Metadata-Version: 2.1
Name: autonml
Version: 0.1.10
Summary: AutonML : CMU's AutoML System
Home-page: UNKNOWN
Author: Saswati Ray, Andrew Williams, Vedant Sanil
Maintainer: Andrew Williams, Vedant Sanil
Maintainer-email: awillia2@andrew.cmu.edu, vsanil@andrew.cmu.edu
License: Apache-2.0
Keywords: datadrivendiscovery,automl,d3m,ta2,cmu
Platform: UNKNOWN
Classifier: Development Status :: 2 - Pre-Alpha
Classifier: License :: OSI Approved :: Apache Software License
Classifier: Topic :: Scientific/Engineering
Requires-Python: >=3.6
Description-Content-Type: text/markdown
License-File: LICENSE
License-File: NOTICE.md

<img src="https://gitlab.com/sray/cmu-ta2/-/raw/master/docs/img/AutonML_logo.png?inline=false" width=30%>


# CMU TA2 (Built using DARPA D3M ecosystem)

Auto<sup>n</sup> ML is an automated machine learning system developed by CMU Auton Lab 
to power data scientists with efficient model discovery and advanced data analytics. 
Auton ML also powers the D3M Subject Matter Expert (SME) User Interfaces such as Two Ravens http://2ra.vn/.

**Taking your machine learning capacity to the nth power.**

  <img src="https://gitlab.com/sray/cmu-ta2/-/raw/master/docs/img/model_pipeline.png?inline=false" width="869" height="489">

We provide a documentation listing the complete set of tasks, data modalities, machine learning models and future supported tasks provided by AutonML [here](https://gitlab.com/sray/cmu-ta2/-/blob/dev/docs/SUPPORTED.md).

### Installation
AutonML can be installed as: `pip install autonml`. We recommend this installation be done in a new virtual environment or conda environment.

Recommended steps to install `autonml`:
```bash
pip install autonml
pip install d3m-common-primitives d3m-sklearn-wrap sri-d3m rpi-d3m-primitives dsbox-primitives dsbox-corex distil-primitives d3m-esrnn d3m-nbeats 
pip install kf-d3m-primitives
```
This installation may take time to complete, owing to the fact that pip's dependecy resolvers may take time resolving potential package conflicts. To make installation faster, you can add pip's legacy resolver as `--use-deprecated=legacy-resolver`. <span style="color:red"> **Caution**: using old resolvers may present unresolved package conflicts.</span>

### D3M dataset
- Any dataset to be used should be in D3M dataset format (directory structure with TRAIN, TEST folders and underlying .json files).
- Example available of a single dataset [here](https://datasets.datadrivendiscovery.org/d3m/datasets/-/tree/master/seed_datasets_current/185_baseball_MIN_METADATA)
- More datasets available [here](https://datasets.datadrivendiscovery.org/d3m/datasets/-/tree/master/seed_datasets_current/)
- Any non-D3M data can be converted to D3M dataset. (See section below on "Convert raw dataset to D3M dataset").

### Run the AutonML pipeline

We can run the AutonML pipeline in two ways. It can be run as a standalone CLI command, accessed via the `autonml_main` command. This command takes five arguments, listed below:
- Path to the data directory (must be in D3M format)
- Output directory where results are to be stored. This directory will be dynamically created if it does not exist.
- Timeout (measured in minutes)
- Number of CPUs to be used (minimum: 4 cores, recommended: 8 cores)
- Path to `problemDoc.json` (see example below)

```bash
INPUT_DIR=/home/<user>/d3m/datasets/185_baseball_MIN_METADATA
OUTPUT_DIR=/output
TIMEOUT=2
NUMCPUS=8
PROBLEMPATH=${INPUT_DIR}/TRAIN/problem_TRAIN/problemDoc.json

autonml_main ${INPUT_DIR} ${OUTPUT_DIR} ${TIMEOUT} ${NUMCPUS} ${PROBLEMPATH} 
```


The above script will do the following-
1. Run search for best pipelines for the specified dataset using TRAIN data.
2. JSON pipelines (with ranks) will be output in JSON format at /output/<search_dir>/pipelines_ranked/
3. CSV prediction files of the pipelines trained on TRAIN data and predicted on TEST data will be available at /output/<search_dir>/predictions/
4. Training data predictions (cross-validated mostly) are produced in the current directory as /output/<search_dir>/training_predictions/<pipeline_id>_train_predictions.csv.
5. Python code equivalent of executing a JSON pipeline on a dataset produced at /output/<search_dir>/executables/

 An example -
```bash
OUTPUT_DIR=output

python ${OUTPUT_DIR}/99211bc3-638a-455b-8d48-0dadc0bf1f10/executables/19908fd3-706a-48da-b13c-dc13da0ed3cc.code.py ${OUTPUT_DIR}/ ${OUTPUT_DIR}/99211bc3-638a-455b-8d48-0dadc0bf1f10/predictions/19908fd3-706a-48da-b13c-dc13da0ed3cc.predictions.csv
```

### Convert raw dataset to D3M dataset

#### D3M dataset
- Any dataset to be used should be in D3M dataset format (directory structure with TRAIN, TEST folders and underlying .json files).
- Example available of a single dataset [here](https://datasets.datadrivendiscovery.org/d3m/datasets/-/tree/master/seed_datasets_current/185_baseball_MIN_METADATA)
- More datasets available [here](https://datasets.datadrivendiscovery.org/d3m/datasets/-/tree/master/seed_datasets_current/)
- Any non-D3M data can be converted to D3M dataset. (See section below on "Convert raw dataset to D3M dataset").

#### [Convert raw dataset to D3M dataset](https://gitlab.com/sray/cmu-ta2/-/blob/master/convert_raw_dataset.rst)

If not done already, run `pip install autonml` before our raw dataset converter.
```bash
create_d3m_dataset <train_data.csv> <test_data.csv> <label> <metric> -t classification <-t ...>
```

Detailed description of dataset type(s), task type(s) and metrics provided **[here](https://gitlab.com/sray/cmu-ta2/-/blob/master/convert_raw_dataset.rst).**


