refactor into transformers

src/pyalibert/__init__.py

@@ -1,4 +1,4 @@
 
 from .parser import Parser, just, regex, end
 from .result import Result, ParseError
-from .forward import FwDeclaration
+from .forward import Declare

src/pyalibert/forward.py

@@ -1,23 +1,39 @@
-from typing import Generic, TypeVar
+from typing import Generic, Optional, TypeVar
+
+from .transformer import Transformer
 from .result import Result
 from .parser import Parser
 
 
 P = TypeVar("P")
-class FwDeclaration(Generic[P]):
-    parser: None | Parser[P]
+class FutureTransform(Transformer[P]):
+    actual: Optional[Transformer[P]]
+    def __init__(self) -> None:
+        self.actual = None
+    
+    def define(self, actual: Transformer[P]):
+        if self.actual is not None:
+            raise Exception("Redefinition of parser.")
+        self.actual = actual
+    
+    def parse(self, stream: str, at_index: int) -> Result[P]:
+        if self.actual is None:
+            raise Exception("Using forwarded definition of parser without defining first.")
+        return self.actual.parse(stream, at_index)
+
+
+P = TypeVar("P")
+class Declare(Generic[P]):
+    parser: Parser[P]
+    transform: FutureTransform[P]
 
     def __init__(self) -> None:
-        self.parser = None
+        transform = FutureTransform[P]()
+        self.parser = Parser(transform)
+        self.transform = transform
 
     def p(self):
-        def inner(stream: str, index: int) -> Result[P]:
-            if self.parser is None:
-                raise Exception("Using forwarded definition of parser without defining first.")
-            return self.parser.inner(stream, index)
-        return Parser(inner)
+        return self.parser
 
     def define(self, parser: Parser[P]):
-        if self.parser is not None:
-            raise Exception("Redefinition of parser.")
-        self.parser = parser
+        self.transform.define(parser.inner)

src/pyalibert/parser.py

@@ -1,165 +1,81 @@
-from typing import Callable, Generic, TypeVar, Union
-from .result import Result, Success, Failure, ParseError
-import re
+from dataclasses import dataclass
+from typing import Callable, Generic, TypeVar
+
+from .result import ParseError, Failure
+from .transformer import Transformer
+from .transformers import (
+    EndTransform,
+    JustTransform,
+    RegexTransform,
+    AndTransform,
+    OrTransform,
+    ListTransform,
+    SepListTransform,
+    OptionTransform,
+    MapTransform,
+    ValueTransform
+)
 
 
 def regex(pattern: str):
-    def inner(stream: str, index: int) -> Result[str]:
-        do_match = re.match(pattern, stream[index:])
-        if do_match is None:
-            return Result.failure(index, set(f"matching /{pattern}/"))
-        else:
-            match = do_match[0]
-            return Result.success(match, index + len(match))
-    return Parser(inner)
-
-def just(text: str):
-    length = len(text)
-    def inner(stream: str, index: int) -> Result[str]:
-        end = index + length
-        if stream[index:end] == text:
-            return Result.success(text, end)
-        else:
-            return Result.failure(index, set([text]))
-    return Parser(inner)
+    return Parser(RegexTransform(pattern))
 
 
-def end() -> "Parser[None]":
-    def inner(stream: str, index: int):
-        if stream[index:] == "":
-            return Result.success(None, index)
-        else:
-            return Result.failure(index, set())
-    return Parser(inner)
+def just(word: str):
+    return Parser(JustTransform(word))
+
+
+def end():
+    return Parser(EndTransform())
 
 
 P = TypeVar("P")
-O = TypeVar("O")
 T = TypeVar("T")
-
+O = TypeVar("O")
+@dataclass
 class Parser(Generic[P]):
-    inner: Callable[[str, int], Result[P]]
+    inner: "Transformer[P]"
 
-    def __init__(self, inner: Callable[[str, int], Result[P]]):
-        self.inner = inner
+    def parse(self, stream: str) -> P:
+        parser = self.and_(end()).map(lambda value : value[0])
+        parsed = parser.inner.parse(stream, 0)
+        if isinstance(parsed, Failure):
+            raise ParseError(parsed, stream)
+        return parsed.value
 
-    def parse(self, stream: str):
-        (result, _rest) = self.and_then(end()).parse_part(stream)
-        (value, _end) = result
-        return value
+    def and_(self, other: "Parser[T]"):
+        return Parser(AndTransform(self.inner, other.inner))
 
