glitchlings 0.1.1__py3-none-any.whl → 0.1.3__py3-none-any.whl

{zoo → glitchlings/zoo}/redactyl.py

@@ -1,73 +1,128 @@
-import re
-import random
-from .core import Glitchling, AttackWave
-
-FULL_BLOCK = "█"
-
-
-def redact_words(
-    text: str,
-    replacement_char: str = FULL_BLOCK,
-    redaction_rate: float = 0.05,
-    merge_adjacent: bool = False,
-    seed: int = 151,
-    rng: random.Random | None = None,
-) -> str:
-    """Redact random words by replacing their characters.
-
-    Parameters
-    - text: Input text.
-    - replacement_char: The character to use for redaction (default FULL_BLOCK).
-    - redaction_rate: Max proportion of words to redact (default 0.05).
-    - merge_adjacent: If True, merges adjacent redactions across intervening non-word chars.
-    - seed: Seed used if `rng` not provided (default 151).
-    - rng: Optional RNG; overrides seed.
-    """
-    if rng is None:
-        rng = random.Random(seed)
-
-    # Preserve exact spacing and punctuation by using regex
-    tokens = re.split(r"(\s+)", text)
-    word_indices = [i for i, token in enumerate(tokens) if i % 2 == 0 and token.strip()]
-    num_to_redact = max(1, int(len(word_indices) * redaction_rate))
-
-    # Sample from the indices of actual words
-    indices_to_redact = rng.sample(word_indices, k=num_to_redact)
-    indices_to_redact.sort()
-
-    for i in indices_to_redact:
-        if i >= len(tokens):
-            break
-
-        word = tokens[i]
-        if not word or word.isspace():  # Skip empty or whitespace
-            continue
-
-        # Check if word has trailing punctuation
-        match = re.match(r"^(\W*)(.*?)(\W*)$", word)
-        if match:
-            prefix, core, suffix = match.groups()
-            tokens[i] = f"{prefix}{replacement_char * len(core)}{suffix}"
-        else:
-            tokens[i] = f"{replacement_char * len(word)}"
-
-    text = "".join(tokens)
-
-    if merge_adjacent:
-        text = re.sub(
-            rf"{replacement_char}\W+{replacement_char}",
-            lambda m: replacement_char * (len(m.group(0)) - 1),
-            text,
-        )
-
-    return text
-
-
-redactyl = Glitchling(
-    name="Redactyl",
-    corruption_function=redact_words,
-    replacement_char=FULL_BLOCK,
-    redaction_rate=0.05,
-    scope=AttackWave.WORD,
-    seed=151,
-)
+import re
+import random
+
+from .core import Glitchling, AttackWave
+
+FULL_BLOCK = "█"
+
+
+try:
+    from glitchlings._zoo_rust import redact_words as _redact_words_rust
+except ImportError:  # pragma: no cover - compiled extension not present
+    _redact_words_rust = None
+
+
+def _python_redact_words(
+    text: str,
+    *,
+    replacement_char: str,
+    redaction_rate: float,
+    merge_adjacent: bool,
+    rng: random.Random,
+) -> str:
+    """Redact random words by replacing their characters.
+
+    Parameters
+    - text: Input text.
+    - replacement_char: The character to use for redaction (default FULL_BLOCK).
+    - redaction_rate: Max proportion of words to redact (default 0.05).
+    - merge_adjacent: If True, merges adjacent redactions across intervening non-word chars.
+    - seed: Seed used if `rng` not provided (default 151).
+    - rng: Optional RNG; overrides seed.
+    """
+    # Preserve exact spacing and punctuation by using regex
+    tokens = re.split(r"(\s+)", text)
+    word_indices = [i for i, token in enumerate(tokens) if i % 2 == 0 and token.strip()]
+    if not word_indices:
+        raise ValueError("Cannot redact words because the input text contains no redactable words.")
+    num_to_redact = max(1, int(len(word_indices) * redaction_rate))
+
+    # Sample from the indices of actual words
+    indices_to_redact = rng.sample(word_indices, k=num_to_redact)
+    indices_to_redact.sort()
+
+    for i in indices_to_redact:
+        if i >= len(tokens):
+            break
+
+        word = tokens[i]
+        if not word or word.isspace():  # Skip empty or whitespace
+            continue
+
+        # Check if word has trailing punctuation
+        match = re.match(r"^(\W*)(.*?)(\W*)$", word)
+        if match:
+            prefix, core, suffix = match.groups()
+            tokens[i] = f"{prefix}{replacement_char * len(core)}{suffix}"
+        else:
+            tokens[i] = f"{replacement_char * len(word)}"
+
+    text = "".join(tokens)
+
+    if merge_adjacent:
+        text = re.sub(
+            rf"{replacement_char}\W+{replacement_char}",
+            lambda m: replacement_char * (len(m.group(0)) - 1),
+            text,
+        )
+
+    return text
+
+
+def redact_words(
+    text: str,
+    replacement_char: str = FULL_BLOCK,
+    redaction_rate: float = 0.05,
+    merge_adjacent: bool = False,
+    seed: int = 151,
+    rng: random.Random | None = None,
+) -> str:
+    """Redact random words by replacing their characters."""
+
+    if rng is None:
+        rng = random.Random(seed)
+
+    if _redact_words_rust is not None:
+        return _redact_words_rust(
+            text,
+            replacement_char,
+            redaction_rate,
+            merge_adjacent,
+            rng,
+        )
+
+    return _python_redact_words(
+        text,
+        replacement_char=replacement_char,
+        redaction_rate=redaction_rate,
+        merge_adjacent=merge_adjacent,
+        rng=rng,
+    )
+
+
+class Redactyl(Glitchling):
+    """Glitchling that redacts words with block characters."""
+
+    def __init__(
+        self,
+        *,
+        replacement_char: str = FULL_BLOCK,
+        redaction_rate: float = 0.05,
+        merge_adjacent: bool = False,
+        seed: int = 151,
+    ) -> None:
+        super().__init__(
+            name="Redactyl",
+            corruption_function=redact_words,
+            scope=AttackWave.WORD,
+            seed=seed,
+            replacement_char=replacement_char,
+            redaction_rate=redaction_rate,
+            merge_adjacent=merge_adjacent,
+        )
+
+
+redactyl = Redactyl()
+
+
+__all__ = ["Redactyl", "redactyl"]

