| Name | safe-mol JSON |
| Version |
0.1.13
JSON |
| download |
| home_page | None |
| Summary | Implementation of the 'Gotta be SAFE: a new framework for molecular design' paper |
| upload_time | 2024-11-27 08:25:46 |
| maintainer | None |
| docs_url | None |
| author | None |
| requires_python | >=3.9 |
| license | Apache-2.0 |
| keywords |
safe
smiles
de novo
design
molecules
|
| VCS |
 |
| bugtrack_url |
|
| requirements |
No requirements were recorded.
|
| Travis-CI |
No Travis.
|
| coveralls test coverage |
No coveralls.
|
<h1 align="center"> :safety_vest: SAFE </h1>
<h4 align="center"><b>S</b>equential <b>A</b>ttachment-based <b>F</b>ragment <b>E</b>mbedding (SAFE) is a novel molecular line notation that represents molecules as an unordered sequence of fragment blocks to improve molecule design using generative models.</h4>
</br>
<div align="center">
<img src="docs/assets/safe-tasks.svg" width="100%">
</div>
</br>
<p align="center">
<a href="https://arxiv.org/pdf/2310.10773.pdf" target="_blank">
Paper
</a> |
<a href="https://safe-docs.datamol.io/" target="_blank">
Docs
</a> |
<a href="https://huggingface.co/datamol-io/safe-gpt" target="_blank">
🤗 Model
</a> |
<a href="https://huggingface.co/datasets/datamol-io/safe-gpt" target="_blank">
🤗 Training Dataset
</a>
</p>
---
</br>
[](https://pypi.org/project/safe-mol/)
[](https://anaconda.org/conda-forge/safe-mol)
[](https://pypi.org/project/safe-mol/)
[](https://anaconda.org/conda-forge/safe-mol)
[](https://github.com/datamol-io/safe/blob/main/LICENSE)
[](https://github.com/datamol-io/safe/blob/main/DATA_LICENSE)
[](https://github.com/datamol-io/safe/stargazers)
[](https://github.com/datamol-io/safe/network/members)
[](https://arxiv.org/pdf/2310.10773.pdf)
[](https://github.com/datamol-io/safe/actions/workflows/test.yml)
[](https://github.com/datamol-io/safe/actions/workflows/release.yml)
[](https://github.com/datamol-io/safe/actions/workflows/code-check.yml)
[](https://github.com/datamol-io/safe/actions/workflows/doc.yml)
## Overview of SAFE
SAFE _is the_ deep learning molecular representation. It's an encoding leveraging a peculiarity in the decoding schemes of SMILES, to allow representation of molecules as a contiguous sequence of connected fragments. SAFE strings are valid SMILES strings, and thus are able to preserve the same amount of information. The intuitive representation of molecules as an ordered sequence of connected fragments greatly simplifies the following tasks often encountered in molecular design:
- _de novo_ design
- superstructure generation
- scaffold decoration
- motif extension
- linker generation
- scaffold morphing.
The construction of a SAFE strings requires defining a molecular fragmentation algorithm. By default, we use [BRICS], but any other fragmentation algorithm can be used. The image below illustrates the process of building a SAFE string. The resulting string is a valid SMILES that can be read by [datamol](https://github.com/datamol-io/datamol) or [RDKit](https://github.com/rdkit/rdkit).
