Merge pull request 'fix/tokenizer-bug' (#1) from fix/tokenizer-bug into main

Reviewed-on: #1

.gitea/workflows/ci.yml

@@ -0,0 +1,29 @@
+name: CI
+
+on:
+  push:
+    branches: ["main"]
+  pull_request:
+    branches: ["main"]
+
+jobs:
+  test:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+
+      - name: Set up Python
+        uses: actions/setup-python@v5
+        with:
+          python-version: "3.11"
+
+      - name: Install Core with Dev Dependencies
+        run: |
+          pip install --upgrade pip
+          pip install -e .[dev]
+
+      - name: Run Quality Checks (Linter)
+        run: ruff check src/
+
+      - name: Execute Test Suite
+        run: pytest tests/ -v

.gitea/workflows/publish.yml

@@ -0,0 +1,77 @@
+name: Publish Core
+
+on:
+  push:
+    tags:
+      - "v*" # Fires directly on v0.1.0, v0.2.0 etc.
+
+jobs:
+  # Enforce that tests MUST pass before release can execute
+  verify:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+      - name: Set up Python
+        uses: actions/setup-python@v5
+        with:
+          python-version: "3.11"
+      - name: Install and Verify
+        run: |
+          pip install -e .[dev]
+          pytest tests/ -v
+
+  publish:
+    needs: verify # Blocks execution if verify job fails
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+
+      - name: Set up Python
+        uses: actions/setup-python@v5
+        with:
+          python-version: "3.11"
+
+      - name: Install Release Tools
+        run: pip install build twine
+
+      - name: Build Wheel and Source Distribution
+        run: python -m build .
+      - name: Publish Package to PyPI
+        env:
+          TWINE_USERNAME: __token__
+          # Inherits your clean Organization level secret
+          TWINE_PASSWORD: ${{ secrets.PYPI_TOKEN }}
+        run: twine upload --skip-existing dist/*
+
+      - name: Build Gitea Release with Assets
+        env:
+          # Pulls your clean Organization level Gitea Token
+          GIT_RELEASE_TOKEN: ${{ secrets.GIT_RELEASE_TOKEN }}
+        run: |
+          TAG=${GITHUB_REF#refs/tags/}
+
+          # Delete existing release block if present
+          EXISTING=$(curl -s -H "Authorization: token $GIT_RELEASE_TOKEN" "${{ github.server_url }}/api/v1/repos/${{ github.repository }}/releases/tags/$TAG")
+          EXISTING_ID=$(echo $EXISTING | python -c "import sys,json; d=json.load(sys.stdin); print(d.get('id',''))" 2>/dev/null || echo "")
+          if [ -n "$EXISTING_ID" ]; then
+            curl -s -X DELETE -H "Authorization: token $GIT_RELEASE_TOKEN" "${{ github.server_url }}/api/v1/repos/${{ github.repository }}/releases/$EXISTING_ID"
+          fi
+
+          # Create fresh production release container
+          RELEASE=$(curl -s -X POST \
+            -H "Authorization: token $GIT_RELEASE_TOKEN" \
+            -H "Content-Type: application/json" \
+            -d "{
+              \"tag_name\": \"$TAG\",
+              \"name\": \"foreignthon $TAG\",
+              \"body\": \"Release version $TAG of foreignthon core compiler engine.\",
+              \"draft\": false,
+              \"prerelease\": false
+            }" "${{ github.server_url }}/api/v1/repos/${{ github.repository }}/releases")
+
+          RELEASE_ID=$(echo $RELEASE | python -c "import sys,json; print(json.load(sys.stdin)['id'])")
+
+          # Upload wheels directly into Gitea Assets tab
+          for FILE in dist/*; do
+            curl -s -X POST -H "Authorization: token $GIT_RELEASE_TOKEN" -F "attachment=@$FILE" "${{ github.server_url }}/api/v1/repos/${{ github.repository }}/releases/$RELEASE_ID/assets"
+          done

