Native Superset of Python

Jac's Native Superset of Python

Jac is designed as a superset of Python, extending the language with additional features for modern software architecture while maintaining full compatibility with the Python ecosystem. Python developers can leverage their existing knowledge while accessing new capabilities for graph-based and object-spatial programming.

How it Works: Transpilation to Native Python

Jac programs execute on the standard Python runtime without requiring custom runtime environments, virtual machines, or interpreters. The Jac compiler transpiles Jac source code into standard Python through a multi-stage compilation pipeline that generates optimized Python bytecode. This approach provides several advantages:

• Standard Python Runtime: Jac programs execute on the Python runtime, utilizing Python's garbage collector, memory management, and threading model.
• Full Ecosystem Access: All packages on PyPI, internal libraries, and Python development tools are compatible with Jac.
• Readable Output: The transpiled Python code is clean and maintainable, enabling inspection, debugging, and understanding.

The relationship between Jac and Python is analogous to that of TypeScript and JavaScript: a superset language that compiles to a widely-adopted base language.

Example: From Jac to Python

The following Jac module demonstrates functions, objects, and an entrypoint:

"""Functions in Jac."""

def factorial(n: int) -> int {
    if n == 0 { return 1; }
    else { return n * factorial(n-1); }
}

obj Person {
    has name: str;
    has age: int;

    def greet() -> None {
        print(f"Hello, my name is {self.name} and I'm {self.age} years old.");
    }
}

with entry {
    person = Person("John", 42);
    person.greet();
    print(f"5! = {factorial(5)}");
}

The Jac compiler converts this code into the following Python implementation:

"""Functions in Jac."""
from __future__ import annotations
from jaclang.lib import Obj

def factorial(n: int) -> int:
    if n == 0:
        return 1
    else:
        return n * factorial(n - 1)

class Person(Obj):
    name: str
    age: int

    def greet(self) -> None:
        print(f"Hello, my name is {self.name} and I'm {self.age} years old.")

person = Person('John', 42)
person.greet()
print(f'5! = {factorial(5)}')

The compiled output demonstrates how Jac's object-oriented features map to standard Python classes inheriting from Obj (Jac's base object archetype), with imports from the jaclang.lib package.

---

Seamless Interoperability: Import Jac Files Like Python Modules

Jac integrates with Python through a simple import mechanism. By adding import jaclang to Python code, developers can import .jac files using standard Python import statements without requiring build steps, compilation commands, or configuration files.

Key Integration Features:

• Bidirectional Module Imports: Python files can import Jac modules, and Jac files can import Python modules using standard import syntax. Modules written in .jac and .py can be used interchangeably within a project.

• Incremental Adoption: Jac can be added to existing Python projects without restructuring the codebase. Python files can remain unchanged while Jac modules are introduced where beneficial.

• Standard Import Syntax: The same import statements used for Python modules work with .jac files, requiring no special syntax or additional tools.

Example: Importing Across Languages

Consider a Jac module containing graph utilities:

# graph_tools.jac
node Task {
    has name: str;
    has priority: int;
}

This module can be imported in Python using standard import syntax:

# main.py
import jaclang  # Enable Jac imports (one-time setup)
from graph_tools import Task  # Import from .jac file

# Use Jac classes in Python
my_task = Task(name="Deploy", priority=1)

Jac files can also import Python libraries:

# analyzer.jac
import pandas as pd;
import numpy as np;

# Use Python libraries in Jac code

Implementation Details: Jac extends Python's native import mechanism using the PEP 302 import hook system. When import jaclang is executed, it registers a custom importer that enables Python to locate and load .jac files. Subsequently, Python's import mechanism automatically checks for .jac files alongside .py files, compiles them transparently, and loads them into the program. This integration makes Jac modules function as first-class citizens within the Python environment.

---

Five Adoption Patterns: Choose Your Integration Level

Jac provides five adoption strategies that accommodate different development requirements, ranging from pure Python implementations with Jac library support to fully Jac-based applications. The following patterns represent the primary integration approaches:

