sonatoki 0.1.3__py3-none-any.whl → 0.1.5__py3-none-any.whl

sonatoki/Configs.py

@@ -9,15 +9,15 @@ from typing_extensions import NotRequired
 from sonatoki.Filters import (
     Filter,
     NimiPu,
-    Numerics,
+    Numeric,
     Syllabic,
     NimiLinku,
     NimiPuAle,
     Alphabetic,
     ProperName,
     Phonotactic,
+    Punctuation,
     NimiLinkuAle,
-    Punctuations,
 )
 from sonatoki.Scorers import Number, Scorer, PassFail, SoftScaling, SoftPassFail
 from sonatoki.Cleaners import Cleaner, ConsecutiveDuplicates
@@ -45,7 +45,7 @@ class IloConfig(TypedDict):
 BaseConfig: IloConfig = {
     "preprocessors": [URLs],
     "cleaners": [ConsecutiveDuplicates],
-    "ignoring_filters": [Numerics, Punctuations],
+    "ignoring_filters": [Numeric, Punctuation],
     "scoring_filters": [],
     "scorer": PassFail,
     "passing_score": 0.8,

sonatoki/Filters.py

@@ -1,10 +1,11 @@
 # STL
+import re
 from abc import ABC, abstractmethod
 from typing import Set
 from functools import lru_cache as cache  # cache comes in 3.9
 
 # PDM
-import regex as re
+import regex
 from typing_extensions import override
 
 # LOCAL
@@ -13,14 +14,16 @@ from sonatoki.constants import (
     CONSONANTS,
     NIMI_PU_SET,
     ALPHABET_SET,
+    UNICODE_PUNCT,
     ALLOWABLES_SET,
     NIMI_LINKU_SET,
     NIMI_PU_ALE_SET,
     NIMI_LINKU_ALE_SET,
+    PRUNED_POSIX_PUNCT,
     NIMI_LINKU_SANDBOX_SET,
 )
 
-re.DEFAULT_VERSION = re.VERSION1
+regex.DEFAULT_VERSION = regex.VERSION1
 
 
 class Filter(ABC):
@@ -41,6 +44,16 @@ class RegexFilter(Filter):
         return not not re.fullmatch(cls.pattern, token)
 
 
+class Regex1Filter(Filter):
+    pattern: "regex.Pattern[str]"
+
+    @classmethod
+    @override
+    @cache(maxsize=None)
+    def filter(cls, token: str) -> bool:
+        return not not regex.fullmatch(cls.pattern, token)
+
+
 class SetFilter(Filter):
     tokens: Set[str]
 
@@ -131,7 +144,7 @@ class Alphabetic(Filter):
         return set(token.lower()).issubset(ALPHABET_SET)
 
 
-class Numerics(Filter):
+class Numeric(Filter):
     """Determine if a given token is entirely numeric.
     Covers all numeric symbols in Unicode.
 
@@ -147,8 +160,8 @@ class Numerics(Filter):
         return msg.isnumeric()
 
 
-class Punctuations(RegexFilter):
-    pattern = re.compile(r"[\p{Punctuation}\p{posix_punct}]+")
+class Punctuation(RegexFilter):
+    pattern = re.compile(rf"[{PRUNED_POSIX_PUNCT}{UNICODE_PUNCT}]+")
 
 
 __all__ = [
@@ -159,6 +172,6 @@ __all__ = [
     "Syllabic",
     "Alphabetic",
     "ProperName",
-    "Punctuations",
-    "Numerics",
+    "Punctuation",
+    "Numeric",
 ]

sonatoki/Preprocessors.py

@@ -17,13 +17,14 @@ It is up to the user to order them appropriately.
 """
 
 # STL
+import re
 from abc import ABC, abstractmethod
 
 # PDM
-import regex as re
+import regex
 from typing_extensions import override
 
-re.DEFAULT_VERSION = re.VERSION1
+regex.DEFAULT_VERSION = regex.VERSION1
 
 
 class Preprocessor(ABC):
@@ -43,6 +44,16 @@ class RegexPreprocessor(Preprocessor):
         return re.sub(cls.pattern, cls.replace, msg)
 
 
+class Regex1Preprocessor(Preprocessor):
+    pattern: "regex.Pattern[str]"
+    replace: str = " "
+
+    @classmethod
+    @override
+    def process(cls, msg: str) -> str:
+        return regex.sub(cls.pattern, cls.replace, msg)
+
+
 """
 The following classes are Ignorables.
 