README.md

@@ -1 +1,56 @@
-# foreignthon
+# ForeignThon
+
+Write Python in any human language.
+
+ForeignThon transpiles `.es.py`, `.ta.py` and more into standard Python — keywords, builtins, exceptions, all translated. Errors come back in your language too.
+
+```python
+# main.es.py
+def saludar(nombre):
+    retornar f"Hola, {nombre}!"
+
+para i en dist(3):
+    escribir(saludar(f"mundo {i}"))
+```
+
+```bash
+fpy run main.es.py
+# Hola, mundo 0!
+# Hola, mundo 1!
+# Hola, mundo 2!
+```
+
+## Install
+
+```bash
+pip install foreignthon      # Core 
+pip install foreignthon-es   # Spanish
+pip install foreignthon-ta   # Tamil
+```
+
+## Quick start
+
+```bash
+fpy new myproject --lang es
+cd myproject
+fpy run src/main.es.py
+```
+
+## Commands
+
+| Command | Description |
+|---|---|
+| `fpy new <name> --lang <code>` | Scaffold a new project |
+| `fpy run <file>` | Transpile and run |
+| `fpy compile <file>` | Transpile to `.compiled.py` |
+| `fpy decompile <file> --lang <code>` | Convert Python back to a language |
+| `fpy check <file>` | Validate without running |
+
+## Documentation
+
+→ [fpy.keshavanand.net](https://foreignthon.keshavanand.net)
+
+## License
+
+GPL v3
+

pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
 
 [project]
 name = "foreignthon"
-version = "0.5.2"
+version = "0.5.4"
 description = "Write Python in any language. Transpiles foreign-language .xx.py files to standard Python."
 license = { text = "GPL v3" }
 requires-python = ">=3.9"

src/foreignthon/__init__.py

@@ -1,4 +1,4 @@
-from importlib.metadata import version, PackageNotFoundError
+from importlib.metadata import PackageNotFoundError, version
 
 try:
     __version__ = version("foreignthon")

src/foreignthon/cli.py

@@ -292,6 +292,7 @@ def check(files: tuple):
 def langs():
     """List all installed language packs with their versions and authors."""
     import json
+
     from .pack import _discover_packs
 
     packs = _discover_packs()

src/foreignthon/transpiler.py

@@ -3,6 +3,8 @@ from __future__ import annotations
 import io
 import re
 import tokenize
+import unicodedata
+from functools import lru_cache
 from pathlib import Path
 
 from .pack import load_pack
@@ -79,7 +81,103 @@ def _get_slice(source_lines: list[str], sr: int, sc: int, er: int, ec: int) -> s
     return "".join(parts)
 
