@scemoon/cdh 1.0.0

package/tui/ansi/_stream_parser.py

@@ -0,0 +1,418 @@
+from functools import lru_cache
+import io
+import re
+
+import rich.repr
+
+from textual.cache import LRUCache
+from typing import Callable, Generator, Iterable
+
+type TokenMatch = tuple[str, str]
+
+type ParseResult[ParseType] = Generator[StreamRead | ParseType, Token, None]
+type PatternCheck = Generator[None, str, TokenMatch | bool | None]
+
+
+@rich.repr.auto
+class Pattern[ValueType]:
+    __slots__ = ["_send", "value"]
+
+    def __init__(self) -> None:
+        self._send: Callable[[str], None] | None = None
+        self.value: ValueType | None = None
+
+    def feed(self, character: str) -> bool | TokenMatch | None:
+        if self._send is None:
+            generator = self.check()
+            self._send = generator.send
+            next(generator)
+        try:
+            self._send(character)
+        except StopIteration as stop_iteration:
+            return stop_iteration.value
+        else:
+            return None
+
+    def check(self) -> PatternCheck:
+        return False
+        yield
+
+
+class StreamRead[ResultType]:
+    pass
+
+
+@rich.repr.auto
+class Read[ResultType](StreamRead[ResultType]):
+    __slots__ = ["remaining"]
+
+    def __init__(self, count: int) -> None:
+        self.remaining = count
+
+
+@rich.repr.auto
+class ReadUntil[ResultType](StreamRead[ResultType]):
+    __slots__ = ["characters", "_regex"]
+
+    def __init__(self, *characters: str) -> None:
+        self.characters = characters
+        self._regex = re.compile(
+            "|".join(re.escape(character) for character in characters)
+        )
+
+    def __rich_repr__(self) -> rich.repr.Result:
+        yield from self.characters
+
+
+@rich.repr.auto
+class ReadRegex[ResultType](StreamRead[ResultType]):
+    __slots__ = ["regex", "max_length", "_buffer"]
+
+    def __init__(self, regex: str, max_length: int | None = None) -> None:
+        self.regex = regex
+        self.max_length = max_length
+        self._buffer = io.StringIO()
+
+    def __rich_repr__(self) -> rich.repr.Result:
+        yield self.regex
+
+    @property
+    def buffer_size(self) -> int:
+        return self._buffer.tell()
+
+
+@rich.repr.auto
+class ReadPatterns[ResultType](StreamRead[ResultType]):
+    __slots__ = ["patterns", "_text"]
+
+    def __init__(self, start: str = "", **patterns: Pattern) -> None:
+        self.patterns = patterns
+        self._text = io.StringIO()
+        self._text.write(start)
+
+    @property
+    def unconsumed_text(self) -> str:
+        return self._text.getvalue()
+
+    def __rich_repr__(self) -> rich.repr.Result:
+        for key, value in self.patterns.items():
+            yield key, value
+
+    @property
+    def is_exhausted(self) -> bool:
+        return not self.patterns
+
+    def feed(self, text: str) -> tuple[int, TokenMatch | None]:
+        consumed = 0
+        new_patterns = patterns = self.patterns
+        for character in text:
+            consumed += 1
+            for name, sequence_validator in patterns.items():
+                if (value := sequence_validator.feed(character)) is False:
+                    new_patterns = patterns.copy()
+                    new_patterns.pop(name)
+                elif value:
+                    return consumed, (name, value)
+            patterns = self._patterns = new_patterns
+        self._text.write(text[:consumed])
+        return consumed, None
+
+
+@rich.repr.auto
+class ReadPattern[ResultType](StreamRead[ResultType]):
+    """Special case for a single pattern."""
