sonatoki 0.5.0__tar.gz → 0.5.1__tar.gz

{sonatoki-0.5.0 → sonatoki-0.5.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: sonatoki
-Version: 0.5.0
+Version: 0.5.1
 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-0.5.0 → sonatoki-0.5.1}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "sonatoki"
-version = "0.5.0"
+version = "0.5.1"
 description = "ilo li moku e toki li pana e sona ni: ni li toki ala toki pona?"
 authors = [
     { name = "jan Kekan San (@gregdan3)", email = "gregory.danielson3@gmail.com" },

{sonatoki-0.5.0 → sonatoki-0.5.1}/src/sonatoki/Configs.py

@@ -73,6 +73,7 @@ PrefConfig: IloConfig = {
     "scoring_filters": [
         Or(NimiLinkuCore, NimiLinkuCommon, NimiUCSUR, Miscellaneous),
         And(LongSyllabic, Not(FalsePosSyllabic)),
+        # NOTE: These are allowed to pass name and alphabetic below, because they *could* be wrong
         LongProperName,
         LongAlphabetic,
     ],

{sonatoki-0.5.0 → sonatoki-0.5.1}/src/sonatoki/Filters.py

@@ -351,6 +351,10 @@ class Or:
         else:
             other_filters.extend(member_filters)
 
+        if len(other_filters) == 1:  # we only had member filters
+            # TODO: this sucks?
+            return other_filters[0]
+
         filter = cls.__generic_filter(*other_filters)
         return filter
 

{sonatoki-0.5.0 → sonatoki-0.5.1}/src/sonatoki/constants.py

@@ -519,8 +519,10 @@ ALLOWABLES = {
     "kxk",  # ken ala ken
     "wxw",  # wile ala wile
     "msa",
+    "anusem",
 }
 
+# NOTE: This is being tracked manually rather than fetched from syllabics.txt until I am convinced that solution is appropriate
 FALSE_POS_SYLLABIC = {
     # ordered by frequency in previous TPT data
     "like",
@@ -540,6 +542,7 @@ FALSE_POS_SYLLABIC = {
     "man",
     # "son",  # sona typo?
     "joke",
+    # pon would go here
     "so",
     "ten",
     "make",
@@ -548,11 +551,14 @@ FALSE_POS_SYLLABIC = {
     # "aka" # in sandbox
     "into",
     "in",
+    "no",
     "some",
+    # "papa",
     "on",
     "me",
     "ipa",
     "sun",
+    "mine",
     "sense",
     "none",
     "meme",
@@ -561,28 +567,101 @@ FALSE_POS_SYLLABIC = {
     "mon",
     "take",
     "luna",
-    "anti",
     "elo",
+    "japanese",
     "an",
+    "anti",
     "win",
     "won",
-    "we",
+    "we",  # word in sandbox
     "men",
     "ton",
     "woke",
+    "sen",  # seen
+    "se",  # see
     "semi",
     "male",
+    # "pen",  # borderline
+    "woman",
+    "line",
+    "meta",
+    "mini",
+    "sine",
+    # "min",  # borderline
+    "oposite",
+    "anime",
+    "potato",
+    # "japan",
+    "nose",
+    "kilo",
+    "alone",
+    "minute",
+    "late",
+    "women",
+    "leson",
+    "amen",
+    "tote",
+    "lame",
+    "online",
+    "tone",
+    "ate",
+    "mile",
+    "melon",
+    "tense",
+    "nonsense",
+    "nine",
+    "emo",
+    "unlike",
+    "lone",
+    # manual additions
+    "alike",
+    "amuse",
+    "antelope",
+    "antena",
+    "apetite",
+    "asasin",
+    "asasinate",
+    "asinine",
+    "asinine",
+    "asume",
+    "atone",
+    "awake",
+    "awaken",
+    "eliminate",
+    "elite",
+    "misuse",
+    "emanate",
+    "iluminate",
+    "imense",
+    "imitate",
+    "insane",
+    "insolate",
+    "insulate",
+    "intense",
+    "lemon",
+    "manipulate",
 }
 