</br>
<div align="center">
<img src="docs/assets/safe-construction.svg" width="100%">
</div>
## News 🚀
#### 💥 2024/01/15 💥
1. [@IanAWatson](https://github.com/IanAWatson) has a C++ implementation of SAFE in [LillyMol](https://github.com/IanAWatson/LillyMol/tree/bazel_version_float) that is quite fast and use a custom fragmentation algorithm. Follow the installation instruction on the repo and checkout the docs of the CLI here: [docs/Molecule_Tools/SAFE.md](https://github.com/IanAWatson/LillyMol/blob/bazel_version_float/docs/Molecule_Tools/SAFE.md)
### Installation
You can install `safe` using pip:
```bash
pip install safe-mol
```
You can use conda/mamba:
```bash
mamba install -c conda-forge safe-mol
```
#### 2024/11/22
NOTE: Installation might cause issues like no detection of GPUs (which can be checked by `torch.cuda.is_available()`) and sengmentation error due to mismatch between installed and driver cuda versions. In that case, follow these steps:
Create a new environment using conda:
```bash
conda create -n env_safe python=3.12
conda activate env_safe
```
Check nvidia driver version on machine by running `nvcc --version` or `nvidia-smi` commands
Install pytorch with compatible cuda versions (from `https://pytorch.org/get-started/locally/`) and safe-mol:
```bash
conda install pytorch torchvision torchaudio pytorch-cuda=12.1 -c pytorch -c nvidia
conda install -c conda-forge safe-mol
```
### Datasets and Models
| Type | Name | Infos | Size | Comment |
| ---------------------- | ------------------------------------------------------------------------------ | ---------- | ----- | -------------------- |
| Model | [datamol-io/safe-gpt](https://huggingface.co/datamol-io/safe-gpt) | 87M params | 350M | Default model |
| Training Dataset | [datamol-io/safe-gpt](https://huggingface.co/datasets/datamol-io/safe-gpt) | 1.1B rows | 250GB | Training dataset |
| Drug Benchmark Dataset | [datamol-io/safe-drugs](https://huggingface.co/datasets/datamol-io/safe-drugs) | 26 rows | 20 kB | Benchmarking dataset |
## Usage
Please refer to the [documentation](https://safe-docs.datamol.io/), which contains tutorials for getting started with `safe` and detailed descriptions of the functions provided, as well as an example of how to get started with SAFE-GPT.
### API
We summarize some key functions provided by the `safe` package below.
| Function | Description |
| ------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| `safe.encode` | Translates a SMILES string into its corresponding SAFE string. |
| `safe.decode` | Translates a SAFE string into its corresponding SMILES string. The SAFE decoder just augment RDKit's `Chem.MolFromSmiles` with an optional correction argument to take care of missing hydrogen bonds. |
| `safe.split` | Tokenizes a SAFE string to build a generative model. |
### Examples
#### Translation between SAFE and SMILES representations
```python
import safe
ibuprofen = "CC(Cc1ccc(cc1)C(C(=O)O)C)C"
# SMILES -> SAFE -> SMILES translation
try:
ibuprofen_sf = safe.encode(ibuprofen) # c12ccc3cc1.C3(C)C(=O)O.CC(C)C2
ibuprofen_smi = safe.decode(ibuprofen_sf, canonical=True) # CC(C)Cc1ccc(C(C)C(=O)O)cc1
except safe.EncoderError:
pass
except safe.DecoderError:
pass
ibuprofen_tokens = list(safe.split(ibuprofen_sf))
```
### Training/Finetuning a (new) model
A command line interface is available to train a new model, please run `safe-train --help`. You can also provide an existing checkpoint to continue training or finetune on you own dataset.
For example:
```bash
safe-train --config <path to config> \
--model-path <path to model> \
--tokenizer <path to tokenizer> \
--dataset <path to dataset> \
--num_labels 9 \
--torch_compile True \
--optim "adamw_torch" \
--learning_rate 1e-5 \
--prop_loss_coeff 1e-3 \
--gradient_accumulation_steps 1 \
--output_dir "<path to outputdir>" \
--max_steps 5
```
## References
If you use this repository, please cite the following related [paper](https://arxiv.org/abs/2310.10773#):
```bib
@misc{noutahi2023gotta,
title={Gotta be SAFE: A New Framework for Molecular Design},
author={Emmanuel Noutahi and Cristian Gabellini and Michael Craig and Jonathan S. C Lim and Prudencio Tossou},
year={2023},
eprint={2310.10773},
archivePrefix={arXiv},
primaryClass={cs.LG}
}
```
## License
The training dataset is licensed under CC BY 4.0. See [DATA_LICENSE](DATA_LICENSE) for details. This code base is licensed under the Apache-2.0 license. See [LICENSE](LICENSE) for details.
Note that the model weights of **SAFE-GPT** are exclusively licensed for research purposes (CC BY-NC 4.0).
