repl glossing

Author: neta-elad

README.md

@@ -226,7 +226,7 @@ will turn out as
   - [ ] Better cache
 - [ ] `translator`: Automatic translation according to gloss
   - [x] Evolve each form (with affixes)
-  - [ ] Match-up forms and glosses
+  - [x] Match-up forms and glosses
   - [ ] Better cache
 - [ ] `book`: Generate a grammar/lexicon reference with less boilerplate
   - [x] Conlang-to-English lexicon, with entries, affixes and templates:

pyconlang/ipa.py

@@ -1,7 +1,7 @@
 from cmd import Cmd
 from dataclasses import dataclass, field
 from operator import attrgetter
-from typing import Callable
+from typing import Callable, Tuple
 
 from prompt_toolkit import PromptSession
 from prompt_toolkit.history import FileHistory
@@ -12,6 +12,8 @@
 from .errors import show_exception
 from .evolve.types import Evolved
 from .translate import Translator
+from .types import Form
+from .unicode import center, length
 
 HISTORY_PATH = PYCONLANG_PATH / "repl.history"
 
@@ -115,6 +117,24 @@ def do_s(self, line: str) -> None:
         """
         self.do_simple(line)
 
+    def do_gloss(self, line: str) -> None:
+        """
+        Translates and glosses the translation.
+        """
+
+        def pair_width(pair: Tuple[Evolved, Form]) -> int:
+            return 2 + max(length(pair[0].modern), length(str(pair[1])))
+
+        gloss = self.translator.gloss_string(line)
+        print(" ".join(center(pair[0].modern, pair_width(pair), " ") for pair in gloss))
+        print(" ".join(center(str(pair[1]), pair_width(pair), " ") for pair in gloss))
+
+    def do_g(self, line: str) -> None:
+        """
+        See gloss.
+        """
+        return self.do_gloss(line)
+
 
 def run(command: str = "") -> None:
     session = ReplSession()

pyconlang/repl.py

@@ -1,5 +1,5 @@
 from dataclasses import dataclass, field
-from typing import List, Mapping
+from typing import List, Mapping, Sequence, Tuple
 
 from .evolve import Evolver
 from .evolve.types import Evolved
@@ -17,10 +17,15 @@ class Translator:
     def evolve(self, form: Form) -> Evolved:
         return self.evolver.evolve_single(self.lexicon.resolve(form))
 
+    def resolve_and_evolve(self, forms: Sequence[Form]) -> List[Evolved]:
+        return self.evolver.evolve([self.lexicon.resolve(form) for form in forms])
+
     def evolve_string(self, string: str) -> List[Evolved]:
-        return self.evolver.evolve(
-            [self.lexicon.resolve(form) for form in parse_sentence(string)]
-        )
+        return self.resolve_and_evolve(parse_sentence(string))
+
+    def gloss_string(self, string: str) -> Sequence[Tuple[Evolved, Form]]:
+        forms = parse_sentence(string)
+        return list(zip(self.resolve_and_evolve(forms), forms))
 
     def evolve_entry(self, entry: Entry) -> List[Evolved]:
         return [

pyconlang/translate.py

@@ -9,11 +9,17 @@
 class Rule:
     name: str
 
+    def __str__(self) -> str:
+        return f"@{self.name}"
+
 
 @dataclass(eq=True, frozen=True)
 class Canonical:
     name: str
 
+    def __str__(self) -> str:
+        return f"<{self.name}>"
+
 
 @dataclass(eq=True, frozen=True)
 class Proto:
@@ -25,6 +31,9 @@ def era_name(self) -> Optional[str]:
             return None
         return self.era.name
 
+    def __str__(self) -> str:
+        return f"*{self.form}{self.era or ''}"
+
 
 @dataclass(eq=True, frozen=True)
 class TemplateName:
@@ -76,6 +85,21 @@ def from_form(cls, form: "Form") -> "Compound":
             case _:
                 return Compound(form)
 
+    def __str__(self) -> str:
+        return (
+            "".join(
+                affix.name + "."
+                for affix in self.affixes
+                if affix.type == AffixType.PREFIX
+            )
+            + str(self.stem)
+            + "".join(
+                "." + affix.name
+                for affix in self.affixes
+                if affix.type == AffixType.SUFFIX
+            )
+        )
+
 
 Form = Union[SimpleForm, Compound]
 

pyconlang/types.py

@@ -0,0 +1,19 @@
+from wcwidth import wcswidth
+
+PRIMARY_STRESS = "\u02c8"
+
+
+def remove_primary_stress(word: str) -> str:
+    return word.replace(PRIMARY_STRESS, "")
+
+
+def length(string: str) -> int:
+    return wcswidth(string)
+
+
+def center(string: str, width: int, fill_char: str = " ") -> str:
+    diff = max(0, width - length(string))
+    add_left = int(diff / 2)
+    add_right = int((diff + 1) / 2)
+
+    return fill_char * add_left + string + fill_char * add_right

pyconlang/unicode.py

@@ -17,7 +17,8 @@ dependencies = [
     "click==8.1.3",
     "toml==0.10.2",
     "pyparsing==3.0.9",
-    "unidecode==1.3.4"
+    "unidecode==1.3.4",
+    "wcwidth==0.2.5 "
 ]
 dynamic = ["version"]
 

pyproject.toml

@@ -0,0 +1,2 @@
+def wcswidth(string: str) -> int:
+    ...