Example: Basic Parsing & Traversal

Last Updated March 27, 2026

The foundation of the Omni-MDX Python engine is its ability to parse Markdown and JSX at lightning speed while keeping the memory footprint near zero. This example demonstrates how to load a document and iterate through its native AST nodes.

ℹ️ Information

Full Source Code: Clone and test this environment directly from omni-mdx-sandbox/python/basic-parsing.

Extracting Node Information

Imagine you want to parse a simple document and isolate specific JSX alerts to check their severity.

1. The Source MDX File

You have a standard MDX string containing custom components:

mdx

# System Status

The system is currently operating normally.


  Database connection **lost** at 10:42 PM.

2. The Python Extraction Script

On the Python side, you use the omni_mdx.parse() function. Because of the Zero-Copy architecture, the attributes are instantly available as a native Python dictionary.

python

import omni_mdx

def main():
    # 1. Parse the document (returns a pointer to Rust memory)
    document_source = """# System Status

The system is currently operating normally.


  Database connection **lost** at 10:42 PM.

"""
    ast = omni_mdx.parse(document_source)
    
    # 2. Filter specific JSX nodes
    # In omni_mdx, JSX component nodes use the component tag name (e.g. "Alert") as node_type.
    alerts = [node for node in ast.nodes if node.node_type == "Alert"]
    
    for alert in alerts:
        # 3. Access the native dictionary for attributes
        attrs = alert.attributes or {}
        level = attrs.get("level", "unknown")
        is_active = "active" in attrs
        
        # 4. Use text_content() to recursively strip Markdown tags (like **lost**)
        clean_message = alert.text_content().strip()
        
        status = "ONGOING" if is_active else "RESOLVED"
        print(f"[{status}] {level.upper()}: {clean_message}")

# Output: [ONGOING] CRITICAL: Database connection lost at 10:42 PM.

By leveraging the AST, you bypass the need for fragile regular expressions, accessing your document’s semantic structure safely and efficiently.

Example: Basic Parsing & Traversal

Last Updated March 27, 2026

ℹ️ Information

Full Source Code: Clone and test this environment directly from omni-mdx-sandbox/python/basic-parsing.

Extracting Node Information

Imagine you want to parse a simple document and isolate specific JSX alerts to check their severity.

1. The Source MDX File

You have a standard MDX string containing custom components:

mdx

# System Status

The system is currently operating normally.


  Database connection **lost** at 10:42 PM.

2. The Python Extraction Script

On the Python side, you use the omni_mdx.parse() function. Because of the Zero-Copy architecture, the attributes are instantly available as a native Python dictionary.

python

import omni_mdx

def main():
    # 1. Parse the document (returns a pointer to Rust memory)
    document_source = """# System Status

The system is currently operating normally.


  Database connection **lost** at 10:42 PM.

"""
    ast = omni_mdx.parse(document_source)
    
    # 2. Filter specific JSX nodes
    # In omni_mdx, JSX component nodes use the component tag name (e.g. "Alert") as node_type.
    alerts = [node for node in ast.nodes if node.node_type == "Alert"]
    
    for alert in alerts:
        # 3. Access the native dictionary for attributes
        attrs = alert.attributes or {}
        level = attrs.get("level", "unknown")
        is_active = "active" in attrs
        
        # 4. Use text_content() to recursively strip Markdown tags (like **lost**)
        clean_message = alert.text_content().strip()
        
        status = "ONGOING" if is_active else "RESOLVED"
        print(f"[{status}] {level.upper()}: {clean_message}")

# Output: [ONGOING] CRITICAL: Database connection lost at 10:42 PM.

By leveraging the AST, you bypass the need for fragile regular expressions, accessing your document’s semantic structure safely and efficiently.