## Development lifecycle
### Setup dev environment
```bash
mamba create -n safe -f env.yml
mamba activate safe
pip install --no-deps -e .
```
### Tests
You can run tests locally with:
```bash
pytest
```
Raw data
{
"_id": null,
"home_page": null,
"name": "safe-mol",
"maintainer": null,
"docs_url": null,
"requires_python": ">=3.9",
"maintainer_email": null,
"keywords": "safe, smiles, de novo, design, molecules",
"author": null,
"author_email": "Emmanuel Noutahi <emmanuel.noutahi@gmail.com>",
"download_url": "https://files.pythonhosted.org/packages/a9/a9/38d14f78a4d15716a8cbe7bfe4348e14d78a44534ca066fe8df101970368/safe_mol-0.1.13.tar.gz",
"platform": null,
"description": "<h1 align=\"center\"> :safety_vest: SAFE </h1>\n<h4 align=\"center\"><b>S</b>equential <b>A</b>ttachment-based <b>F</b>ragment <b>E</b>mbedding (SAFE) is a novel molecular line notation that represents molecules as an unordered sequence of fragment blocks to improve molecule design using generative models.</h4>\n\n</br>\n<div align=\"center\">\n <img src=\"docs/assets/safe-tasks.svg\" width=\"100%\">\n</div>\n</br>\n\n<p align=\"center\">\n <a href=\"https://arxiv.org/pdf/2310.10773.pdf\" target=\"_blank\">\n Paper\n </a> |\n <a href=\"https://safe-docs.datamol.io/\" target=\"_blank\">\n Docs\n </a> |\n <a href=\"https://huggingface.co/datamol-io/safe-gpt\" target=\"_blank\">\n \ud83e\udd17 Model\n </a> |\n <a href=\"https://huggingface.co/datasets/datamol-io/safe-gpt\" target=\"_blank\">\n \ud83e\udd17 Training Dataset\n </a>\n</p>\n\n---\n\n</br>\n\n[](https://pypi.org/project/safe-mol/)\n[](https://anaconda.org/conda-forge/safe-mol)\n[](https://pypi.org/project/safe-mol/)\n[](https://anaconda.org/conda-forge/safe-mol)\n[](https://github.com/datamol-io/safe/blob/main/LICENSE)\n[](https://github.com/datamol-io/safe/blob/main/DATA_LICENSE)\n[](https://github.com/datamol-io/safe/stargazers)\n[](https://github.com/datamol-io/safe/network/members)\n[](https://arxiv.org/pdf/2310.10773.pdf)\n\n[](https://github.com/datamol-io/safe/actions/workflows/test.yml)\n[](https://github.com/datamol-io/safe/actions/workflows/release.yml)\n[](https://github.com/datamol-io/safe/actions/workflows/code-check.yml)\n[](https://github.com/datamol-io/safe/actions/workflows/doc.yml)\n\n## Overview of SAFE\n\nSAFE _is the_ deep learning molecular representation. It's an encoding leveraging a peculiarity in the decoding schemes of SMILES, to allow representation of molecules as a contiguous sequence of connected fragments. SAFE strings are valid SMILES strings, and thus are able to preserve the same amount of information. The intuitive representation of molecules as an ordered sequence of connected fragments greatly simplifies the following tasks often encountered in molecular design:\n\n- _de novo_ design\n- superstructure generation\n- scaffold decoration\n- motif extension\n- linker generation\n- scaffold morphing.\n\nThe construction of a SAFE strings requires defining a molecular fragmentation algorithm. By default, we use [BRICS], but any other fragmentation algorithm can be used. The image below illustrates the process of building a SAFE string. The resulting string is a valid SMILES that can be read by [datamol](https://github.com/datamol-io/datamol) or [RDKit](https://github.com/rdkit/rdkit).