Peng task2

[PengLin-Zilv]

Summary

Implements task 2 as described in https://sumailsyr-my.sharepoint.com/:w:/g/personal/jpatil01_syr_edu/IQARlfe48KDiRrHCOmFnAYmHAV98Gnc9SjifRRNSky2_5zE?e=HC7FGk

Add a tutorial and example demonstrating how to convert natural language paragraphs into PyReason facts and rules using an open source local LLM.

Changes

• New tutorial: docs/source/tutorials/natural_language_to_pyreason.rst
• New example: examples/natural_language_to_pyreason_ex.py
• Updated tutorial index to include the new tutorial

Pipeline

User input structured paragraph

• Step 1: LLM extracts facts/rules from English paragraph
• Step 2: LLM converts English to PyReason syntax
• Step 3: Parser extracts structured facts/rules
• Step 4: Validate each rule with pr.Rule()

Testing

Tested with different models, chose qwen3:14b for its speed and stability on four paragraphs spanning education, medical,
legal, animal, and relational (edge rule) scenarios. All pass rule validation.

Outputs: for different structured paragraphs, each tested 2+ times to ensure the stability

Test 1: study/school

Test 2: Medical

Test 3: Legal

Test 4: Animal

Test 5: Workforce Relational (Edge Rule)

[ColtonPayne]

LGTM