compoe I/F を導入してみた

naivesearch/indexer/chunker.py

@@ -1,14 +1,19 @@
-from typing import Callable, Protocol, List
+from typing import Callable, Protocol, List, Optional
 
 from .formatter import Formatter
 
 
-Chunker = Callable[[str], List[str]]
+class Chunker(Protocol):
+    def __init__(self, formatter: Optional[Formatter]):
+        ...
+
+    def __call__(self, x: str) -> List[str]:
+        ...
 
 
 class CharacterChunker:
-    def __init__(self, formatter: Formatter):
+    def __init__(self, formatter: Optional[Formatter]):
         self.formatter = formatter
 
     def __call__(self, x: str) -> List[str]:
-        return list(self.formatter(x))
+        return list(self.formatter(x) if self.formatter else x)

naivesearch/indexer/formatter.py

@@ -1,22 +1,29 @@
 import unicodedata
-from typing import Callable, Optional
+from typing import Callable, Optional, Protocol
 
-Formatter = Callable[[str], str]
+
+class Formatter(Protocol):
+    def __init__(self, other: Optional['Formatter'] = None, **kwargs):
+        ...
+
+    def __call__(self, x: str) -> str:
+        ...
 
 
 class UnicodeNormalizer:
-    def __init__(self, other: Optional[Formatter] = None):
+    def __init__(self, other: Optional[Formatter] = None, form: str = 'NFKC'):
         self.other = other
+        self.form = form
 
-    def __call__(self, x: str):
+    def __call__(self, x: str) -> str:
         x = self.other(x) if self.other else x
-        return unicodedata.normalize('NFKC', x)
+        return unicodedata.normalize(self.form, x)
 
 
 class LowerCaseNormalizer:
     def __init__(self, other: Optional[Formatter] = None):
         self.other = other
 
-    def __call__(self, x: str):
+    def __call__(self, x: str) -> str:
         x = self.other(x) if self.other else x
         return x.lower()

naivesearch/main.py

@@ -1,22 +1,39 @@
+from returns.curry import partial
+from typing import Callable, Type
+
 from naivesearch.indexer import InvertedIndex
-from naivesearch.indexer.formatter import UnicodeNormalizer, LowerCaseNormalizer
+from naivesearch.indexer.formatter import UnicodeNormalizer, LowerCaseNormalizer, Formatter
 from naivesearch.indexer.converter import BigramConverter
-from naivesearch.indexer.chunker import CharacterChunker
+from naivesearch.indexer.chunker import CharacterChunker, Chunker
 
 
 def naivesearch(filepath: str):
-
-    def file_reader(filepath):
-        with open(filepath) as f:
-            for line in f.readlines():
-                yield line.strip()
-
     index = InvertedIndex(
         file_reader(filepath),
         [
-            BigramConverter(CharacterChunker(
-                UnicodeNormalizer(LowerCaseNormalizer()),
-            ))
+            composed(
+                BigramConverter,
+                CharacterChunker,
+                LowerCaseNormalizer,
+                partial(UnicodeNormalizer, form='NFKC'),
+            )
         ]
     )
     return index
+
+
+def composed(
+        converter: Type[BigramConverter],
+        chunker: Type[Chunker],
+        *formatters: Callable[..., Formatter],
+):
+    result = None
+    for x in reversed(formatters):
+        result = x(result)
+    return converter(chunker(result))
+
+
+def file_reader(filepath):
+    with open(filepath) as f:
+        for line in f.readlines():
+            yield line.strip()

pyproject.toml

@@ -6,6 +6,7 @@ authors = ["Your Name <[email protected]>"]
 
 [tool.poetry.dependencies]
 python = "3.8.6"
+returns = "^0.15.0"
 
 [tool.poetry.dev-dependencies]
 mypy = "^0.790"