| Name | ml-copilot-agent JSON |
| Version |
0.1.9
JSON |
| download |
| home_page | None |
| Summary | "It is a Machine Learning Copilot Agent that can execute code for end-to-end ML Cycle" |
| upload_time | 2024-11-03 22:33:54 |
| maintainer | None |
| docs_url | None |
| author | VatsalPatel18 |
| requires_python | <4.0,>=3.9 |
| license | CC-BY-NC-ND-4.0 |
| keywords |
|
| VCS |
|
| bugtrack_url |
|
| requirements |
No requirements were recorded.
|
| Travis-CI |
No Travis.
|
| coveralls test coverage |
No coveralls.
|
# ML-Copilot-Agent
ML-Copilot is an interactive machine learning assistant that streamlines the process of data preprocessing, model training, evaluation, plotting results, and generating documentation—all through a command-line interface powered by OpenAI's GPT 4o.
The framework is build as an llm-agent with llama-index workflow, which is able to execute realtime code through code-intepreter which is present as a tool with the llm-agenmt.
[Full Documentation available on Read the Docs](https://ml-copilot-agent.readthedocs.io/en/latest/)
### How to use ? : watch short video (Recommended)
<!-- ![ML-Copilot-Agent Usage]("assets/ml_copilot_use2_720.gif") -->
[](https://youtu.be/rci7WLu7Lw8)
<!-- [](https://youtube.com/embed/rci7WLu7Lw8) -->
## LLM-Agent-Features
- **List Files**: View files in the current directory.
- **Data Preprocessing**: Preprocess data for binary classification tasks with customizable instructions.
- **Model Training**: Train binary classification models using algorithms like Logistic Regression, SVM, or Random Forest.
- **Model Evaluation**: Evaluate trained models and obtain metrics such as accuracy, precision, recall, F1-score, and AUC.
- **Plotting**: Generate various plots (e.g., bar plots, PCA plots, correlation matrices) from data or evaluation results.
<!-- - **Documentation**: Automatically generate a documentation report summarizing the entire workflow. -->
- **Interactive Workflow**: Seamlessly navigate through different steps with an intuitive command-line interface.
- **Custom Instruction**: Can provide custom instruction to execute code, for example "Perform pca on temp_csv, save the pca results, plot the pca and save plot inside results folder
Please make a conda environment before begin
## Direct Installation
```bash
pip install ml-copilot-agent
```
### Manual Installation
1. **Clone the repository**:
```bash
git clone https://github.com/VatsalPatel18/ml-copilot-agent.git
```
2. **Navigate to the project directory**:
```bash
cd ml-copilot
```
3. **Install the required dependencies**:
```bash
pip install -r requirements.txt
```
Ensure you have Python 3.7 or higher installed on your system.
## Usage
### Download the file for test run
```bash
wget https://raw.githubusercontent.com/VatsalPatel18/ml-copilot-agent/master/temp_csv1.data
```
1. **Get Your OpenAI API Key:**:
To use the OpenAI API, you need to obtain your API key. If you haven't done so yet, follow these steps:
- Go to the [OpenAI API keys page](https://platform.openai.com/account/api-keys)
- Log in to your OpenAI account (or sign up if you don't have one).
- Create a new API key and copy it
Remember to delete the key after use.
2. **Run ML-Copilot**:
```bash
python -m ml_copilot_agent paste-your-openai-api-key
```
Replace `paste-your-openai-api-key` with your actual OpenAI API key.
3. **Interact with ML-Copilot**:
Once started, ML-Copilot will prompt you for commands. You can enter any of the following commands:
- `show files`: Show files in the current directory.
- `preprocess`: Preprocess data for a binary classification task.
- `train`: Train a binary classification model.
- `evaluate`: Evaluate the trained model.
- `plot`: Generate plots from data or evaluation results.
- `document`: Generate a documentation report. (Under Development)
- `exit`: Terminate the workflow.
### Example Workflow
**Step 1: List Files**
```
list files
```
View all files in the current directory to ensure your dataset is available.