@@ -62,6 +73,13 @@ class URLs(RegexPreprocessor):
     pattern = re.compile(r"https?:\/\/\S+")
 
 
+class Reference(RegexPreprocessor):
+    """Remove text contained in double brackets.
+    Often used to fetch articles on Wikipedia, or Magic the Gathering cards."""
+
+    pattern = re.compile(r"\[\[.+\]\]")
+
+
 class DiscordEmotes(RegexPreprocessor):
     """Remove text-formatted Discord emotes `<flags:name:id>`"""
 
@@ -80,6 +98,13 @@ class DiscordSpecial(RegexPreprocessor):
     pattern = re.compile(r"<id:[a-zA-Z0-9_]{4,}>")
 
 
+class AngleBracketObject(RegexPreprocessor):
+    """A generalized version of the Discord-specific angle bracket objects.
+    Removes any contiguous (not broken by whitespace) text in angle brackets."""
+
+    pattern = re.compile(r"<[^<>\s]+>")
+
+
 """
 The following classes are Containers.
 
@@ -92,23 +117,23 @@ would likely be using a language other than Toki Pona.
 
 
 class SingleQuotes(RegexPreprocessor):
-    pattern = re.compile(r"'[^']+'", flags=re.S)  # . matches newline
+    pattern = re.compile(r"'[^']+'", flags=re.DOTALL)
 
 
 class DoubleQuotes(RegexPreprocessor):
-    pattern = re.compile(r'"[^"]+"', flags=re.S)
+    pattern = re.compile(r'"[^"]+"', flags=re.DOTALL)
 
 
 class Backticks(RegexPreprocessor):
     """Remove paired backticks and their contents `like this`"""
 
-    pattern = re.compile(r"`[^`]+`", flags=re.S)
+    pattern = re.compile(r"`[^`]+`", flags=re.DOTALL)
 
 
 class Spoilers(RegexPreprocessor):
     """Remove paired double bars and their contents `||like this||`"""
 
-    pattern = re.compile(r"\|\|(?:(?!\|\|).)+\|\|", flags=re.S)
+    pattern = re.compile(r"\|\|(?:(?!\|\|).)+\|\|", flags=re.DOTALL)
 
 
 class ArrowQuote(RegexPreprocessor):
@@ -117,7 +142,22 @@ class ArrowQuote(RegexPreprocessor):
     pattern = re.compile(r"^>\ .+$", re.MULTILINE)
 
 
+class AllQuotes(RegexPreprocessor):
+    pattern = re.compile(
+        "|".join(
+            [
+                SingleQuotes.pattern.pattern,
+                DoubleQuotes.pattern.pattern,
+                Backticks.pattern.pattern,
+                ArrowQuote.pattern.pattern,
+            ]
+        ),
+        flags=re.MULTILINE | re.DOTALL,
+    )
+
+
 __all__ = [
+    "AngleBracketObject",
     "DiscordChannels",
     "DiscordMentions",
     "DiscordSpecial",
@@ -125,7 +165,9 @@ __all__ = [
     "SingleQuotes",
     "DoubleQuotes",
     "ArrowQuote",
+    "AllQuotes",
     "Backticks",
+    "Reference",
     "Spoilers",
     "URLs",
 ]

sonatoki/Scorers.py

@@ -10,8 +10,6 @@ from typing_extensions import override
 # LOCAL
 from sonatoki.Filters import Filter
 
-LOG = logging.getLogger(__name__)
-
 Number = Union[int, float]
 Weights = Dict[str, Number]
 
@@ -37,12 +35,7 @@ class PassFail(Scorer):
     def score_token(cls, token: str, filters: List[Type[Filter]]) -> Number:
         for f in filters:
             if f.filter(token):
-                score = 1
-                LOG.debug(
-                    "%12s.%s('%s') = %.2f", cls.__name__, f.__name__, token, score
-                )
-                return score
-        LOG.debug("%12s('%s') = 0.00", cls.__name__, token)
+                return 1
         return 0
 