 FALSE_POS_ALPHABETIC: Set[str] = {
     "t",
     "is",
+    "as",
     "not",
+    "link",
+    "wait",
     "lol",
+    "new",
     "also",
     "isn",  # TODO: tokenizer....
     "mean",
     "means",
+    "it",
+    "moment",
+    "its",
+    "lmao",
+    "new",
+    "wel",
+    "makes",
 }
 
 UCSUR_RANGES = [

{sonatoki-0.5.0 → sonatoki-0.5.1}/tests/test_filters.py

@@ -21,6 +21,7 @@ from sonatoki.Filters import (
     Punctuation,
     AlphabeticRe,
     LongSyllabic,
+    MemberFilter,
     NimiLinkuCore,
     PunctuationRe,
     LongAlphabetic,
@@ -207,6 +208,8 @@ def test_OrFilter(s: str):
 @given(st.sampled_from(list(NIMI_PU | NIMI_LINKU_OBSCURE)))
 def test_MemberFilters_OrFilter(s: str):
     filter = Or(NimiPu, NimiLinkuObscure)
+    assert issubclass(filter, MemberFilter)
+
     res = filter.filter(s)
     res_pu = NimiPu.filter(s)
     res_obscure = NimiLinkuObscure.filter(s)

{sonatoki-0.5.0 → sonatoki-0.5.1}/tests/test_ilo.py

@@ -73,6 +73,7 @@ SYLLABIC_MATCHES = [
 
 ALPHABETIC_MATCHES = [
     "mi mtue o kama sona",
+    "mi mute o kma son",  # this one is odd because `son` is an unintended phonetic match
     "mi mute o kama kne snoa a",
     "ni li tptpt",
     "mi wile pana lon sptp",
@@ -120,6 +121,7 @@ EXCESSIVE_SYLLABICS = [
     "a ton of insolate puke. make no amen, no joke.",
     "I elope so, to an elite untaken tune, some unwise tone",
     "insane asinine lemon awesome atone joke",
+    "insane asinine lemon awesome atone",  # i got more clever
 ]
 
 EXCESSIVE_ALPHABETICS = [
@@ -177,7 +179,6 @@ KNOWN_BAD = (
 FALSE_NEGATIVES = [
     # emoticon should not be a problem
     # a token that is one edit off a known word should be allowed
-    "mi mute o kma son",  # this one is obnoxious because `son` did match phonetically before
     "mi pnoa",
     "tok",
     "mut",
@@ -187,7 +188,6 @@ FALSE_NEGATIVES = [
 ]
 
 FALSE_POSITIVES = [
-    "insane asinine lemon awesome atone",
     "lete li ike x.x",  # this is an emoticon but passes because 'x' is in Filters.Miscellaneous
 ]
 

{sonatoki-0.5.0 → sonatoki-0.5.1}/tests/test_properties.py

@@ -25,9 +25,11 @@ from sonatoki.constants import (
     NIMI_LINKU_CORE,
     NIMI_PU_SYNONYMS,
     NIMI_LINKU_COMMON,
+    FALSE_POS_SYLLABIC,
     NIMI_LINKU_OBSCURE,
     NIMI_LINKU_SANDBOX,
     NIMI_LINKU_UNCOMMON,
+    FALSE_POS_ALPHABETIC,
 )
 
 
@@ -76,3 +78,10 @@ def test_nimi_linku_properties(s: str):
     assert Syllabic.filter(s), repr(s)
     assert Phonotactic.filter(s), repr(s)
     # Passing phonotactic implies all of the above
+
+
+@given(st.sampled_from(list(FALSE_POS_ALPHABETIC)))
+def test_false_pos_properties(s: str):
+    res_syllabic = Syllabic.filter(s)
+    res_alphabetic = Alphabetic.filter(s)
+    assert res_alphabetic and not res_syllabic