Filtering works on multiple values

src/GEval/DataSource.hs

@@ -22,7 +22,9 @@ import Data.Conduit.SmartSource
 import Data.Conduit.Header
 import GEval.Selector
 
-data Filter = NoFilter | FilterByFeatures (Maybe TabularHeader) String
+import qualified Data.Set as S
+
+data Filter = NoFilter | FilterByFeatures (Maybe TabularHeader) (S.Set String)
 
 noFilter :: Filter
 noFilter = NoFilter
@@ -33,19 +35,19 @@ applyFilter (FilterByFeatures mInHeader featureSpec) tR = applyFeatureFilter mIn
 
 getFilterForScheme :: Maybe TabularHeader -> EvaluationScheme -> Filter
 getFilterForScheme mTabHeader (EvaluationScheme _ ops) = case findFilter ops of
-  Nothing -> NoFilter
+  [] -> NoFilter
-  Just f -> FilterByFeatures mTabHeader (unpack $ fixIndex f)
+  fs -> FilterByFeatures mTabHeader (S.fromList $ Prelude.map (unpack . fixIndex) fs)
 
 fixIndex = replace "[" "<" . replace "]" ">"
 
-findFilter :: [PreprocessingOperation] -> Maybe Text
+findFilter :: [PreprocessingOperation] -> [Text]
-findFilter [] = Nothing
+findFilter [] = []
-findFilter ((FeatureFilter f):_) = Just f
+findFilter ((FeatureFilter f):ops) = (f:(findFilter ops))
 findFilter (_:ops) = findFilter ops
 
 
-applyFeatureFilter :: Maybe TabularHeader -> String -> TargetRecord -> Bool
+applyFeatureFilter :: Maybe TabularHeader -> S.Set String -> TargetRecord -> Bool
-applyFeatureFilter mInHeader featureSpec tR = featureSpec `elem` (Prelude.map show fs)
+applyFeatureFilter mInHeader featureSpec tR = featureSpec `S.isSubsetOf` (S.fromList $ Prelude.map show fs)
   where fs = getFeatures Nothing
                          BlackBoxDebuggingOptions {
                            bbdoMinFrequency = 0,

test/Spec.hs

@@ -137,6 +137,8 @@ main = hspec $ do
       runGEvalTest "accuracy-on-sorted" `shouldReturnAlmost` 0.75
     it "with filtering" $
       runGEvalTest "accuracy-filtering" `shouldReturnAlmost` 0.6666
+    it "with filtering 2" $
+      runGEvalTest "accuracy-multiple-filtering" `shouldReturnAlmost` 0.5
   describe "F-measure" $ do
     it "simple example" $
       runGEvalTest "f-measure-simple" `shouldReturnAlmost` 0.57142857

test/accuracy-multiple-filtering/accuracy-multiple-filtering-solution/test-A/out.tsv

@@ -0,0 +1,8 @@
+foo
+bar
+bar
+foo
+foo
+foo
+foo
+bar

test/accuracy-multiple-filtering/accuracy-multiple-filtering/config.txt

@@ -0,0 +1 @@
+--metric Accuracy:f<in[1]:en>N<xyz>f<in[1]:general>

test/accuracy-multiple-filtering/accuracy-multiple-filtering/test-A/expected.tsv

@@ -0,0 +1,8 @@
+foo
+bar
+foo
+foo
+foo
+foo
+foo
+bar

test/accuracy-multiple-filtering/accuracy-multiple-filtering/test-A/in.tsv

@@ -0,0 +1,8 @@
+general
+de specific
+general en
+
+general
+general
+en specific
+en general