nltx 0.2.0

Offline NLP/NLT lab reference library with packaged JSON datasets

nltx

nltx is an offline Python package for serving lab content converted from raw Jupyter notebooks into structured JSON.

Install

From the project root:

pip install .

After publishing to PyPI:

pip install nltx

Usage

from nltx import (
    details,
    cell,
    description,
    working,
    explain,
    guide,
    inputGuide,
    cautions,
    help,
)

print(details("lab1"))
print(cell("lab1", 1, "A"))
print(description("lab1", "A"))      # variant-level description
print(working("lab1", "A"))          # variant-level working mechanism
print(description("lab1", 1, "A"))   # optional cell-level description (backward compatible)
print(working("lab1", 1, "A"))       # optional cell-level working (backward compatible)
print(explain("lab1", 1, "A"))
print(guide("lab1", "A"))
print(inputGuide("lab1", "A"))
print(cautions("lab1", "A"))
print(help())

Lab Data Structure

Each lab JSON file (nltx/data/<lab>.json) uses:

{
  "title": "<lab-name>",
  "variants": {
    "A": {
      "title": "Default variant",
      "cells": [
        {
          "code": "<code>",
          "description": "<what this cell does>",
          "working": "<how this cell works>",
          "explanation": "<combined description + working>",
          "metadata": {}
        }
      ],
      "description": "<variant-level summary>",
      "differenceFromOtherVariants": "<how this variant differs from sibling variants>",
      "workingMechanism": "<step-by-step variant execution summary>",
      "cellWorkings": [
        {
          "cellNumber": 1,
          "description": "<what cell does>",
          "working": "<how cell works>"
        }
      ],
      "guide": "",
      "inputGuide": "",
      "cautions": [],
      "sourceNotebook": "raw/lab1A.ipynb",
      "introMarkdown": "",
      "rawNotebookMetadata": {}
    }
  }
}

Data Workflow

• Notebook files are maintained in the project root raw/ directory.
• JSON files in nltx/data/ are curated manually from those notebooks.
• The package reads only packaged JSON data at runtime.
• The project does not use an in-package converter module.

Current Scope

• Current packaged labs: lab1, lab2, lab3, lab4, lab5, lab7, lab8
• Multiple variants are supported per lab (A/B/C based on availability)
• Markdown parsing and advanced enrichment are intentionally out of scope for now