Merge pull request #19 from interscript/fix/remaining-singletons

fix: quote escapes, named dependency stages, not_word, library string classes

Author: ronaldtse

src/interscript/engine.py

@@ -139,7 +139,15 @@ def _run_op(self, op: dict, text: str) -> str:
                 target = deps[alias]
             if self._loader is None:
                 raise ExecutionError(f"run {op['map']!r}: no map loader configured")
-            return self._loader(target).transliterate(text)
+            dep = self._loader(target)
+            stage_name = op.get("stage")
+            if not stage_name:
+                return dep.transliterate(text)
+            out = text
+            for st in dep.tree.get("stages", []):
+                if st.get("name") == stage_name:
+                    out = dep._run_stage(st, out)
+            return out
         if kind == "separate":
             separator = op.get("separator", " ")
             return separator.join(text)

src/interscript/expr.py

@@ -17,7 +17,7 @@
     r'|any\(\s*"(?P<rlo>(?:[^"\\]|\\.)*)"\s*\.\.\s*"(?P<rhi>(?:[^"\\]|\\.)*)"\s*\)'
     r'|any\(\s*"(?P<cls>(?:[^"\\]|\\.)*)"\s*\)'
     r"|(?P<space>\bspace\b)|(?P<boundary>\bboundary\b)"
-    r"|(?P<stdin_kw>\b(?:alpha|digit|word|any_character)\b)"
+    r"|(?P<stdin_kw>\b(?:alpha|digit|word|not_word|any_character)\b)"
     r"|(?P<nwb>\bnon_word_boundary\b)"
     r'|capture\(\s*(?P<grp>(?:[^()\\]|\\.|\([^()]*\))*)\s*\)' 
     r"|(?P<line_end>\bline_end\b)|(?P<line_start>\bline_start\b)"
@@ -74,7 +74,7 @@ def _read_bracketed(expr: str, pos: int) -> tuple[str, int]:
                 return expr[pos + 1 : i], i + 1
         i += 1
     raise ValueError(f"unterminated any([ in {expr!r}")
-_UNESC = re.compile(r"\\u([0-9a-fA-F]{4})")
+_UNESC = re.compile(r"\\(u[0-9a-fA-F]{4}|.)", re.S)
 
 SPACE = re.escape(" ")
 
@@ -84,6 +84,7 @@ def _read_bracketed(expr: str, pos: int) -> tuple[str, int]:
     "alpha": "[a-zA-Z]",
     "digit": "[0-9]",
     "word": "[a-zA-Z0-9_]",
+    "not_word": "[^a-zA-Z0-9_]",
     "any_character": ".",
 }
 
@@ -100,7 +101,13 @@ def _read_bracketed(expr: str, pos: int) -> tuple[str, int]:
 
 
 def _unesc(s: str) -> str:
-    return _UNESC.sub(lambda m: chr(int(m.group(1), 16)), s)
+    def repl(m: re.Match) -> str:
+        g = m.group(1)
+        if g[0] == "u" and len(g) == 5:
+            return chr(int(g[1:], 16))
+        return {"n": "\n", "t": "\t", "r": "\r"}.get(g, g)
+
+    return _UNESC.sub(repl, s)
 
 
 _ANY_LIST = re.compile(r"any\(\s*\[")