+
+    __slots__ = ["name", "pattern", "_text", "_exhaused"]
+
+    def __init__(self, start: str, name: str, pattern: Pattern) -> None:
+        self.name = name
+        self.pattern: Pattern = pattern
+        self._text = io.StringIO()
+        self._text.write(start)
+        self._exhaused = False
+
+    @property
+    def unconsumed_text(self) -> str:
+        return self._text.getvalue()
+
+    def __rich_repr__(self) -> rich.repr.Result:
+        yield self.name
+        yield self.pattern
+
+    @property
+    def is_exhausted(self) -> bool:
+        return self._exhaused
+
+    def feed(self, text: str) -> tuple[int, TokenMatch | None]:
+        consumed = 0
+        feed = self.pattern.feed
+        for character in text:
+            consumed += 1
+            if (value := feed(character)) is False:
+                self._exhaused = True
+                break
+            elif value:
+                self._exhaused = True
+                return consumed, ("pattern", value)
+        self._text.write(text[:consumed])
+        return consumed, None
+
+
+@rich.repr.auto
+class Token:
+    """A token containing text."""
+
+    __slots__ = "text"
+
+    def __init__(self, text: str = "") -> None:
+        self.text = text
+
+    def __rich_repr__(self) -> rich.repr.Result:
+        yield self.text
+
+    def __str__(self) -> str:
+        return self.text
+
+
+class SeparatorToken(Token):
+    pass
+
+
+class MatchToken(Token):
+    __slots__ = ["match"]
+
+    def __init__(self, text: str, match: re.Match) -> None:
+        self.match = match
+        super().__init__(text)
+
+    def __rich_repr__(self) -> rich.repr.Result:
+        yield self.match
+
+
+class EOFToken(Token):
+    pass
+
+
+class PatternToken(Token):
+    __slots__ = ["name", "value"]
+
+    def __init__(self, name: str, value: TokenMatch) -> None:
+        self.name = name
+        self.value = value
+        super().__init__("")
+
+    def __rich_repr__(self) -> rich.repr.Result:
+        yield self.name
+        yield None, self.value
+
+
+class StreamParser[ParseType]:
+    """Parses a stream of text into tokens."""
+
+    def __init__(self):
+        self._gen = self.parse()
+        self._reading: StreamRead | ParseType = next(self._gen)
+        self._cache = LRUCache(1024 * 4)
+
+    def read(self, count: int) -> Read:
+        """Read a specific number of bytes.
+
+        Args:
+            count: Number of bytes to read.
+        """
+        return Read(count)
+
+    @lru_cache(1024)
+    def read_until(self, *characters: str) -> ReadUntil:
+        """Read until the given characters.
+
+        Args:
+            characters: Set of characters to stop read.
+
+        """
+        return ReadUntil(*characters)
+
+    def read_regex(self, regex: str) -> ReadRegex:
+        """Search for the matching regex.
+
+        Args:
+            regex: Regular expression.
+        """
+        return ReadRegex(regex)
+
+    def read_patterns(self, start: str = "", **patterns) -> ReadPattern | ReadPatterns:
+        """Read until a pattern matches, or the patterns have been exhausted.
+
+        Args:
+            start: Initial part of the string.
+            **patterns: One or more patterns.
+        """
+        if len(patterns) == 1:
+            name, pattern = patterns.popitem()
+            return ReadPattern(start, name, pattern)
+        return ReadPatterns(start, **patterns)
+
+    def feed(self, text: str) -> Iterable[Token | ParseType]:
+        sequences = text.splitlines(keepends=True)
+        # TODO: Cache
+        for sequence in sequences:
+            yield from self._feed(sequence)
+
+    def _feed(self, text: str) -> Iterable[Token | ParseType]:
+        """Feed text in to parser.
+
+        Args:
+            text: Text from stream.
+
+        Returns:
+            A generator of tokens or the parse type.
