Merge pull request #13 from PyCampES/handle_pipes

add remove_pipes
PyCampES · Mar 31, 2024 · d2315c0 · d2315c0
2 parents cfa562f + 2226e22
commit d2315c0
Show file tree

Hide file tree

Showing 2 changed files with 18 additions and 0 deletions.
diff --git a/src/ficamp/classifier/preprocessing.py b/src/ficamp/classifier/preprocessing.py
@@ -17,10 +17,15 @@ def remove_digits(s: str) -> str:
     return " ".join(clean)
 
 
+def remove_pipes(s: str) -> str:
+    return " ".join(s.split("|"))
+
+
 def preprocess(s: str) -> str:
     "Clean up transaction description"
     steps = (
         lambda s: s.lower(),
+        remove_pipes,
         remove_digits,
     )
     out = s

diff --git a/tests/test_preprocessing.py b/tests/test_preprocessing.py
@@ -3,6 +3,7 @@
 from ficamp.classifier.preprocessing import (
     preprocess,
     remove_digits,
+    remove_pipes,
 )
 
 
@@ -23,6 +24,17 @@ def test_remove_digits(inp, exp):
     assert remove_digits(inp) == exp
 
 
+@pytest.mark.parametrize(
+    ("inp,exp"),
+    (
+        ("SEPA 1231AMSTERDAM", "SEPA 1231AMSTERDAM"),  # nothing to do
+        ("SEPA 1231|AMSTERDAM", "SEPA 1231 AMSTERDAM"),
+    ),
+)
+def test_remove_pipes(inp, exp):
+    assert remove_pipes(inp) == exp
+
+
 @pytest.mark.parametrize(
     ("inp,exp"),
     (
@@ -34,6 +46,7 @@ def test_remove_digits(inp, exp):
         ("SEPA 123", "sepa"),
         ("SEPA 12312321 bic", "sepa bic"),
         ("SEPA 12312321 123BIC", "sepa"),
+        ("SEPA 1231|AMSTERDAM 123BIC", "sepa amsterdam"),
     ),
 )
 def test_preprocess(inp, exp):