     @classmethod
@@ -86,12 +79,7 @@ class Scaling(Scorer):
     def score_token(cls, token: str, filters: List[Type[Filter]], scale: int):
         for i, f in enumerate(filters):
             if f.filter(token):
-                score = scale - i
-                LOG.debug(
-                    "%12s.%s('%s') = %.2f", cls.__name__, f.__name__, token, score
-                )
-                return score
-        LOG.debug("%12s('%s') = 0.00", cls.__name__, token)
+                return scale - i
         return 0
 
     @classmethod

sonatoki/Tokenizers.py

@@ -1,11 +1,15 @@
 # STL
+import re
 from abc import ABC, abstractmethod
 from typing import List
 
 # PDM
-import regex as re
+import regex
 from typing_extensions import override
 
+# LOCAL
+from sonatoki.constants import UNICODE_PUNCT, PRUNED_POSIX_PUNCT
+
 try:
     # PDM
     import nltk
@@ -15,7 +19,7 @@ except ImportError as e:
     nltk = e
 
 
-LANGUAGE = "english"  # for NLTK
+regex.DEFAULT_VERSION = regex.VERSION1
 
 
 class Tokenizer(ABC):
@@ -42,15 +46,26 @@ class RegexTokenizer(Tokenizer):
         return [clean for word in re.split(cls.pattern, s) if (clean := word.strip())]
 
 
+class Regex1Tokenizer(Tokenizer):
+    pattern: "regex.Pattern[str]"
+
+    @classmethod
+    @override
+    def tokenize(cls, s: str) -> List[str]:
+        return [
+            clean for word in regex.split(cls.pattern, s) if (clean := word.strip())
+        ]
+
+
 class WordTokenizerTok(RegexTokenizer):
-    pattern = re.compile(r"""([\p{Punctuation}\p{posix_punct}]+|\s+)""")
-    # TODO: are <> or {} that common as *sentence* delims? [] are already a stretch
-    # TODO: do the typography characters matter?
-    # NOTE: | / and , are *not* sentence delimiters for my purpose
+    pattern = re.compile(rf"""([{PRUNED_POSIX_PUNCT}{UNICODE_PUNCT}]+|\s+)""")
 
 
 class SentTokenizerTok(RegexTokenizer):
-    pattern = re.compile(r"""(?<=[.?!:;·…“”"'()\[\]\-]|$)""")
+    pattern = re.compile(r"""(?<=[.?!:;·…“”"'()\[\]\-])|$""", flags=re.MULTILINE)
+    # TODO: are <> or {} that common as *sentence* delims? [] are already a stretch
+    # TODO: do the typography characters matter?
+    # NOTE: | / and , are *not* sentence delimiters for my purpose
 
 
 class WordTokenizerRe(RegexTokenizer):

sonatoki/constants.py

@@ -11,6 +11,22 @@ CONSONANTS = "jklmnpstw"
 ALPHABET = VOWELS + CONSONANTS
 ALPHABET_SET = set(ALPHABET)
 