 
+@lru_cache(maxsize=2048)
+def _is_safe_token(s: str) -> bool:
+    """Return True if the tokenizer produces `s` as a single NAME token.
+
+    Some Unicode scripts (e.g. Tamil) contain combining characters that the
+    tokenize module's regex treats as token boundaries, even though the full
+    string passes str.isidentifier().  Keys that fail this check need a
+    pre-pass string replacement before tokenization.
+    """
+    try:
+        toks = [
+            t for t in tokenize.generate_tokens(io.StringIO(s + "\n").readline)
+            if t.type not in (
+                tokenize.ENDMARKER, tokenize.NEWLINE, tokenize.NL,
+                tokenize.COMMENT, tokenize.INDENT, tokenize.DEDENT,
+            )
+        ]
+        return len(toks) == 1 and toks[0].type == tokenize.NAME and toks[0].string == s
+    except tokenize.TokenError:
+        return False
+
+
+def _is_id_char(c: str) -> bool:
+    """True if `c` can be part of an identifier (letter, mark, digit, or underscore)."""
+    cat = unicodedata.category(c)
+    return cat.startswith(("L", "M", "N")) or c == "_"
+
+
+def _code_region_replace(source: str, unsafe_mapping: dict) -> str:
+    """Replace unsafe keys in source only in code regions (not string literals or comments).
+
+    The tokenizer correctly identifies STRING/COMMENT boundaries even when NAME
+    tokens are mangled by combining characters, so we use it to find protected spans.
+    Identifier-boundary checks prevent partial matches inside longer words.
+    """
+    lines = source.splitlines(keepends=True)
+    cumlen = [0]
+    for line in lines:
+        cumlen.append(cumlen[-1] + len(line))
+
+    protected: list[tuple[int, int]] = []
+    try:
+        for tok in tokenize.generate_tokens(io.StringIO(source).readline):
+            if tok.type in (tokenize.STRING, tokenize.COMMENT):
+                sr, sc = tok.start
+                er, ec = tok.end
+                protected.append((cumlen[sr - 1] + sc, cumlen[er - 1] + ec))
+    except tokenize.TokenError:
+        pass
+
+    sorted_keys = sorted(unsafe_mapping, key=len, reverse=True)
+    result: list[str] = []
+    pos = 0
+    span_idx = 0
+    n = len(source)
+
+    while pos < n:
+        if span_idx < len(protected) and pos >= protected[span_idx][0]:
+            end = protected[span_idx][1]
+            result.append(source[pos:end])
+            pos = end
+            span_idx += 1
+            continue
+
+        code_end = protected[span_idx][0] if span_idx < len(protected) else n
+
+        matched = False
+        for k in sorted_keys:
+            klen = len(k)
+            if pos + klen > code_end:
+                continue
+            if source[pos:pos + klen] != k:
+                continue
+            before_ok = pos == 0 or not _is_id_char(source[pos - 1])
+            after_ok = (pos + klen >= n) or not _is_id_char(source[pos + klen])
+            if before_ok and after_ok:
+                result.append(unsafe_mapping[k])
+                pos += klen
+                matched = True
+                break
+
+        if not matched:
+            result.append(source[pos])
+            pos += 1
+
+    return "".join(result)
+
+
 def _swap_tokens(source: str, mapping: dict) -> str:
+    safe_mapping: dict[str, str] = {}
+    unsafe_mapping: dict[str, str] = {}
+    for k, v in mapping.items():
+        (safe_mapping if _is_safe_token(k) else unsafe_mapping)[k] = v
+
+    if unsafe_mapping:
+        source = _code_region_replace(source, unsafe_mapping)
+
     source_lines = source.splitlines(keepends=True)
     tokens = list(tokenize.generate_tokens(io.StringIO(source).readline))
 
@@ -94,8 +192,8 @@ def _swap_tokens(source: str, mapping: dict) -> str:
         gap = _get_slice(source_lines, prev_end[0], prev_end[1], s_row, s_col)
         result.append(gap)
 
-        if tok_type == tokenize.NAME and tok_string in mapping:
+        if tok_type == tokenize.NAME and tok_string in safe_mapping:
-            result.append(mapping[tok_string])
+            result.append(safe_mapping[tok_string])
         else:
             result.append(tok_string)
 