src/interscript/isc.py

@@ -59,7 +59,7 @@ def __init__(self, subs: list[dict], capture_rules: list[dict]) -> None:
 _CONSTRAINTS = {"before", "after", "not_before", "not_after"}
 # Stdlib aliases usable as bare names (Ruby Stdlib::ALIASES, resolved
 # before doc-local aliases, mirroring the interpreter's lookup order).
-_STDLIB_EXPR = {"alpha", "digit", "word", "any_character"}
+_STDLIB_EXPR = {"alpha", "digit", "word", "not_word", "any_character"}
 # Tokens that terminate an item inside a rule; a bare word equal to one
 # of these is a keyword, never an alias reference.
 _KEYWORDS = _CONSTRAINTS | {"to", "from", "note"}
@@ -768,7 +768,7 @@ def _regex_of(item: dict, aliases: dict[str, str]) -> str:
         if item.get("map"):
             return _qualified_regex(item["map"], name)
         if name in _STDLIB_EXPR:
-            return {"alpha": "[a-zA-Z]", "digit": "[0-9]", "word": "[a-zA-Z0-9_]", "any_character": "."}[name]
+            return {"alpha": "[a-zA-Z]", "digit": "[0-9]", "word": "[a-zA-Z0-9_]", "not_word": "[^a-zA-Z0-9_]", "any_character": "."}[name]
         if name in _imported:
             return _qualified_regex(None, name)
         if name not in aliases:
@@ -875,7 +875,10 @@ def _stage_tree(
         dep = body_item.get("dependency")
         if dep:
             target = dep_aliases.get(dep, dep)
-            return {"kind": "run", "map": target}
+            op = {"kind": "run", "map": target}
+            if body_item.get("stage"):
+                op["stage"] = body_item["stage"]
+            return op
         return {"kind": "run_stage", "name": body_item["stage"]}
     if kind in ("compose", "decompose"):
         return {"kind": kind}
@@ -979,7 +982,14 @@ def _library_aliases(name: str, libs_dir: Path) -> dict[str, str]:
                 # Library aliases may reference aliases defined earlier
                 # in the same file (var-kor: jamo = any([jamo_leading_cons,
                 # ...])); substitute them, outside string literals only.
-                aliases[m.group(1)] = _subst_alias_refs(value, aliases)
+                value = _subst_alias_refs(value, aliases)
+                # A library alias holding a bare string acts as a
+                # character class inside any() (measured: the unicode
+                # library's greek); an in-map string alias would be a
+                # literal sequence instead.
+                if re.fullmatch(r'"(?:[^"\\]|\\.)*"', value):
+                    value = f"any({value})"
+                aliases[m.group(1)] = value
     _LIB_CACHE[name] = aliases
     return aliases
 

tests/test_engine.py

@@ -357,3 +357,42 @@ def test_any_character_renders():
     e = _load("kp-kor-Hang-Latn-2002")
     assert e.transliterate("고비리") == "Kobi-ri"
     assert e.transliterate("교구동") == "Kyogu-dong"
+
+
+@pytest.mark.skipif(not MAPS.is_dir(), reason="interscript maps repo not present")
+def test_quote_escape_survives_the_round_trip():
+    """gost-rus maps the hard sign to a double quote (to "\\""); the
+    value round-trips through expression source, where the unescaper
+    only decoded \\uXXXX — \\" stayed a backslash-quote and съезд came
+    out s\\"ezd."""
+    e = _load("gost-rus-Cyrl-Latn-7.79-2000-2002")
+    assert e.transliterate("съезд") == 's"ezd'
+
+
+@pytest.mark.skipif(not MAPS.is_dir(), reason="interscript maps repo not present")
+def test_run_executes_the_named_dependency_stage():
+    """mvd-rus-2010 runs the 2008 dependency's translit stage — not its
+    main, which ends in compose. Running main precomposed l+U+0301 into
+    ĺ before the postrule could strip the acute (Vasiĺeva)."""
+    e = _load("mvd-rus-Cyrl-Latn-2010")
+    assert e.transliterate("Васiльева") == "Vasileva"
+
+
+@pytest.mark.skipif(not MAPS.is_dir(), reason="interscript maps repo not present")
+def test_not_word_stdlib_alias_renders():
+    """odni-che maps the digit 1 (a palochka stand-in) guarded by
+    not_word; the alias had no stdlib entry and the rules were
+    dropped — Ахмадк1ант kept its 1."""
+    e = _load("odni-che-Cyrl-Latn-2015")
+    assert e.transliterate("Ахмадк1ант") == "Akhmadkant"
+
+
+@pytest.mark.skipif(not MAPS.is_dir(), reason="interscript maps repo not present")
+def test_library_string_alias_is_a_class_inside_any():
+    """Measured in Ruby: an alias imported from a library whose value
+    is a plain string acts as a CHARACTER CLASS inside any() (the
+    unicode library's greek), while an in-map string alias is a literal
+    sequence. alalc-ell's γ-nasal rules guard on any(greek)."""
+    e = _load("alalc-ell-Grek-Latn-2010")
+    assert e.transliterate("γκέγκε") == "gkenke"
+    assert e.transliterate("Λαγκαδάς") == "Lankadas"

tests/test_isc.py

@@ -72,7 +72,7 @@ def test_tests_render_as_pairs():
 def test_dependencies_resolve_alias_for_run():
     tree = isc_to_tree(SAMPLE)
     stage_children = tree["stages"][0]["children"]
-    assert stage_children[0] == {"kind": "run", "map": "other-map"}
+    assert stage_children[0] == {"kind": "run", "map": "other-map", "stage": "main"}
 
 
 def test_parallel_subs_render_as_expr_source():