+
+        """
+        if not text or self._gen is None:
+            yield EOFToken()
+            return
+
+        def send(token: Token) -> Iterable[Token]:
+            try:
+                while True:
+                    new_token = self._gen.send(token)
+                    if isinstance(new_token, StreamRead):
+                        self._reading = new_token
+                        break
+                    else:
+                        token = new_token
+                        yield token
+
+            except StopIteration:
+                self._gen.close()
+                self._gen = None
+
+        while text:
+            if isinstance(self._reading, (ReadPattern, ReadPatterns)):
+                consumed, pattern_match = self._reading.feed(text)
+
+                if pattern_match is not None:
+                    name, value = pattern_match
+                    yield from send(PatternToken(name, value))
+                    text = text[consumed:]
+                else:
+                    if self._reading.is_exhausted:
+                        unconsumed_text = self._reading.unconsumed_text
+                        yield from send(Token(unconsumed_text))
+                        text = text[consumed:]
+                    else:
+                        text = ""
+
+            elif isinstance(self._reading, Read):
+                if self._reading.remaining:
+                    read_text = text[: self._reading.remaining]
+                    read_text_length = len(read_text)
+                    self._reading.remaining -= read_text_length
+                    text = text[read_text_length:]
+                    yield from send(Token(read_text))
+                else:
+                    yield from send(Token(""))
+
+            elif isinstance(self._reading, ReadUntil):
+                if (match := self._reading._regex.search(text)) is not None:
+                    start, end = match.span(0)
+                    read_text = text[:start]
+
+                    if read_text:
+                        yield from send(Token(read_text))
+                        text = text[start:]
+                    else:
+                        yield from send(SeparatorToken(text[start:end]))
+                        text = text[end:]
+                else:
+                    yield from send(Token(text))
+                    text = ""
+
+            elif isinstance(self._reading, ReadRegex):
+                self._reading._buffer.write(text)
+                match_text = self._reading._buffer.getvalue()
+                if (
+                    match := re.search(self._reading.regex, match_text, re.VERBOSE)
+                ) is not None:
+                    token_text = match_text[: match.start(0)]
+                    if token_text:
+                        yield from send(Token(token_text))
+                    end = match.end(0)
+                    yield from send(MatchToken(match.group(0), match))
+                    text = text[end:]
+                else:
+                    yield from send(Token(match_text))
+                    text = ""
+
+    def parse(self) -> ParseResult[ParseType]:
+        yield from ()
+
+
+if __name__ == "__main__":
+    # from rich import print
+
+    import string
+
+    class KeyValue(Pattern):
+        def check(self) -> PatternCheck:
+            """Parses text in the form key:'value'
+
+            e.g
+
+            """
+            key: str = ""
+            value: str = ""
+            is_letter = string.ascii_lowercase.__contains__
+            if not is_letter(character := (yield)):
+                return False
+            key += character
+            while is_letter(character := (yield)):
+                key += character
+            if character != ":":
+                return False
+            if (yield) != "'":
+                return False
+            while is_letter(character := (yield)):
+                value += character
+            if character != "'":
+                return False
+            self.value = (key, value)
+            return True
+
+    class TestParser(StreamParser):
+        def parse(self) -> ParseResult:
+            token = yield self.read_patterns(key_value=KeyValue())
+            print("!", repr(token))
+            yield token
+            while token := (yield self.read(1)):
+                print(repr(token))
+            # while True:
+            #     token = yield self.read_until(":")
+            #     if not token:
+            #         break
+            #     yield token
+            #     if isinstance(token, SeparatorToken):
+            #         break
+            #     key += token.text
+
+            # while token := (yield self.read_regex(r"\'.*?\'")):
+            #     yield token
+            #     if isinstance(token, MatchToken):
+            #         break
+
+            # string = yield self.read_regex("'.*?'")
+            # print("VALUE=", string)
+
+            # yield (yield self.read(3))
+            # while (text := (yield self.read_until("'"))) != "'":
+            #     yield text
+            # yield text
+            # while (text := (yield self.read_until("'"))) != "'":
+            #     yield text
+            # yield text
+
+    parser = TestParser()
+
+    for chunk in ["foo", ":", "'bar", "asd", "asdasd", "';"]:
+        for token in parser.feed(chunk):
+            print(repr(token))

package/tui/answer.py

@@ -0,0 +1,12 @@
+from typing import NamedTuple
+
+
+class Answer(NamedTuple):
+    """An answer to a question posed by the agent."""
+
+    text: str
+    """The textual response."""
+    id: str
+    """The id of the response."""
+    kind: str | None = None
+    """Optional hint: permission kind or keyboard shortcut key for UI display."""