Metadata-Version: 2.4
Name: identa-agent
Version: 2.0.0
Summary: Cross-model prompt migration CLI tool
Project-URL: Homepage, https://github.com/srodriguez/identa-agent
Project-URL: Repository, https://github.com/srodriguez/identa-agent
Project-URL: Issues, https://github.com/srodriguez/identa-agent/issues
Project-URL: Changelog, https://github.com/srodriguez/identa-agent/blob/main/CHANGELOG.md
Author: Identa Author
License: Apache-2.0
License-File: LICENSE
Keywords: calibration,llm,migration,nlp,prompt-engineering
Classifier: Development Status :: 4 - Beta
Classifier: Intended Audience :: Developers
Classifier: License :: OSI Approved :: Apache Software License
Classifier: Operating System :: OS Independent
Classifier: Programming Language :: Python :: 3.11
Classifier: Programming Language :: Python :: 3.12
Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
Requires-Python: >=3.11
Requires-Dist: aiohttp>=3.9
Requires-Dist: alembic>=1.13
Requires-Dist: anthropic>=0.35
Requires-Dist: litellm>=1.40
Requires-Dist: openai>=1.40
Requires-Dist: pydantic>=2.0
Requires-Dist: rich>=13.0
Requires-Dist: sqlalchemy>=2.0
Requires-Dist: tomli>=2.0
Requires-Dist: typer>=0.12
Provides-Extra: dev
Requires-Dist: mypy>=1.10; extra == 'dev'
Requires-Dist: pytest-asyncio>=0.23; extra == 'dev'
Requires-Dist: pytest-cov>=5.0; extra == 'dev'
Requires-Dist: pytest>=8.0; extra == 'dev'
Requires-Dist: ruff>=0.5; extra == 'dev'
Provides-Extra: dspy
Requires-Dist: dspy-ai>=2.4; extra == 'dspy'
Provides-Extra: google
Requires-Dist: google-genai>=0.5.0; extra == 'google'
Provides-Extra: langchain
Requires-Dist: langchain-core>=0.2; extra == 'langchain'
Provides-Extra: mistral
Requires-Dist: mistralai>=1.0; extra == 'mistral'
Description-Content-Type: text/markdown

# Identa — Cross-Model Prompt Migration CLI (v2.0.0)

Identa is a powerful tool for managing, migrating, and calibrating LLM prompts across different model architectures. It implements the research-backed **PromptBridge** methodology and the **MAP-RPE** iterative refinement engine.

## 🚀 Key Features

*   **Dual-Mode Calibration (MAP-RPE)**: Optimize prompts using **Static** (zero-cost) or **Agent** (dynamic LLM-generated) question sets.
*   **PromptBridge Migration**: Automated cross-model transfer using mapping extraction and adapter layers.
*   **Statistical Analysis**: Measure the "Transfer Gap" (drift) between models with standard deviation and risk metrics.
*   **Persistent Storage**: Automatic SQLite-based prompt versioning, cross-model history, and calibration results.
*   **Multi-Provider Support**: Integrated adapters for **OpenAI**, **Anthropic**, **Google**, **OpenRouter**, **Mistral**, and **Ollama**.
*   **Terminal Native**: Rich logging, tables, and progress monitoring for long-running evolutionary runs.

## 🧠 Research Basis
Based on *PromptBridge: Cross-Model Transfer of LLM Prompts* (arXiv:2512.01420v1) — Accenture / UC Santa Cruz.

## 📦 Installation

```bash
pip install -e .
```

## 🛠 Quick Start

1.  **Configure Environment**: Create a `.env` file with your API keys:
    ```bash
    OPENAI_API_KEY=sk-...
    OPENROUTER_API_KEY=sk-or-...
    ```

2.  **Run Calibration**:
    ```bash
    # Optimize a prompt for gpt-4o using the Software Developer domain
    identa -v calibrate run --domain software_developer --target gpt-4o
    ```

3.  **Migrate Prompts**:
    ```bash
    # Transfer a prompt from gpt-4o to claude-3-5-sonnet and save to a file
    identa migrate --input my-coding-prompt.txt --from gpt-4o --to claude-3-5-sonnet -o results/migrated.txt
    ```

4.  **Inspect Store**:
    ```bash
    # List all successful migrations and calibrations
    identa store list
    ```

## ⚙️ Configuration (`identa.toml`)

You can set project-wide defaults to avoid repetitive CLI flags:

```toml
[calibration]
target_model = "gpt-4o"
reflection_model = "openrouter/google/gemini-2.0-flash-001"
calibration_questions = 20
```

## 📖 Documentation
- [Quickstart Guide](docs/quickstart.md)
- [Calibration Modes (Static vs Agent)](docs/CALIBRATION_MODES.md)
- [Project Changelog](CHANGELOG.md)

---
*Created for the research community and prompt engineering teams.*