-    def parse_part(self, stream: str):
-        result = self.inner(stream, 0)
-        if isinstance(result.actual, Success):
-            rest = stream[result.actual.next_index:]
-            return (result.actual.value, rest)
-        else:
-            raise ParseError(result.actual, stream)
-
-    def map(self, transform: Callable[[P], O]) -> "Parser[O]":
-        def inner(stream: str, index: int):
-            result = self.inner(stream, 0)
-            if isinstance(result.actual, Success):
-                mapped = transform(result.actual.value)
-                return Result.success(mapped, result.actual.next_index)
-            return result
-        return Parser(inner)
-
-    def and_then(self, other: "Parser[T]"):
-        def inner(stream: str, index: int) -> Result[tuple[P, T]]:
-            result = self.inner(stream, index)
-            if isinstance(result.actual, Failure):
-                return Result(result.actual)
-            value_left = result.actual.value
-            next_index = result.actual.next_index
-
-            result = other.inner(stream, next_index)
-            if isinstance(result.actual, Failure):
-                return Result(result.actual)
-            value_right = result.actual.value
-            next_index = result.actual.next_index
-
-            return Result.success((value_left, value_right), next_index)
-        return Parser(inner)
-
-    def or_else(self, other: "Parser[T]"):
-        def inner(stream: str, index: int) -> Result[Union[P, T]]:
-            result_left = self.inner(stream, index)
-            if isinstance(result_left.actual, Success):
-                return Result.success(result_left.actual.value, result_left.actual.next_index)
-            result_right = other.inner(stream, index)
-            if isinstance(result_right.actual, Success):
-                return Result.success(result_right.actual.value, result_right.actual.next_index)
-            return Result.failure(index, result_left.actual.expected.union(result_right.actual.expected))
-        return Parser(inner)
-
-    def repeat(self):
-        def inner(stream: str, index: int) -> Result[list[P]]:
-            values = list[P]()
-            while True:
-                result = self.inner(stream, index)
-                if isinstance(result.actual, Failure):
-                    break
-                values.append(result.actual.value)
-                if result.actual.next_index == index:
-                    raise Exception("Parsing empty patterns repeatedly.")
-                index = result.actual.next_index
-            return Result.success(values, index)
-        return Parser(inner)
+    def or_(self, other: "Parser[T]"):
+        return Parser(OrTransform(self.inner, other.inner))
 
     def or_not(self):
-        def inner(stream: str, index: int) -> Result[Union[P, None]]:
-            result = self.inner(stream, index)
-            if isinstance(result.actual, Failure):
-                return Result.success(None, index)
-            return Result.success(result.actual.value, result.actual.next_index)
-        return Parser(inner)
+        return Parser(OptionTransform(self.inner))
 
-    def value(self, value: T) -> "Parser[T]":
-        return self.map(lambda _: value)
+    def repeat(self):
+        return Parser(ListTransform(self.inner))
 
-    def sep_by(self, other: "Parser[T]"):
-        parser = self.or_not().and_then(other.and_then(self).repeat())
-        def mapping(value: tuple[P | None, list[tuple[T, P]]]):
-            (first, rest) = value
-            if first is None: return list[P]()
-            return [first, *(value for (_sep, value) in rest)]
-        mapped = parser.map(mapping)
-        return mapped
+    def map(self, transform: Callable[[P], O]):
+        return Parser(MapTransform(self.inner, transform))
 
-    # def skip_until(self, other: "Parser[T]"):
-    #     def inner(stream: str, index: int) -> Result[tuple[P, str, T]]:
-    #         pass
-    #     return Parser(inner)
+    def set(self, value: T) -> "Parser[T]":
+        return Parser(ValueTransform(self.inner, value))
+
+    def sep_by(self, sep: "Parser[T]"):
+        return Parser(SepListTransform(self.inner, sep.inner))
 
     # |
     def __or__(self, other: "Parser[T]"):
-        return self.or_else(other)
+        return self.or_(other)
 
     # &
     def __and__(self, other: "Parser[T]"):
-        return self.and_then(other)
+        return self.and_(other)
 
     # >>
     def __rshift__(self, other: "Parser[T]"):
-        return self.and_then(other).map(lambda v: v[1])
+        return self.and_(other).map(lambda v: v[1])
 
     # <<
     def __lshift__(self, other: "Parser[T]"):
-        return self.and_then(other).map(lambda v: v[0])
-
-P = TypeVar("P")
-class Transformer(Generic[P]):
-    def parse(self, stream: str, at_index: int) -> Result[P]:
-        raise Exception("Abstract method.")
-
-L = TypeVar("L")
-R = TypeVar("R")
-class AndParser(Transformer[tuple[L, R]]):
-    def __init__(self, left: Transformer[L], right: Transformer[R]):
-        self.left = left
-        self.right = right
-
-    def parse(self, stream: str, at_index: int) -> Result[tuple[L, R]]:
-        result_left = self.left.parse(stream, at_index)
-        if result_left.actual
+        return self.and_(other).map(lambda v: v[0])

src/pyalibert/py.typed

@@ -0,0 +1 @@
+# Marker file for PEP 561.  The mypy package uses inline types.