Pattern Comparison Table

Pattern	Use Case	Jac Content	Python Content	Key Benefits	Example Scenario
1. Pure Jac	New projects, microservices	100%	0%	Full Jac language features, modern syntax	Building a new graph-based application with only .jac files
2. Jac + Inline Python	Inline Python in Jac files	Mixed (::py:: blocks)	Embedded inline	Gradual migration, use Python syntax when needed	.jac files with embedded Python for legacy logic or complex imports
3. Mostly Jac	Import Python modules into Jac	80-95% .jac	5-20% .py	Jac architecture with existing Python utilities	Project with .jac files importing your existing .py utility modules
4. Mostly Python	Import Jac modules into Python	5-20% .jac	80-95% .py	Python codebase with select Jac features	Python project with .py files importing specialized .jac modules for graphs/AI
5. Pure Python + Jac Library	Conservative adoption	0%	100%	No new syntax, just Jac runtime capabilities	Pure .py project using Jac runtime via imports and decorators

graph LR
    A[Pure Jac<br/>100% .jac files] --> B[Jac + Inline Python<br/>::py:: blocks]
    B --> C[Mostly Jac<br/>import .py files]
    C --> D[Mostly Python<br/>import .jac files]
    D --> E[Pure Python<br/>+ Jac Library]

    style A fill:#2d5f3f,stroke:#4CAF50,stroke-width:2px,color:#fff
    style B fill:#3d4f2d,stroke:#8BC34A,stroke-width:2px,color:#fff
    style C fill:#5f4d1a,stroke:#FFC107,stroke-width:2px,color:#fff
    style D fill:#5f3d1a,stroke:#FF9800,stroke-width:2px,color:#fff
    style E fill:#5f2d2d,stroke:#FF5722,stroke-width:2px,color:#fff

---

Pattern Details and Examples

Example Project: The following examples demonstrate a task manager application that stores tasks and generates AI-powered task descriptions.

Core Features: - Task storage with graph-based relationships - Task validation (title length check) - AI-generated task descriptions

Each pattern demonstrates a different approach to implementing this application.

---

Pattern 1: Pure Jac

This pattern uses exclusively .jac files with no Python files required.

Use Case: New projects requiring full access to Jac language features

Directory Structure:

project/
├── main.jac
├── models.jac
└── utils.jac

main.jacmodels.jacutils.jac

"""Main application."""
import models, utils;

walker TaskCreator {
    has title: str;

    can create with `root entry {
        if utils.validate_title(self.title) {
            task = models.Task(title=self.title);
            here ++> task;
            desc = utils.generate_desc(self.title);
            print(f"✓ Created: {task.title}");
            print(f"  AI: {desc}");
        } else {
            print("✗ Title too short!");
        }
    }
}

with entry {
    root spawn TaskCreator(title="Build API");
}

"""Task node definition."""

node Task {
    has title: str;
    has done: bool = False;
}

"""Validation and AI utilities."""

def validate_title(title: str) -> bool {
    return len(title) > 3;
}

def generate_desc(title: str) -> str {
    return f"Task description for: {title}";
}

---

Pattern 2: Jac + Inline Python

This pattern embeds Python code directly within .jac files using ::py:: blocks, enabling the use of Python-specific libraries or preservation of existing Python code.

Use Case: Incremental migration of Python codebases while maintaining legacy utilities

Directory Structure:

project/
├── main.jac
└── models.jac

main.jacmodels.jac

"""Application with inline Python validation."""
import models;

def generate_desc(title: str) -> str {
    return f"Task description for: {title}";
}

::py::
# Legacy Python validation - kept as-is
def validate_title(title):
    """Complex validation logic from old codebase."""
    return len(title) > 3 and title.strip() != ""

def get_sample_task():
    """Helper from legacy code."""
    return {"title": "Build API"}
::py::

walker TaskCreator {
    can create with `root entry {
        # Use inline Python functions
        task_data = get_sample_task();

        if validate_title(task_data["title"]) {
            task = models.Task(title=task_data["title"]);
            here ++> task;
            desc = generate_desc(task.title);
            print(f"✓ Created: {task.title}");
            print(f"  AI: {desc}");
        } else {
            print("✗ Title invalid!");
        }
    }
}

with entry {
    root spawn TaskCreator();
}

"""Task node definition."""

node Task {
    has title: str;
    has done: bool = False;
}

This approach preserves tested Python code while introducing Jac features, supporting incremental migration strategies.

---

Pattern 3: Mostly Jac

This pattern implements the primary application logic in Jac while importing Python utilities from separate .py files.

Use Case: Jac-first development that leverages existing Python utilities or shared modules

Directory Structure:

project/
├── main.jac
├── models.jac
└── validators.py

main.jacmodels.jacvalidators.py

"""Main application - imports Python module."""
import models;
import validators;

def generate_desc(title: str) -> str {
    return f"Task description for: {title}";
}

walker TaskCreator {
    has title: str;

    can create with `root entry {
        # Call Python module functions
        if validators.validate_title(self.title) {
            task = models.Task(title=self.title);
            here ++> task;
            desc = generate_desc(task.title);
            print(f"✓ Created: {task.title}");
            print(f"  AI: {desc}");
        } else {
            print("✗ Title too short!");
        }
    }
}

with entry {
    root spawn TaskCreator(title="Build API");
}

"""Task node definition."""

node Task {
    has title: str;
    has done: bool = False;
}

"""Python validation utilities - shared with other projects."""

def validate_title(title: str) -> bool:
    """Validator used across multiple projects."""
    return len(title) > 3

def get_sample_title():
    """Helper to load sample data."""
    return "Build API"

Jac imports Python modules using standard import mechanisms without requiring configuration.