{zoo → glitchlings/zoo}/reduple.py

@@ -1,54 +1,100 @@
-import re
-import random
-from .core import Glitchling, AttackWave
-
-
-def reduplicate_words(
-    text: str,
-    reduplication_rate: float = 0.05,
-    seed: int | None = None,
-    rng: random.Random | None = None,
-) -> str:
-    """Randomly reduplicate words in the text.
-
-    Parameters
-    - text: Input text.
-    - reduplication_rate: Max proportion of words to reduplicate (default 0.05).
-    - seed: Optional seed if `rng` not provided.
-    - rng: Optional RNG; overrides seed.
-
-    Notes
-    - Preserves spacing and punctuation by tokenizing with separators.
-    - Deterministic when run with a fixed seed or via Gaggle.
-    """
-    if rng is None:
-        rng = random.Random(seed)
-
-    # Preserve exact spacing and punctuation by using regex
-    tokens = re.split(r"(\s+)", text)  # Split but keep separators
-
-    for i in range(0, len(tokens), 2):  # Every other token is a word
-        if i >= len(tokens):
-            break
-
-        word = tokens[i]
-        if not word or word.isspace():  # Skip empty or whitespace
-            continue
-
-        # Only consider actual words for reduplication
-        if rng.random() < reduplication_rate:
-            # Check if word has trailing punctuation
-            match = re.match(r"^(\W*)(.*?)(\W*)$", word)
-            if match:
-                prefix, core, suffix = match.groups()
-                # Reduplicate with a space: "word" -> "word word"
-                tokens[i] = f"{prefix}{core} {core}{suffix}"
-            else:
-                tokens[i] = f"{word} {word}"
-
-    return "".join(tokens)
-
-
-reduple = Glitchling(
-    name="Reduple", corruption_function=reduplicate_words, scope=AttackWave.WORD
-)
+import re
+import random
+
+from .core import Glitchling, AttackWave
+
+try:
+    from glitchlings._zoo_rust import reduplicate_words as _reduplicate_words_rust
+except ImportError:  # pragma: no cover - compiled extension not present
+    _reduplicate_words_rust = None
+
+
+def _python_reduplicate_words(
+    text: str,
+    *,
+    reduplication_rate: float,
+    rng: random.Random,
+) -> str:
+    """Randomly reduplicate words in the text.
+
+    Parameters
+    - text: Input text.
+    - reduplication_rate: Max proportion of words to reduplicate (default 0.05).
+    - seed: Optional seed if `rng` not provided.
+    - rng: Optional RNG; overrides seed.
+
+    Notes
+    - Preserves spacing and punctuation by tokenizing with separators.
+    - Deterministic when run with a fixed seed or via Gaggle.
+    """
+    # Preserve exact spacing and punctuation by using regex
+    tokens = re.split(r"(\s+)", text)  # Split but keep separators
+
+    for i in range(0, len(tokens), 2):  # Every other token is a word
+        if i >= len(tokens):
+            break
+
+        word = tokens[i]
+        if not word or word.isspace():  # Skip empty or whitespace
+            continue
+
+        # Only consider actual words for reduplication
+        if rng.random() < reduplication_rate:
+            # Check if word has trailing punctuation
+            match = re.match(r"^(\W*)(.*?)(\W*)$", word)
+            if match:
+                prefix, core, suffix = match.groups()
+                # Reduplicate with a space: "word" -> "word word"
+                tokens[i] = f"{prefix}{core} {core}{suffix}"
+            else:
+                tokens[i] = f"{word} {word}"
+    return "".join(tokens)
+
+
+def reduplicate_words(
+    text: str,
+    reduplication_rate: float = 0.05,
+    seed: int | None = None,
+    rng: random.Random | None = None,
+) -> str:
+    """Randomly reduplicate words in the text.
+
+    Falls back to the Python implementation when the optional Rust
+    extension is unavailable.
+    """
+
+    if rng is None:
+        rng = random.Random(seed)
+
+    if _reduplicate_words_rust is not None:
+        return _reduplicate_words_rust(text, reduplication_rate, rng)
+
+    return _python_reduplicate_words(
+        text,
+        reduplication_rate=reduplication_rate,
+        rng=rng,
+    )
+
+
+class Reduple(Glitchling):
+    """Glitchling that repeats words to simulate stuttering speech."""
+
+    def __init__(
+        self,
+        *,
+        reduplication_rate: float = 0.05,
+        seed: int | None = None,
+    ) -> None:
+        super().__init__(
+            name="Reduple",
+            corruption_function=reduplicate_words,
+            scope=AttackWave.WORD,
+            seed=seed,
+            reduplication_rate=reduplication_rate,
+        )
+
+
+reduple = Reduple()
+
+
+__all__ = ["Reduple", "reduple"]