tests/test_engine.py

@@ -0,0 +1,154 @@
+from __future__ import annotations
+
+import ast
+import json
+import textwrap
+from pathlib import Path
+
+import pytest
+
+from foreignthon.transpiler import _check_shebang, _detect_lang, detranspile, transpile
+
+# ---------------------------------------------------------------------------
+# Setup: Load the local test JSON fixture
+# ---------------------------------------------------------------------------
+
+TEST_PACK_PATH = Path(__file__).parent / "test_pack.json"
+TEST_PACK = json.loads(TEST_PACK_PATH.read_text(encoding="utf-8"))
+
+
+def core_transpile(src: str) -> str:
+    # We pass "es" to match the JSON's meta code, but we feed it the local pack
+    return transpile(textwrap.dedent(src).strip() + "\n", "es", pack=TEST_PACK)
+
+
+def core_detranspile(src: str, postfix: bool = False) -> str:
+    return detranspile(
+        textwrap.dedent(src).strip() + "\n", "es", postfix=postfix, pack=TEST_PACK
+    )
+
+
+def valid(src: str) -> bool:
+    try:
+        ast.parse(src)
+        return True
+    except SyntaxError:
+        return False
+
+
+# ---------------------------------------------------------------------------
+# 1. Core Mechanics: Translation & AST Validity
+# ---------------------------------------------------------------------------
+
+
+def test_engine_basic_translation():
+    # Ensures the engine reads the dictionary and swaps the words
+    src = """
+        para i en dist(5):
+            escribir(i)
+    """
+    out = core_transpile(src)
+    assert "for" in out and "in" in out and "range" in out and "print" in out
+    assert valid(out)
+
+
+# ---------------------------------------------------------------------------
+# 2. Core Mechanics: Safety Boundaries (Strings & Comments)
+# ---------------------------------------------------------------------------
+
+
+def test_engine_preserves_strings():
+    # The engine MUST NOT translate keywords hidden inside strings
+    out = core_transpile('mensaje = "si para mientras def clase"')
+    assert '"si para mientras def clase"' in out
+
+
+def test_engine_preserves_comments():
+    # The engine MUST NOT translate keywords hidden in comments
+    out = core_transpile("# si para mientras\nx = 1")
+    assert "# si para mientras" in out
+
+
+def test_engine_preserves_fstrings():
+    out = core_transpile('escribir(f"valor si={42}")')
+    assert "si" in out  # 'si' survives because it is inside the string
+
+
+# ---------------------------------------------------------------------------
+# 3. Core Mechanics: Postfix Syntax (@@)
+# ---------------------------------------------------------------------------
+
+
+def test_engine_postfix_reversal():
+    # Tests the engine's ability to move the keyword to the front
+    src = """
+        x = 5
+        x > 0 @@si:
+            escribir(x)
+    """
+    out = core_transpile(src)
+    assert "if x > 0:" in out
+    assert "@@" not in out
+    assert valid(out)
+
+
+def test_engine_mixed_prefix_postfix():
+    src = """
+        si x > 0:
+            escribir(x)
+        y < 0 @@si:
+            escribir(y)
+    """
+    out = core_transpile(src)
+    assert out.count("if") == 2
+    assert "@@" not in out
+
+
+# ---------------------------------------------------------------------------
+# 4. Core Mechanics: Decompilation (Round Trip)
+# ---------------------------------------------------------------------------
+
+
+def test_engine_detranspile():
+    # Standard Python should turn back into foreignthon syntax
+    src = """
+        if x > 0:
+            pass
+    """
+    out = core_detranspile(src)
+    assert "si" in out and "pasar" in out
+
+
+def test_engine_roundtrip():
+    # foreignthon -> Python -> foreignthon
+    original = "para i en dist(5):\n    escribir(i)\n"
+    compiled = core_transpile(original)
+    assert valid(compiled)
+
+    back = core_detranspile(compiled)
+    assert "para" in back and "dist" in back
+    # Accept either valid translation for 'print'
+    assert "escribir" in back or "imprimir" in back
+
+
+# ---------------------------------------------------------------------------
+# 5. Core Utilities: Detection & Shebangs
+# ---------------------------------------------------------------------------
+
+
+def test_detect_lang_from_extension():
+    assert _detect_lang(Path("script.es.py")) == "es"
+    assert _detect_lang(Path("script.ta.py")) == "ta"
+
+
+def test_detect_lang_bad_extension():
+    with pytest.raises(ValueError):
+        _detect_lang(Path("script.py"))
+
+
+def test_shebang_override():
+    assert _check_shebang("# foreignthon: fr\nsi x:\n    pasar", "es") == "fr"
+
+
+def test_shebang_default_when_absent():
+    assert _check_shebang("si x:\n    pasar", "es") == "es"

