Tsuchinoko

Supported Features

This document lists all Python features currently supported by Tsuchinoko transpiler.

Core Syntax

• Variable declarations with type hints (x: int = 10)
• Basic types: int, float, str, bool, None
• Collection types: list[T], dict[K, V], tuple[...]
• Optional types: Optional[T], T | None
• Optional patterns: x or default → unwrap_or, ternary with None check (V1.5.0)
• Arithmetic operators: +, -, *, /, //, %, **, @ (V1.3.0)
• Comparison operators: ==, !=, <, >, <=, >=
• Chained comparisons (0 < x < 10 → 0 < x && x < 10) (V1.6.0)
• Logical operators: and, or, not
• Membership operators: in, not in (V1.3.0)
• Identity operators: is, is not (with None comparison)
• Bitwise operators: &, |, ^, ~, <<, >> (V1.3.0)
• Augmented assignment: +=, -=, *=, /=, //=, %=, **=, &=, |=, ^=, <<=, >>= (V1.3.0)
• Docstrings: Triple-quoted strings converted to Rust comments

Control Flow

• If/elif/else statements
• For loops with range(), collection iteration
• While loops
• Break/Continue statements
• Conditional expressions (x if cond else y)
• Early Return handling (refined in v1.2.0)

Functions

• Function definitions with type hints
• Return statements with optional values
• Recursion support
• Nested functions (closure conversion to Rust closures)
• Lambda expressions (lambda x: x + 1)
• Higher-order functions (passing functions as arguments)
• Named arguments (func(name="value"))
• Default argument values (def func(x=10)) (V1.2.0)
• **kwargs (def func(**kwargs) → HashMap<String, Value>) (V1.6.0)

Data Structures

• List literals and operations
• List comprehensions (basic and conditional)
• Dict comprehensions ({k: v for k, v in items}) (V1.3.0)
• Dictionary literals and operations
• Tuple literals and unpacking
• Set literals ({1, 2, 3} → HashSet) (V1.5.0)
• Set comprehensions ({x*2 for x in nums} → HashSet) (V1.6.0)
• Struct definitions (via class syntax)
• Negative indexing (nums[-1])
• Slice notation ([:3], [-3:], [1:n-1])
• Step slicing ([::2], [::-1]) (V1.5.0)
• Index swap (a[i], a[j] = a[j], a[i] → a.swap())
• List copy (l.copy() → l.to_vec()) (V1.2.0)
• Multi-assignment (a, b, c = 1, 2, 3) (V1.3.0)
• List methods: pop, insert, remove, extend, clear (V1.5.0)
• Dict methods: keys, values, get, pop, update (V1.5.0)
• Set methods: add, remove, discard, union, intersection (V1.5.0)

Classes & Objects

• Basic class definitions with __init__
• Instance attributes (self.attr)
• Method definitions
• Static methods (@staticmethod)
• Dataclasses (@dataclass) (V1.2.0 Partial)
• Single inheritance (class Child(Parent)) → Composition (V1.6.0)
• super() calls (super().method() → self.base.method()) (V1.6.0)
• @property decorator → getter/setter methods (V1.6.0)

Resource Management (V1.6.0)

• with statement → RAII scope (with open(...) as f: → { let f = ...; })
• Automatic resource cleanup via Rust’s Drop trait

Resident Python Worker (V1.2.0) 🆕

Tsuchinoko V1.2.0 introduces a Resident Python Worker to support libraries that are difficult to transpile directly to Rust.

• NumPy (import numpy as np)
• Pandas (import pandas as pd)
• OpenCV (import cv2) (V1.4.0)
• SciPy
• Any other library accessible in your Python environment via IPC calls.

from Import Syntax (V1.4.0) 🆕

• Direct function import: from numpy import mean, std
• Automatically converted to py_bridge.call_json("numpy.mean", ...) calls

Persistent Object Handles 🆕

Tsuchinoko now supports persisting Python objects across bridge calls. This allows for:

• Complex object state: Keep DataFrames, NumPy arrays, or custom class instances in memory.
• Method chaining: Call multiple methods on the same object handle.
• Index access: Direct access to Python object elements via handles (df["column"]).
• Handle integration: Seamlessly pass handles back to other Python library functions.

Built-in Functions

• len() - get length
• range() - numeric range iteration
• print() - console output (supports f-string debug "{x=}" / "{:?}")
• list() - convert to list
• min(), max() - min/max values
• abs() - absolute value
• int(), float(), str(), bool() - type conversions
• enumerate() - indexed iteration (V1.3.0)
• zip() - parallel iteration (V1.3.0)
• sorted() - sorted list generation (V1.3.0)
• reversed() - reverse iteration (V1.3.0)
• sum() - sum calculation (V1.3.0)
• all(), any() - boolean check (V1.3.0)
• map(), filter() - functional transformation (V1.3.0)
• assert - assertion statement (V1.3.0)
• input() - user input with optional prompt (V1.5.0)
• round() - rounding with precision (V1.5.0)
• chr(), ord() - character/code point conversion (V1.5.0)
• bin(), hex(), oct() - number format conversion (V1.5.0)
• isinstance() - type checking → DynamicValue enum + match (V1.6.0)

Math Module (V1.3.0 / V1.4.0)

• Functions: math.sqrt, sin, cos, tan, asin, acos, atan, exp, log, log10, log2, abs, floor, ceil, round
• Constants (V1.4.0): math.pi, math.e, math.tau, math.inf, math.nan → Native Rust constants

String Features

• String literals (single/double quotes)
• F-strings (f"Hello {name}")
  • Debug format "{x=}" / "{:?}" supported (V1.2.0)
• String methods: .upper(), .lower(), .strip(), .split(), .join(), etc.
• String methods (V1.5.0): .replace(), .startswith(), .endswith(), .find(), .rfind(), .index(), .count()
• String predicates (V1.5.0): .isdigit(), .isalpha(), .isalnum()

Error Handling

• try/except blocks (converted to catch_unwind)
• try/except with multiple exception types (except (ValueError, TypeError):) (V1.5.0)
• except with variable (except ValueError as e:) (V1.5.0)
• try/except/finally blocks (V1.5.0)
• try/except/else blocks (else runs when no exception) (V1.5.2)
• raise statements (converted to Err(TsuchinokoError) or panic!)
• raise from (raise A from B) - exception chaining with cause (V1.5.2)
• Result type unification - exception functions return Result<T, TsuchinokoError> (V1.5.2)
• Error line numbers - Python source line included in error messages (V1.5.2)
• ValueError, TypeError (converted to TsuchinokoError)

Type System

• Type aliases (MyType = list[int])
• Callable types (Callable[[T], U])
• Function type inference
• Automatic type coercion (Auto-Ref, Auto-Deref, Auto-Clone)
• Type Narrowing (if x is None / if x is not None)

Experimental: PyO3 Direct Integration

[!NOTE] Direct PyO3 calls are still supported but the Resident Worker is recommended for compatibility.

This site is open source. Improve this page.