glitchlings/zoo/rushmore.py

@@ -0,0 +1,97 @@
+import math
+import random
+import re
+
+from .core import Glitchling, AttackWave
+
+try:
+    from glitchlings._zoo_rust import delete_random_words as _delete_random_words_rust
+except ImportError:  # pragma: no cover - compiled extension not present
+    _delete_random_words_rust = None
+
+
+def _python_delete_random_words(
+    text: str,
+    *,
+    max_deletion_rate: float,
+    rng: random.Random,
+) -> str:
+    """Delete random words from the input text while preserving whitespace."""
+
+    tokens = re.split(r"(\s+)", text)  # Split but keep separators for later rejoin
+
+    candidate_indices: list[int] = []
+    for i in range(2, len(tokens), 2):  # Every other token is a word, skip the first word
+        word = tokens[i]
+        if not word or word.isspace():
+            continue
+
+        candidate_indices.append(i)
+
+    allowed_deletions = math.floor(len(candidate_indices) * max_deletion_rate)
+    if allowed_deletions <= 0:
+        return text
+
+    for i in candidate_indices:
+        if rng.random() < max_deletion_rate:
+            word = tokens[i]
+            match = re.match(r"^(\W*)(.*?)(\W*)$", word)
+            if match:
+                prefix, _, suffix = match.groups()
+                tokens[i] = f"{prefix.strip()}{suffix.strip()}"
+            else:
+                tokens[i] = ""
+
+    text = "".join(tokens)
+    text = re.sub(r"\s+([.,;:])", r"\1", text)
+    text = re.sub(r"\s{2,}", " ", text).strip()
+
+    return text
+
+
+def delete_random_words(
+    text: str,
+    max_deletion_rate: float = 0.01,
+    seed: int | None = None,
+    rng: random.Random | None = None,
+) -> str:
+    """Delete random words from the input text.
+
+    Uses the optional Rust implementation when available.
+    """
+
+    if rng is None:
+        rng = random.Random(seed)
+
+    if _delete_random_words_rust is not None:
+        return _delete_random_words_rust(text, max_deletion_rate, rng)
+
+    return _python_delete_random_words(
+        text,
+        max_deletion_rate=max_deletion_rate,
+        rng=rng,
+    )
+
+
+class Rushmore(Glitchling):
+    """Glitchling that deletes words to simulate missing information."""
+
+    def __init__(
+        self,
+        *,
+        max_deletion_rate: float = 0.01,
+        seed: int | None = None,
+    ) -> None:
+        super().__init__(
+            name="Rushmore",
+            corruption_function=delete_random_words,
+            scope=AttackWave.WORD,
+            seed=seed,
+            max_deletion_rate=max_deletion_rate,
+        )
+
+
+rushmore = Rushmore()
+
+
+__all__ = ["Rushmore", "rushmore"]