AutoDoc AI

Semantic-Aware File Converter & Automatic Documentation Engine

---

AutoDoc AI is designed to revolutionize the way developers maintain and present their codebases by automating the most tedious parts of software development: documentation and reporting.

Objective 1 — Automate Docstring Generation

Leverage local Large Language Models (LLMs) to analyze code context and automatically generate high-quality Python docstrings. The system uses background RPA (Ghost Typing) to insert documentation directly into the active editor or in-place via the web interface.

Objective 2 — Transform Repositories Semantically

Enable high-volume project documentation by processing entire .zip repositories. The engine recursively scans deep directory structures, applies semantic analysis to every module, and produces a fully "documented" version of the codebase while preserving the original logic.

Objective 3 — Professional Automated Reporting

Convert raw source code, Markdown files, or Jupyter Notebooks (.ipynb) into professional PDF and DOCX reports. These reports feature automated AI Summaries that describe complexity, project scope, and learning paths, making them ideal for academic submissions or project handovers.

---

AutoDoc AI Suite
│
├── Backend (app/)                 ← FastAPI core and background services
│   ├── api/                       ← REST API endpoints for all tools
│   ├── services/                  ← LLM, RPA, Parser, and Git processing logic
│   ├── pdf_converter/             ← Native report generation (PDF/DOCX)
│   └── templates/                 ← Jinja2-powered dynamic frontend engine
│
├── Frontend (static/)             ← Advanced CSS/JS client interface
│   └── components/                ← Modular UI cards, bars, and layouts
│
└── Conversion Pipeline            ← Native Markdown to Document flow

Data Flow

Raw Code / .zip Repo
    │
    ▼
[Parser Service]  ──extract──▶  Functions, Classes, & Metadata
    │
    ▼
[LLM Heuristic Engine]  ──analyze──▶  Semantic Context (llama3.2)
    │
    ▼
[DocGen / RPA Service]  ──insert──▶  Documented Code
    │
    ▼
[Document Converter]  ──render──▶  Professional PDF / DOCX

---

Core Backend & AI

Layer	Technology	Purpose
Language	Python 3.10+	Primary application runtime
Web Framework	FastAPI	High-performance asynchronous API
LLM Provider	Ollama	Local hosting of llama3.2 inference model
RPA Engine	PyAutoGUI	Automated "Ghost Typing" of docstrings
PDF Generation	ReportLab	Native high-fidelity PDF rendering
DOCX Generation	python-docx	Native Microsoft Word document assembly

Web Interface & Templating

Layer	Technology	Purpose
Templating	Jinja2	Dynamic HTML generation and partials
Styling	Vanilla CSS	Custom modern dark-themed aesthetics
Interaction	JavaScript	Real-time websocket-style updates and UI logic
Static Files	StaticFiles	Managed delivery of CSS/JS/Image assets

---

combine_auto_doc/
│
├── 📂 app/
│   ├── 📂 api/                # FastAPI routers and endpoints
│   ├── 📂 core/               # App configuration and global settings
│   ├── 📂 pdf_converter/      # PDF/DOCX conversion logic & templates
│   ├── 📂 services/           # LLM, Git, Parser, and RPA services
│   ├── 📂 static/             # CSS components and client-side assets
│   ├── 📂 templates/          # Jinja2 HTML templates for all modes
│   └── main.py                # Main application entry point
│
├── 📂 client/
│   └── listener.py            # Local agent for RPA hotkey listening
│
├── 📂 images/                 # Product screenshots and brand assets
│
├── 📂 merged_app/             # Consolidated logic for standalone deployment
│
├── 📂 project documents/      # External documentation and project briefs
│
├── 📂 temp_repos/             # Working directory for ZIP repo processing
│
├── 📂 tests/                  # Unit and integration testing suites
│
├── requirements.txt           # Python dependency manifest
└── README.md                  # This file

---

Prerequisites

# Python 3.10 or higher required
python --version

# Ollama required (running llama3.2)
ollama run llama3.2

1. Setup Environment

# Clone the repository
git clone https://github.com/Shashwath-K/Python_auto_documenter.git
cd combine_auto_doc

# Install dependencies
pip install -r requirements.txt

2. Launch the Application

# Start the FastAPI server
uvicorn app.main:app --reload

The application will be accessible at http://127.0.0.1:8000.

3. Run the Client Listener (Optional for RPA)

# Start the background hotkey listener
python client/listener.py

---

1. Project Dashboard

The central hub for all documentation modes featuring a professional dark UI.

2. Live Generation Editor

Real-time code analysis with semantic-aware docstring insertion.

3. Repository Documentation

Scan, process, and download entire projects with automated documentation.

---

• v1.0: Live Editor and individual file upload.
• v1.1: Repository Mode (.zip support) and Native PDF/DOCX.
• v1.2: AI-Generated Summary Reports (Complexity & Insights).
• v2.0: Direct VS Code Extension integration.
• v2.1: Support for multi-language LLM analysis (Go, Rust, etc.).

---

Step 1 — Semantic Parsing

services/parser_service.py uses Abstract Syntax Trees (AST) to identify functions, classes, and method signatures that lack documentation.

Step 2 — LLM Generation

The parsed code chunks are sent to the llm_service.py, which prompts the local Ollama instance. It generates docstrings that match the code's logic, parameters, and return types.

Step 3 — RPA Ghost Typing

For real-time insertion, services/rpa_service.py uses keyboard automation to "type" the generated docstrings directly into your source file without manual copy-pasting.

Step 4 — Document Assembly

pdf_converter/ modules take the documented markdown or notebooks and render them into professional reports using distinct theme templates (Classic, Modern, etc.).

---

Distributed under the Unlicense License. See LICENSE.txt for more information.

(back to top)