tests/test_pack.json

@@ -0,0 +1,147 @@
+{
+  "meta": {
+    "name": "Spanish",
+    "native_name": "Español",
+    "code": "es"
+  },
+  "keywords": {
+    "si": "if",
+    "sino": "else",
+    "osi": "elif",
+    "para": "for",
+    "mientras": "while",
+    "def": "def",
+    "clase": "class",
+    "importar": "import",
+    "de": "from",
+    "como": "as",
+    "retornar": "return",
+    "parar": "break",
+    "continuar": "continue",
+    "pasar": "pass",
+    "intentar": "try",
+    "excepto": "except",
+    "finalmente": "finally",
+    "lanzar": "raise",
+    "con": "with",
+    "en": "in",
+    "es": "is",
+    "y": "and",
+    "o": "or",
+    "no": "not",
+    "elim": "del",
+    "global": "global",
+    "nolocal": "nonlocal",
+    "afirmar": "assert",
+    "generar": "yield",
+    "esperar": "await",
+    "asinc": "async",
+    "lambda": "lambda",
+    "Verda": "True",
+    "Falso": "False",
+    "Nada": "None"
+  },
+  "builtins": {
+    "escribir": "print",
+    "imprimir": "print",
+    "entrada": "input",
+    "lon": "len",
+    "dist": "range",
+    "tipo": "type",
+    "ent": "int",
+    "dec": "float",
+    "texto": "str",
+    "lista": "list",
+    "dicc": "dict",
+    "conj": "set",
+    "tupla": "tuple",
+    "bool": "bool",
+    "abrir": "open",
+    "enumerar": "enumerate",
+    "map": "map",
+    "filtrar": "filter",
+    "ordenado": "sorted",
+    "invertido": "reversed",
+    "sum": "sum",
+    "min": "min",
+    "max": "max",
+    "abs": "abs",
+    "redondear": "round",
+    "rnd": "round",
+    "todos": "all",
+    "alguno": "any",
+    "esinstancia": "isinstance",
+    "teneatri": "hasattr",
+    "obtatri": "getattr",
+    "estabatri": "setattr",
+    "repr": "repr",
+    "formatear": "format",
+    "vars": "vars",
+    "sigue": "next",
+    "id": "id",
+    "car": "chr",
+    "hex": "hex",
+    "bin": "bin",
+    "oct": "oct"
+  },
+  "exceptions": {
+    "Excepcion": "Exception",
+    "ExcepcionBase": "BaseException",
+    "ErrorDeValor": "ValueError",
+    "ErrorDeTipo": "TypeError",
+    "ErrorDeClave": "KeyError",
+    "ErrorDeIndice": "IndexError",
+    "ErrorDeAtributo": "AttributeError",
+    "ErrorDeNombre": "NameError",
+    "ErrorDeImportacion": "ImportError",
+    "ErrorDelSistema": "OSError",
+    "ArchivoNoEncontrado": "FileNotFoundError",
+    "ErrorDeEjecucion": "RuntimeError",
+    "DetenerIteracion": "StopIteration",
+    "SalidaDelSistema": "SystemExit",
+    "InterrupcionDeTeclado": "KeyboardInterrupt",
+    "ErrorNoImplementado": "NotImplementedError",
+    "ErrorDeDivisionCero": "ZeroDivisionError",
+    "ErrorDeRecursion": "RecursionError",
+    "ErrorDeSintaxis": "SyntaxError",
+    "ErrorDeAfirmacion": "AssertionError",
+    "ErrorDeDesbordamiento": "OverflowError",
+    "ErrorDeMemoria": "MemoryError",
+    "ErrorDePermiso": "PermissionError",
+    "ErrorDeTiempoAgotado": "TimeoutError"
+  },
+  "error_messages": {
+    "SyntaxError": "Error de sintaxis",
+    "ValueError": "Error de valor",
+    "TypeError": "Error de tipo",
+    "KeyError": "Error de clave",
+    "IndexError": "Error de índice",
+    "AttributeError": "Error de atributo",
+    "NameError": "Error de nombre",
+    "ImportError": "Error de importación",
+    "FileNotFoundError": "Archivo no encontrado",
+    "ZeroDivisionError": "Error división por cero",
+    "RecursionError": "Error de recursión",
+    "RuntimeError": "Error de ejecución",
+    "MemoryError": "Error de memoria",
+    "OverflowError": "Error de desbordamiento",
+    "AssertionError": "Error de afirmación",
+    "NotImplementedError": "Error no implementado",
+    "StopIteration": "Detener iteración",
+    "KeyboardInterrupt": "Interrupción de teclado",
+    "PermissionError": "Error de permiso",
+    "TimeoutError": "Error de tiempo agotado"
+  },
+  "stdlib": {
+    "mate": "math",
+    "sis": "sys",
+    "fechahora": "datetime",
+    "tiempo": "time",
+    "aleatorio": "random",
+    "aleatoria": "random",
+    "colecciones": "collections",
+    "ruta": "pathlib",
+    "er": "re"
+  },
+  "postfix_keywords": []
+}