talkcode 0.1.0

talkcode: Talk to your codebase using AI and static analysis.

Talkcode

Talk to your legacy Python codebase using AI and static analysis.
talkcode helps developers explore, understand, and query large codebases using natural language and intelligent indexing.

🚀 What Is It?

talkcode is a command-line tool that combines static analysis with AI to make legacy Python codebases more accessible. Whether you're onboarding a new team member or reverse-engineering a complex system, talkcode lets you ask questions like:

• “Where is process_order() used?”
• “What does UserManager depend on?”
• “Show me all functions that modify global state.”

---

🔍 Features

• 📦 Static analysis of Python codebases
• 🔍 Function and class indexing
• 🔁 Call flow tracing
• 🔑 Keyword-based code search
• 🤖 Optional AI integration for natural language Q&A
• 🖥️ Terminal UI powered by [textual]

---

📦 Installation

pip install talkcode

---

🛠️ Usage

talkcode init
talkcode index --path /path/to/codebase
talkcode chat --ai
talkcode ui --ai

---

🤖 AI Integration

Enable AI-powered answers with:

talkcode chat --ai

Supported Providers

• ✅ Azure OpenAI

Coming Soon

• 🔜 Google Gemini
• 🔜 Anthropic Claude

To configure Azure OpenAI, create a config file at:

# ~/.talkcode/config.toml

[azure]
provider = "azure"
endpoint = "https://your-resource-name.openai.azure.com/"
deployment = "your-model-deployment-name"
api_version = "2023-07-01-preview"
api_key = "your-azure-openai-api-key"

---

🖥️ Textual UI

Launch the interactive terminal UI:

talkcode ui

Enable AI in the UI:

talkcode ui --ai

📄 License

This project is licensed under the MIT License.
You are free to use, modify, and distribute this software with proper attribution.

---

💬 Example Responses

Here are anonymized examples of the kinds of insights talkcode can generate:

🔗 Dependencies

This module depends on:
  - A secrets manager client for secure credential access
  - A GPT-based AI interface for natural language processing

📡 Downstream Consumers

This component is invoked by:
  - A markdown conversion utility that formats analysis results
  - An evaluation module that scores AI-generated outputs

🧠 AI-Powered Answers

Q: What does the authentication handler rely on?
A: It depends on a token validator, a configuration loader, and a secrets manager.

Q: Which modules call the data transformation pipeline?
A: The pipeline is triggered by the ingestion service and the reporting engine.