**Step 2: Preprocess Data**
```
preprocess
```
- **Dataset Path**: Provide the path to your dataset (e.g., `data/dataset.csv`).
- **Target Column Name**: Specify the name of the target column in your dataset.
- **Save Path**: Choose where to save the preprocessed data (default is `data/preprocessed_data.csv`).
- **Additional Instructions**: (Optional) Add any specific preprocessing instructions (e.g., "use standard scaler").
**Step 3: Train Model**
```
train
```
- **Model Save Path**: Specify where to save the trained model (default is `models/model.pkl`).
- **Additional Instructions**: (Optional) Specify model preferences (e.g., "use SVM classifier").
**Step 4: Evaluate Model**
```
evaluate
```
- **Evaluation Save Path**: Specify where to save the evaluation results (default is `results/evaluation.txt`).
**Step 5: Plot Results**
```
plot
```
- **Data File Path**: Provide the data file path or press Enter to use default evaluation results or preprocessed data.
- **Additional Plotting Instructions**: (Optional) Specify the type of plot (e.g., "make a bar plot of accuracy and precision").
**Step 7: Custom Instructions**
```
custom instruction
```
- Provide any kind of custom instruction that you would like to execute code.
**Step 8: Exit**
```
exit
```
Terminate the workflow when you are done.
## Dependencies
- **Python 3.7 or higher**
- **OpenAI GPT Models**
- **LlamaIndex**
- **Pandas**
- **Scikit-learn**
- **Matplotlib**
- **Seaborn**
Install all dependencies using:
```bash
pip install -r requirements.txt
```
## Project Structure
- `ml_copilot/`
- `__init__.py`: Initialization and configuration.
- `__main__.py`: Entry point of the application.
- `workflow.py`: Defines the MLWorkflow class and all associated steps and events.
- `data/`: Directory where preprocessed data is saved.
- `models/`: Directory where trained models are saved.
- `results/`: Directory where evaluation results and plots are saved.
- `reports/`: Directory where documentation reports are saved.
- `requirements.txt`: Contains all Python dependencies.
## Contributing
Contributions are welcome! Please follow these steps:
1. **Fork the repository**.
2. **Create a new branch** for your feature or bug fix.
3. **Commit your changes** with clear and descriptive messages.
4. **Push to your fork** and submit a pull request.
## License
This project is licensed under the CC-BY-NC-ND-4.0 License.
## Acknowledgments
- Thanks to the developers of OpenAI and LlamaIndex for providing the foundational tools that make this project possible.
## Contact
For any questions or suggestions, feel free to open an issue or contact vatsal1804@gmail.com.
Raw data
{
"_id": null,
"home_page": null,
"name": "ml-copilot-agent",
"maintainer": null,
"docs_url": null,
"requires_python": "<4.0,>=3.9",
"maintainer_email": null,
"keywords": null,
"author": "VatsalPatel18",
"author_email": "vatsal1804@gmail.com",
"download_url": "https://files.pythonhosted.org/packages/c3/d1/a8696454702ef73d3d792c80987deff69c2d4e2e46232827f1e7d81e7ba8/ml_copilot_agent-0.1.9.tar.gz",
"platform": null,
"description": "# ML-Copilot-Agent\n\nML-Copilot is an interactive machine learning assistant that streamlines the process of data preprocessing, model training, evaluation, plotting results, and generating documentation\u2014all through a command-line interface powered by OpenAI's GPT 4o.\n\nThe framework is build as an llm-agent with llama-index workflow, which is able to execute realtime code through code-intepreter which is present as a tool with the llm-agenmt. \n\n[Full Documentation available on Read the Docs](https://ml-copilot-agent.readthedocs.io/en/latest/)\n\n### How to use ? : watch short video (Recommended)\n<!-- ![ML-Copilot-Agent Usage](\"assets/ml_copilot_use2_720.gif\") -->\n[](https://youtu.be/rci7WLu7Lw8)\n<!