+LANGUAGE = "english"  # for NLTK
+
+# `\p{posix_punct}` character class
+POSIX_PUNCT = r"""-!"#$%&'()*+,./:;<=>?@[\]^_`{|}~"""
+PRUNED_POSIX_PUNCT = r"""$+<=>^`|~"""  # only those that are not in UNICODE_PUNCT
+
+# `\p{Punctuation}` character class
+UNICODE_PUNCT = r"""!"#%&'()*,-./:;?@\[\\\]_{}¡§«¶·»¿;·՚՛՜՝՞՟։֊־׀׃׆׳״؉؊،؍؛؝؞؟٪٫٬٭۔܀܁܂܃܄܅܆܇܈܉܊܋܌܍߷߸߹࠰࠱࠲࠳࠴࠵࠶࠷࠸࠹࠺࠻࠼࠽࠾࡞।॥॰৽੶૰౷಄෴๏๚๛༄༅༆༇༈༉༊་༌།༎༏༐༑༒༔༺༻༼༽྅࿐࿑࿒࿓࿔࿙࿚၊။၌၍၎၏჻፠፡።፣፤፥፦፧፨᐀᙮᚛᚜᛫᛬᛭᜵᜶។៕៖៘៙៚᠀᠁᠂᠃᠄᠅᠆᠇᠈᠉᠊᥄᥅᨞᨟᪠᪡᪢᪣᪤᪥᪦᪨᪩᪪᪫᪬᪭᭚᭛᭜᭝᭞᭟᭠᭽᭾᯼᯽᯾᯿᰻᰼᰽᰾᰿᱾᱿᳀᳁᳂᳃᳄᳅᳆᳇᳓‐‑‒–—―‖‗‘’‚‛“”„‟†‡•‣․‥…‧‰‱′″‴‵‶‷‸‹›※‼‽‾‿⁀⁁⁂⁃⁅⁆⁇⁈⁉⁊⁋⁌⁍⁎⁏⁐⁑⁓⁔⁕⁖⁗⁘⁙⁚⁛⁜⁝⁞⁽⁾₍₎⌈⌉⌊⌋〈〉❨❩❪❫❬❭❮❯❰❱❲❳❴❵⟅⟆⟦⟧⟨⟩⟪⟫⟬⟭⟮⟯⦃⦄⦅⦆⦇⦈⦉⦊⦋⦌⦍⦎⦏⦐⦑⦒⦓⦔⦕⦖⦗⦘⧘⧙⧚⧛⧼⧽⳹⳺⳻⳼⳾⳿⵰⸀⸁⸂⸃⸄⸅⸆⸇⸈⸉⸊⸋⸌⸍⸎⸏⸐⸑⸒⸓⸔⸕⸖⸗⸘⸙⸚⸛⸜⸝⸞⸟⸠⸡⸢⸣⸤⸥⸦⸧⸨⸩⸪⸫⸬⸭⸮⸰⸱⸲⸳⸴⸵⸶⸷⸸⸹⸺⸻⸼⸽⸾⸿⹀⹁⹂⹃⹄⹅⹆⹇⹈⹉⹊⹋⹌⹍⹎⹏⹒⹓⹔⹕⹖⹗⹘⹙⹚⹛⹜⹝、。〃〈〉《》「」『』【】〔〕〖〗〘〙〚〛〜〝〞〟〰〽゠・꓾꓿꘍꘎꘏꙳꙾꛲꛳꛴꛵꛶꛷꡴꡵꡶꡷꣎꣏꣸꣹꣺꣼꤮꤯꥟꧁꧂꧃꧄꧅꧆꧇꧈꧉꧊꧋꧌꧍꧞꧟꩜꩝꩞꩟꫞꫟꫰꫱꯫﴾﴿︐︑︒︓︔︕︖︗︘︙︰︱︲︳︴︵︶︷︸︹︺︻︼︽︾︿﹀﹁﹂﹃﹄﹅﹆﹇﹈﹉﹊﹋﹌﹍﹎﹏﹐﹑﹒﹔﹕﹖﹗﹘﹙﹚﹛﹜﹝﹞﹟﹠﹡﹣﹨﹪﹫！＂＃％＆＇（）＊，－．／：；？＠［＼］＿｛｝｟｠｡｢｣､･𐄀𐄁𐄂𐎟𐏐𐕯𐡗𐤟𐤿𐩐𐩑𐩒𐩓𐩔𐩕𐩖𐩗𐩘𐩿𐫰𐫱𐫲𐫳𐫴𐫵𐫶𐬹𐬺𐬻𐬼𐬽𐬾𐬿𐮙𐮚𐮛𐮜𐺭𐽕𐽖𐽗𐽘𐽙𐾆𐾇𐾈𐾉𑁇𑁈𑁉𑁊𑁋𑁌𑁍𑂻𑂼𑂾𑂿𑃀𑃁𑅀𑅁𑅂𑅃𑅴𑅵𑇅𑇆𑇇𑇈𑇍𑇛𑇝𑇞𑇟𑈸𑈹𑈺𑈻𑈼𑈽𑊩𑑋𑑌𑑍𑑎𑑏𑑚𑑛𑑝𑓆𑗁𑗂𑗃𑗄𑗅𑗆𑗇𑗈𑗉𑗊𑗋𑗌𑗍𑗎𑗏𑗐𑗑𑗒𑗓𑗔𑗕𑗖𑗗𑙁𑙂𑙃𑙠𑙡𑙢𑙣𑙤𑙥𑙦𑙧𑙨𑙩𑙪𑙫𑙬𑚹𑜼𑜽𑜾𑠻𑥄𑥅𑥆𑧢𑨿𑩀𑩁𑩂𑩃𑩄𑩅𑩆𑪚𑪛𑪜𑪞𑪟𑪠𑪡𑪢𑬀𑬁𑬂𑬃𑬄𑬅𑬆𑬇𑬈𑬉𑱁𑱂𑱃𑱄𑱅𑱰𑱱𑻷𑻸𑽃𑽄𑽅𑽆𑽇𑽈𑽉𑽊𑽋𑽌𑽍𑽎𑽏𑿿𒑰𒑱𒑲𒑳𒑴𒿱𒿲𖩮𖩯𖫵𖬷𖬸𖬹𖬺𖬻𖭄𖺗𖺘𖺙𖺚𖿢𛲟𝪇𝪈𝪉𝪊𝪋𞥞𞥟"""
+# NOTE: This list diverges slightly from the raw list, since []\ must be escaped
+# The [] need to be escaped to avoid prematurely closing the regex character class
+# The \ needs to be escaped to be considered as a raw \
+
+# https://www.compart.com/en/unicode/category
+# https://unicode.org/Public/UNIDATA/UnicodeData.txt
+
+
 """Commonly occurring strings which are some kind of valid Toki Pona or external token"""
 ALLOWABLES = {
     "cw",  # Content Warning

