup "state" r `shouldBe` Just "ny" + + describe "teardown" $ do + it "language classifier teardown" $ do + r <- liftIO tearDownLanguageClassifier + r `shouldBe` () + it "parser teardown" $ do + r <- liftIO tearDownParser + r `shouldBe` () + it "teardown" $ do + r <- liftIO tearDown + r `shouldBe` () diff --git a/test/Spec.hs b/test/Spec.hs new file mode 100644 --- /dev/null +++ b/test/Spec.hs @@ -0,0 +1,1 @@ +{-# OPTIONS_GHC -F -pgmF hspec-discover #-}