-- [](https://youtube.com/embed/rci7WLu7Lw8) -->\n\n## LLM-Agent-Features\n\n- **List Files**: View files in the current directory.\n- **Data Preprocessing**: Preprocess data for binary classification tasks with customizable instructions.\n- **Model Training**: Train binary classification models using algorithms like Logistic Regression, SVM, or Random Forest.\n- **Model Evaluation**: Evaluate trained models and obtain metrics such as accuracy, precision, recall, F1-score, and AUC.\n- **Plotting**: Generate various plots (e.g., bar plots, PCA plots, correlation matrices) from data or evaluation results.\n<!-- - **Documentation**: Automatically generate a documentation report summarizing the entire workflow. -->\n- **Interactive Workflow**: Seamlessly navigate through different steps with an intuitive command-line interface.\n- **Custom Instruction**: Can provide custom instruction to execute code, for example \"Perform pca on temp_csv, save the pca results, plot the pca and save plot inside results folder \n\nPlease make a conda environment before begin\n\n## Direct Installation\n\n```bash\npip install ml-copilot-agent\n```\n\n### Manual Installation\n\n1. **Clone the repository**:\n```bash\ngit clone https://github.com/VatsalPatel18/ml-copilot-agent.git\n```\n\n2. **Navigate to the project directory**:\n```bash\ncd ml-copilot\n```\n\n3. **Install the required dependencies**:\n```bash\npip install -r requirements.txt\n```\nEnsure you have Python 3.7 or higher installed on your system.\n\n\n## Usage\n\n### Download the file for test run \n```bash\nwget https://raw.githubusercontent.com/VatsalPatel18/ml-copilot-agent/master/temp_csv1.data\n```\n\n1. **Get Your OpenAI API Key:**:\n\nTo use the OpenAI API, you need to obtain your API key. If you haven't done so yet, follow these steps:\n- Go to the [OpenAI API keys page](https://platform.openai.com/account/api-keys)\n- Log in to your OpenAI account (or sign up if you don't have one).\n- Create a new API key and copy it\n\nRemember to delete the key after use. \n\n2. **Run ML-Copilot**:\n```bash\npython -m ml_copilot_agent paste-your-openai-api-key\n```\n\nReplace `paste-your-openai-api-key` with your actual OpenAI API key.\n\n3. **Interact with ML-Copilot**:\n\nOnce started, ML-Copilot will prompt you for commands. You can enter any of the following commands:\n\n- `show files`: Show files in the current directory.\n- `preprocess`: Preprocess data for a binary classification task.\n- `train`: Train a binary classification model.\n- `evaluate`: Evaluate the trained model.\n- `plot`: Generate plots from data or evaluation results.\n- `document`: Generate a documentation report. (Under Development)\n- `exit`: Terminate the workflow.\n\n### Example Workflow\n\n**Step 1: List Files**\n\n```\nlist files\n```\n\nView all files in the current directory to ensure your dataset is available.\n\n**Step 2: Preprocess Data**\n```\npreprocess\n```\n\n- **Dataset Path**: Provide the path to your dataset (e.g., `data/dataset.csv`).\n- **Target Column Name**: Specify the name of the target column in your dataset.\n- **Save Path**: Choose where to save the preprocessed data (default is `data/preprocessed_data.csv`).\n- **Additional Instructions**: (Optional) Add any specific preprocessing instructions (e.g., \"use standard scaler\").\n\n**Step 3: Train Model**\n```\ntrain\n```\n\n- **Model Save Path**: Specify where to save the trained model (default is `models/model.pkl`).\n- **Additional Instructions**: (Optional) Specify model preferences (e.g., \"use SVM classifier\").\n\n**Step 4: Evaluate Model**\n```\nevaluate\n```\n\n- **Evaluation Save Path**: Specify where to save the evaluation results (default is `results/evaluation.txt`).