sonatoki/ilo.py

@@ -1,5 +1,4 @@
 # STL
-import logging
 from typing import List, Type, Tuple
 
 # LOCAL
@@ -9,8 +8,6 @@ from sonatoki.Cleaners import Cleaner
 from sonatoki.Tokenizers import Tokenizer
 from sonatoki.Preprocessors import Preprocessor
 
-LOG = logging.getLogger(__name__)
-
 
 class Ilo:
     __preprocessors: List[Type[Preprocessor]]
@@ -20,7 +17,6 @@ class Ilo:
     __scoring_filters: List[Type[Filter]]
     __scorer: Type[Scorer]
     __passing_score: Number
-    logging_threshold: Number = -1
 
     def __init__(
         self,
@@ -104,14 +100,6 @@ class Ilo:
         score = self.score_tokens(cleaned)
         result = score >= self.__passing_score
 
-        if score <= self.logging_threshold:
-            LOG.debug("msg: %.2f  %s", score, repr(message))
-            LOG.debug("preproc:   %s", repr(preprocessed))
-            LOG.debug("tokenized: %s", tokenized)
-            LOG.debug("filtered:  %s", filtered)
-            LOG.debug("cleaned:   %s", cleaned)
-        # TODO: Move to each function? Loses ability to control when logging occurs by threshold
-
         return preprocessed, tokenized, filtered, cleaned, score, result
 
     def is_toki_pona(self, message: str) -> bool:

{sonatoki-0.1.3.dist-info → sonatoki-0.1.5.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: sonatoki
-Version: 0.1.3
+Version: 0.1.5
 Summary: ilo li moku e toki li pana e sona ni: ni li toki ala toki pona?
 Author-Email: "jan Kekan San (@gregdan3)" <gregory.danielson3@gmail.com>
 License: AGPL-3.0-or-later