src/pyalibert/result.py

@@ -1,8 +1,11 @@
 from dataclasses import dataclass
 from typing import Union, Generic, TypeVar
 
+from .utils import TypeInfo
+
 
 P = TypeVar("P")
+T = TypeVar("T")
 @dataclass
 class Success(Generic[P]):
     value: P
@@ -13,7 +16,7 @@ class Success(Generic[P]):
 class Failure:
     at_index: int
     expected: set[str]
-    depth: int | None
+    depth: int | None = None
 
 
 P = TypeVar("P")
@@ -33,5 +36,5 @@ class ParseError(BaseException):
 Parsing failed at position {self.failure.at_index} of stream :
 ${fail_section}
 Expected one of: {failure.expected}
-"""
+""".strip()
         super().__init__(message)

src/pyalibert/transformer.py

@@ -0,0 +1,11 @@
+from abc import abstractmethod
+from typing import Generic, TypeVar
+
+from .result import Result
+
+
+P = TypeVar("P")
+class Transformer(Generic[P]):
+    @abstractmethod
+    def parse(self, stream: str, at_index: int) -> Result[P]:
+        raise Exception("Abstract method.")

src/pyalibert/transformers/__init__.py

@@ -0,0 +1,13 @@
+
+from .end import EndTransform
+from .just import JustTransform
+from .regex import RegexTransform
+
+from .and_ import AndTransform
+from .or_ import OrTransform
+from .list import ListTransform
+from .sep_list import SepListTransform
+from .option import OptionTransform
+
+from .map import MapTransform
+from .value import ValueTransform

src/pyalibert/transformers/and_.py

@@ -0,0 +1,23 @@
+from dataclasses import dataclass
+from typing import TypeVar
+
+from ..result import Result, Success, Failure
+from ..transformer import Transformer
+
+
+L = TypeVar("L")
+R = TypeVar("R")
+@dataclass
+class AndTransform(Transformer[tuple[L, R]]):
+    def __init__(self, left: Transformer[L], right: Transformer[R]):
+        self.left = left
+        self.right = right
+
+    def parse(self, stream: str, at_index: int) -> Result[tuple[L, R]]:
+        result_left = self.left.parse(stream, at_index)
+        if isinstance(result_left, Failure):
+            return result_left
+        result_right = self.right.parse(stream, result_left.next_index)
+        if isinstance(result_right, Failure):
+            return result_right
+        return Success((result_left.value, result_right.value), result_right.next_index)

src/pyalibert/transformers/end.py

@@ -0,0 +1,15 @@
+from dataclasses import dataclass
+
+from ..result import Result, Success, Failure
+from ..transformer import Transformer
+
+
+@dataclass
+class EndTransform(Transformer[None]):
+    def __init__(self) -> None:
+        pass
+    
+    def parse(self, stream: str, at_index: int) -> Result[None]:
+        if len(stream) == at_index:
+            return Success(None, at_index)
+        return Failure(at_index, set("<end>"))

src/pyalibert/transformers/just.py

@@ -0,0 +1,19 @@
+from dataclasses import dataclass
+
+from ..result import Result, Success, Failure
+from ..transformer import Transformer
+
+
+@dataclass
+class JustTransform(Transformer[str]):
+    def __init__(self, word: str):
+        self.word = word
+        self.word_len = len(self.word)
+
+    def parse(self, stream: str, at_index: int) -> Result[str]:
+        end_index = at_index + self.word_len
+        prefix = stream[at_index:end_index]
+        if prefix == self.word:
+            return Success(self.word, end_index)
+        return Failure(at_index, set([self.word]))
+

src/pyalibert/transformers/list.py

@@ -0,0 +1,24 @@
+from dataclasses import dataclass
+from typing import TypeVar
+
+from ..result import Result, Success, Failure
+from ..transformer import Transformer
+
+
+T = TypeVar("T")
+@dataclass
+class ListTransform(Transformer[list[T]]):
+    def __init__(self, item: Transformer[T]):
+        self.item = item
+
+    def parse(self, stream: str, at_index: int) -> Result[list[T]]:
+        values = list[T]()
+        while True:
+            result = self.item.parse(stream, at_index)
+            if isinstance(result, Failure):
+                break
+            values.append(result.value)
+            if result.next_index == at_index:
+                raise Exception("Parsing empty patterns repeatedly.")
+            at_index = result.next_index
+        return Success(values, at_index)