\n\n**Step 5: Plot Results**\n\n```\nplot\n```\n\n- **Data File Path**: Provide the data file path or press Enter to use default evaluation results or preprocessed data.\n- **Additional Plotting Instructions**: (Optional) Specify the type of plot (e.g., \"make a bar plot of accuracy and precision\").\n\n**Step 7: Custom Instructions**\n\n```\ncustom instruction\n```\n\n- Provide any kind of custom instruction that you would like to execute code.\n\n**Step 8: Exit**\n\n```\nexit\n```\n\nTerminate the workflow when you are done.\n\n## Dependencies\n\n- **Python 3.7 or higher**\n- **OpenAI GPT Models**\n- **LlamaIndex**\n- **Pandas**\n- **Scikit-learn**\n- **Matplotlib**\n- **Seaborn**\n\nInstall all dependencies using:\n```bash\npip install -r requirements.txt\n```\n## Project Structure\n\n- `ml_copilot/`\n- `__init__.py`: Initialization and configuration.\n- `__main__.py`: Entry point of the application.\n- `workflow.py`: Defines the MLWorkflow class and all associated steps and events.\n- `data/`: Directory where preprocessed data is saved.\n- `models/`: Directory where trained models are saved.\n- `results/`: Directory where evaluation results and plots are saved.\n- `reports/`: Directory where documentation reports are saved.\n- `requirements.txt`: Contains all Python dependencies.\n\n## Contributing\n\nContributions are welcome! Please follow these steps:\n\n1. **Fork the repository**.\n2. **Create a new branch** for your feature or bug fix.\n3. **Commit your changes** with clear and descriptive messages.\n4. **Push to your fork** and submit a pull request.\n\n## License\n\nThis project is licensed under the CC-BY-NC-ND-4.0 License.\n\n## Acknowledgments\n\n- Thanks to the developers of OpenAI and LlamaIndex for providing the foundational tools that make this project possible.\n\n## Contact\n\nFor any questions or suggestions, feel free to open an issue or contact vatsal1804@gmail.com.\n\n\n",
"bugtrack_url": null,
"license": "CC-BY-NC-ND-4.0",
"summary": "\"It is a Machine Learning Copilot Agent that can execute code for end-to-end ML Cycle\"",
"version": "0.1.9",
"project_urls": null,
"split_keywords": [],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "6ee6c27580c4eafe339c9c6952ed68a5c09b45a0cd076eac9a849155548097e7",
"md5": "ddef0d50d9e2ab154a234054b1062ffa",
"sha256": "828417f226f2ec637ab3e3d8a2d634ccb32227c71376f4ea2524b3d8ebc1ce64"
},
"downloads": -1,
"filename": "ml_copilot_agent-0.1.9-py3-none-any.whl",
"has_sig": false,
"md5_digest": "ddef0d50d9e2ab154a234054b1062ffa",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": "<4.0,>=3.9",
"size": 14192,
"upload_time": "2024-11-03T22:33:51",
"upload_time_iso_8601": "2024-11-03T22:33:51.853713Z",
"url": "https://files.pythonhosted.org/packages/6e/e6/c27580c4eafe339c9c6952ed68a5c09b45a0cd076eac9a849155548097e7/ml_copilot_agent-0.1.9-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": "",
"digests": {
"blake2b_256": "c3d1a8696454702ef73d3d792c80987deff69c2d4e2e46232827f1e7d81e7ba8",
"md5": "014ee8ef2dc8fee92c0d5dae78c4ae38",
"sha256": "2f33342f8c1de1d9870c33a72ea331cf1f0a253c13f76d4a7607f49deab9eeae"
},
"downloads": -1,
"filename": "ml_copilot_agent-0.1.9.tar.gz",
"has_sig": false,
"md5_digest": "014ee8ef2dc8fee92c0d5dae78c4ae38",
"packagetype": "sdist",
"python_version": "source",
"requires_python": "<4.0,>=3.9",
"size": 14909,
"upload_time": "2024-11-03T22:33:54",
"upload_time_iso_8601": "2024-11-03T22:33:54.348626Z",
"url": "https://files.pythonhosted.org/packages/c3/d1/a8696454702ef73d3d792c80987deff69c2d4e2e46232827f1e7d81e7ba8/ml_copilot_agent-0.1.9.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2024-11-03 22:33:54",
"github": false,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"lcname": "ml-copilot-agent"
}