sonatoki-0.1.5.dist-info/RECORD

@@ -0,0 +1,16 @@
+sonatoki-0.1.5.dist-info/METADATA,sha256=wJBa9CKSni9dcfQGQZp_-FSfANJHmTfZdYSCMH0Wolg,5225
+sonatoki-0.1.5.dist-info/WHEEL,sha256=vnE8JVcI2Wz7GRKorsPArnBdnW2SWKWGow5gu5tHlRU,90
+sonatoki-0.1.5.dist-info/licenses/LICENSE,sha256=DZak_2itbUtvHzD3E7GNUYSRK6jdOJ-GqncQ2weavLA,34523
+sonatoki/Cleaners.py,sha256=gTZ9dSsnvKVUtxM_ECSZ-_2heh--nD5A9dCQR1ATb1c,1160
+sonatoki/Configs.py,sha256=iY6Lyn1rMi7iF0M62yx0ET4pEb35-QAd1FS0tkyUfSc,1935
+sonatoki/Filters.py,sha256=xanTOKxasW_2OpQXsnk5pzbM1FmG8y46pakuTfMz9Hw,4470
+sonatoki/Preprocessors.py,sha256=FqBcHirsXV_91mj99ju9AnbsHaCFctSnuz_vca9ckSY,4441
+sonatoki/Scorers.py,sha256=w5p4qPzpEhR-xHaOXzqulN01OKtLcSPsTrgKyMhfAaQ,3658
+sonatoki/Tokenizers.py,sha256=Wqyf36d1OST7sVrANlLixDIOYWSctD5rVg1_MlaPrCw,2310
+sonatoki/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+sonatoki/__main__.py,sha256=6xc-wIrrFo9wTyn4zRQNAmqwmJBtVvCMwV-CrM-hueA,82
+sonatoki/constants.py,sha256=LZVI0322kTsqwrGQIJfgSl6IIZS-ste1rd9isRzcNqM,4883
+sonatoki/ilo.py,sha256=yyLgNPI0Hmb4f1BzX6IRHr11FPChfL2xDR_9odlr8_8,3849
+sonatoki/linku.json,sha256=B5KNdhyM5UEfMciROgh1ECHr3i-ASBeMvwrkzNJX47c,271013
+sonatoki/sandbox.json,sha256=hx6LRsfvmmTtqXcXIyCsfSaGK3DZ-GCdbM8xhZQBHoA,77650
+sonatoki-0.1.5.dist-info/RECORD,,

sonatoki-0.1.3.dist-info/RECORD

@@ -1,16 +0,0 @@
-sonatoki-0.1.3.dist-info/METADATA,sha256=ivcjgCnmdW1Typsn01RgqHX4PePGx6r4U_Ms5h5ksYo,5225
-sonatoki-0.1.3.dist-info/WHEEL,sha256=vnE8JVcI2Wz7GRKorsPArnBdnW2SWKWGow5gu5tHlRU,90
-sonatoki-0.1.3.dist-info/licenses/LICENSE,sha256=DZak_2itbUtvHzD3E7GNUYSRK6jdOJ-GqncQ2weavLA,34523
-sonatoki/Cleaners.py,sha256=gTZ9dSsnvKVUtxM_ECSZ-_2heh--nD5A9dCQR1ATb1c,1160
-sonatoki/Configs.py,sha256=yzTbjEyWS7sysaQ_9mIQMmO-acpEgQBzv4foP_J5x1o,1939
-sonatoki/Filters.py,sha256=RgbOXLat30WdsJW4y1DlMNttmGfVtLM6T7cD_qK-ASo,4194
-sonatoki/Preprocessors.py,sha256=uJ8-Y51gcgu5Wrri9BiP1F1YT-cYiLeWhrquFbYi9AI,3347
-sonatoki/Scorers.py,sha256=V293DBiupBiujzuc4yMrKOAiuNTLltIsiCzIAlLeokA,4129
-sonatoki/Tokenizers.py,sha256=fvqxpubs2F63va2RzZKZQhZbFnVaC_9haXIA9Mqznis,1942
-sonatoki/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-sonatoki/__main__.py,sha256=6xc-wIrrFo9wTyn4zRQNAmqwmJBtVvCMwV-CrM-hueA,82
-sonatoki/constants.py,sha256=m0Z4At6MfbqZRio2glT3J3zT9x_itcWZBT_G82mpaVc,1647
-sonatoki/ilo.py,sha256=oN14iYFKxgjFjjOslgqBrMaIgpnvS5gO6MscbS0JS5A,4343
-sonatoki/linku.json,sha256=B5KNdhyM5UEfMciROgh1ECHr3i-ASBeMvwrkzNJX47c,271013
-sonatoki/sandbox.json,sha256=hx6LRsfvmmTtqXcXIyCsfSaGK3DZ-GCdbM8xhZQBHoA,77650
-sonatoki-0.1.3.dist-info/RECORD,,