src/pyalibert/transformers/map.py

@@ -0,0 +1,21 @@
+from dataclasses import dataclass
+from typing import Callable, TypeVar
+
+from ..result import Result, Success, Failure
+from ..transformer import Transformer
+
+
+I = TypeVar("I")
+O = TypeVar("O")
+@dataclass
+class MapTransform(Transformer[O]):
+    def __init__(self, value: Transformer[I], transform: Callable[[I], O]):
+        self.value = value
+        self.transform = transform
+
+    def parse(self, stream: str, at_index: int) -> Result[O]:
+        result = self.value.parse(stream, at_index)
+        if isinstance(result, Failure):
+            return result
+        transformed = self.transform(result.value)
+        return Success(transformed, result.next_index)

src/pyalibert/transformers/option.py

@@ -0,0 +1,18 @@
+from dataclasses import dataclass
+from typing import Optional, TypeVar
+
+from ..result import Result, Success
+from ..transformer import Transformer
+
+
+T = TypeVar("T")
+@dataclass
+class OptionTransform(Transformer[Optional[T]]):
+    def __init__(self, value: Transformer[T]):
+        self.value = value
+    
+    def parse(self, stream: str, at_index: int) -> Result[Optional[T]]:
+        result = self.value.parse(stream, at_index)
+        if isinstance(result, Success):
+            return Success(result.value, result.next_index)
+        return Success(None, at_index)

src/pyalibert/transformers/or_.py

@@ -0,0 +1,23 @@
+from dataclasses import dataclass
+from typing import TypeVar, Union
+
+from ..result import Result, Success, Failure
+from ..transformer import Transformer
+
+
+L = TypeVar("L")
+R = TypeVar("R")
+@dataclass
+class OrTransform(Transformer[Union[L, R]]):
+    def __init__(self, left: Transformer[L], right: Transformer[R]):
+        self.left = left
+        self.right = right
+
+    def parse(self, stream: str, at_index: int) -> Result[Union[L, R]]:
+        result_left = self.left.parse(stream, at_index)
+        if isinstance(result_left, Success):
+            return Success(result_left.value, result_left.next_index)
+        result_right = self.right.parse(stream, at_index)
+        if isinstance(result_right, Success):
+            return Success(result_right.value, result_right.next_index)
+        return Failure(at_index, result_left.expected.union(result_right.expected))

src/pyalibert/transformers/regex.py

@@ -0,0 +1,19 @@
+from dataclasses import dataclass
+import re
+
+from ..result import Result, Success, Failure
+from ..transformer import Transformer
+
+
+@dataclass
+class RegexTransform(Transformer[str]):
+    def __init__(self, pattern: str):
+        self.pattern = pattern
+
+    def parse(self, stream: str, at_index: int) -> Result[str]:
+        do_match = re.match(self.pattern, stream[at_index:])
+        if do_match is None:
+            return Failure(at_index, set(f"<matching '{self.pattern}'>"))
+        else:
+            match = do_match[0]
+            return Success(match, at_index + len(match))

src/pyalibert/transformers/sep_list.py

@@ -0,0 +1,35 @@
+from dataclasses import dataclass
+from typing import Optional, TypeVar
+
+from ..result import Result
+from ..transformer import Transformer
+
+from .map import MapTransform
+from .option import OptionTransform
+from .and_ import AndTransform
+from .list import ListTransform
+
+
+T = TypeVar("T")
+S = TypeVar("S")
+@dataclass
+class SepListTransform(Transformer[list[T]]):
+    def __init__(self, items: Transformer[T], sep: Transformer[S]):
+        self.items = items
+        self.sep = sep
+        first_transformer = OptionTransform(items)
+        rest_transformer = ListTransform(AndTransform(sep, items))
+        total_transformer = AndTransform(first_transformer, rest_transformer)
+        self.actual = MapTransform(total_transformer,SepListTransform.mapping)
+
+    def parse(self, stream: str, at_index: int) -> Result[list[T]]:
+        return self.actual.parse(stream, at_index)
+
+    @staticmethod
+    def mapping(value: tuple[Optional[T], list[tuple[S, T]]]):
+        items = list[T]()
+        if value[0] is not None:
+            items.append(value[0])
+        for (_sep, item) in value[1]:
+            items.append(item)
+        return items

src/pyalibert/transformers/value.py

@@ -0,0 +1,18 @@
+from dataclasses import dataclass
+from typing import TypeVar
+
+from ..result import Result
+from ..transformer import Transformer
+
+from .map import MapTransform
+
+
+I = TypeVar("I")
+T = TypeVar("T")
+@dataclass
+class ValueTransform(Transformer[T]):
+    def __init__(self, ignored: Transformer[I], value: T):
+        self.actual = MapTransform(ignored, lambda _: value)
+
+    def parse(self, stream: str, at_index: int) -> Result[T]:
+        return self.actual.parse(stream, at_index)

src/pyalibert/utils.py

@@ -0,0 +1,5 @@
+from typing import Generic, TypeVar
+
+T = TypeVar("T")
+class TypeInfo(Generic[T]):
+    pass