---

Pattern 4: Mostly Python

This pattern maintains a Python-first application structure while importing .jac modules for graph-based and AI features.

Use Case: Existing Python projects incorporating Jac's graph-native and AI capabilities

Directory Structure:

project/
├── main.py
├── validators.py
└── task_graph.jac

main.pyvalidators.pytask_graph.jac

"""Python application importing Jac modules."""
import jaclang  # Enable Jac imports

from validators import validate_title
from task_graph import Task, TaskCreator, generate_desc
from jaclang.lib import spawn, root

def create_task(title: str):
    """Python function using Jac features."""
    if not validate_title(title):
        print("✗ Title too short!")
        return

    # Use Jac walker
    creator = TaskCreator(title=title)
    spawn(creator, root())

    # Use Jac's AI
    desc = generate_desc(title)
    print(f"  AI: {desc}")

if __name__ == "__main__":
    create_task("Build API")

"""Python validation utilities."""

def validate_title(title: str) -> bool:
    """Title validator."""
    return len(title) > 3

"""Jac module with graph and AI features."""

node Task {
    has title: str;
    has done: bool = False;
}

walker TaskCreator {
    has title: str;

    can create with `root entry {
        task = Task(title=self.title);
        here ++> task;
        print(f"✓ Created: {task.title}");
    }
}

def generate_desc(title: str) -> str {
    return f"Task description for: {title}";
}

This approach maintains familiar Python syntax while providing access to Jac's graph-based and AI features.

---

Pattern 5: Pure Python + Jac Library

This pattern uses pure Python with Jac's runtime as a library, without any .jac files.

Use Case: Conservative adoption paths, teams preferring Python syntax, or existing Python projects

Directory Structure:

project/
├── main.py
└── validators.py

main.pyvalidators.py

"""Pure Python using Jac runtime."""
from jaclang.lib import Node, Walker, on_entry, connect, spawn, root
from validators import validate_title

# Define Task node using Jac base class
class Task(Node):
    title: str
    done: bool

    def __init__(self, title: str):
        super().__init__()
        self.title = title
        self.done = False

# Define walker using Jac decorators
class TaskCreator(Walker):
    def __init__(self, title: str):
        super().__init__()
        self.title = title

    @on_entry
    def create(self, here) -> None:
        """Entry point - creates task."""
        if validate_title(self.title):
            task = Task(title=self.title)
            connect(here, task)
            print(f"✓ Created: {task.title}")
            # Note: AI features require .jac syntax
        else:
            print("✗ Title too short!")

if __name__ == "__main__":
    creator = TaskCreator(title="Build API")
    spawn(creator, root())

"""Python validation utilities."""

def validate_title(title: str) -> bool:
    """Title validator."""
    return len(title) > 3

This pattern provides graph-based capabilities in pure Python without introducing new syntax, utilizing Jac's object-spatial model through library imports.

---

Key Takeaways

Jac's design as a Python superset enables complementary use of both languages rather than requiring a choice between them. Key characteristics include:

• Incremental Adoption: Projects can begin with Pattern 5 (pure Python + Jac library) and progressively adopt Pattern 1 (pure Jac) as requirements evolve
• Full Ecosystem Access: All Python libraries, frameworks, and development tools remain compatible without modification
• Flexible Integration: Five adoption patterns accommodate different team preferences and project requirements
• No Vendor Lock-in: Transpiled Python code is readable and maintainable, providing migration paths if needed
• Transparent Interoperability: PEP 302 import hooks enable seamless bidirectional imports between .jac and .py files

Adoption Pattern	Learning Curve	Migration Effort	Feature Access	Risk Level
Pattern 1: Pure Jac	Higher	Higher	100% Jac features	Low
Pattern 2: Jac + Inline Python	Medium	Low	100% Jac features	Low
Pattern 3: Mostly Jac	Medium-High	Medium	100% Jac features	Low
Pattern 4: Mostly Python	Low-Medium	Low	Select Jac features	Low
Pattern 5: Pure Python + Library	Low	Very Low	Core runtime only	Very Low

Jac accommodates both new application development and enhancement of existing Python codebases, providing structured approaches to graph-based and object-spatial programming while maintaining full Python ecosystem compatibility.