\n\n</br>\n<div align=\"center\">\n <img src=\"docs/assets/safe-construction.svg\" width=\"100%\">\n</div>\n\n## News \ud83d\ude80\n\n#### \ud83d\udca5 2024/01/15 \ud83d\udca5\n1. [@IanAWatson](https://github.com/IanAWatson) has a C++ implementation of SAFE in [LillyMol](https://github.com/IanAWatson/LillyMol/tree/bazel_version_float) that is quite fast and use a custom fragmentation algorithm. Follow the installation instruction on the repo and checkout the docs of the CLI here: [docs/Molecule_Tools/SAFE.md](https://github.com/IanAWatson/LillyMol/blob/bazel_version_float/docs/Molecule_Tools/SAFE.md)\n\n\n### Installation\n\nYou can install `safe` using pip:\n\n```bash\npip install safe-mol\n```\n\nYou can use conda/mamba:\n\n```bash\nmamba install -c conda-forge safe-mol\n```\n\n#### 2024/11/22\nNOTE: Installation might cause issues like no detection of GPUs (which can be checked by `torch.cuda.is_available()`) and sengmentation error due to mismatch between installed and driver cuda versions. In that case, follow these steps:\n\nCreate a new environment using conda:\n\n```bash\nconda create -n env_safe python=3.12\nconda activate env_safe\n```\n\nCheck nvidia driver version on machine by running `nvcc --version` or `nvidia-smi` commands\n\nInstall pytorch with compatible cuda versions (from `https://pytorch.org/get-started/locally/`) and safe-mol:\n\n```bash\nconda install pytorch torchvision torchaudio pytorch-cuda=12.1 -c pytorch -c nvidia\nconda install -c conda-forge safe-mol\n```\n\n### Datasets and Models\n\n| Type | Name | Infos | Size | Comment |\n| ---------------------- | ------------------------------------------------------------------------------ | ---------- | ----- | -------------------- |\n| Model | [datamol-io/safe-gpt](https://huggingface.co/datamol-io/safe-gpt) | 87M params | 350M | Default model |\n| Training Dataset | [datamol-io/safe-gpt](https://huggingface.co/datasets/datamol-io/safe-gpt) | 1.1B rows | 250GB | Training dataset |\n| Drug Benchmark Dataset | [datamol-io/safe-drugs](https://huggingface.co/datasets/datamol-io/safe-drugs) | 26 rows | 20 kB | Benchmarking dataset |\n\n## Usage\n\nPlease refer to the [documentation](https://safe-docs.datamol.io/), which contains tutorials for getting started with `safe` and detailed descriptions of the functions provided, as well as an example of how to get started with SAFE-GPT.\n\n### API\n\nWe summarize some key functions provided by the `safe` package below.\n\n| Function | Description |\n| ------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |\n| `safe.encode` | Translates a SMILES string into its corresponding SAFE string. |\n| `safe.decode` | Translates a SAFE string into its corresponding SMILES string. The SAFE decoder just augment RDKit's `Chem.MolFromSmiles` with an optional correction argument to take care of missing hydrogen bonds. |\n| `safe.split` | Tokenizes a SAFE string to build a generative model. |\n\n### Examples\n\n#### Translation between SAFE and SMILES representations\n\n```python\nimport safe\n\nibuprofen = \"CC(Cc1ccc(cc1)C(C(=O)O)C)C\"\n\n# SMILES -> SAFE -> SMILES translation\ntry:\n ibuprofen_sf = safe.encode(ibuprofen) # c12ccc3cc1.C3(C)C(=O)O.CC(C)C2\n ibuprofen_smi = safe.decode(ibuprofen_sf, canonical=True) # CC(C)Cc1ccc(C(C)C(=O)O)cc1\nexcept safe.EncoderError:\n pass\nexcept safe.DecoderError:\n pass\n\nibuprofen_tokens = list(safe.split(ibuprofen_sf))\n```\n\n### Training/Finetuning a (new) model\n\nA command line interface is available to train a new model, please run `safe-train --help`. You can also provide an existing checkpoint to continue training or finetune on you own dataset.\n\nFor example:\n\n```bash\nsafe-train --config <path to config> \\\n --model-path <path to model> \\\n --tokenizer <path to tokenizer> \\\n --dataset <path to dataset> \\\n --num_labels 9 \\\n --torch_compile True \\\n --optim \"adamw_torch\" \\\n --learning_rate 1e-5 \\\n --prop_loss_coeff 1e-3 \\\n --gradient_accumulation_steps 1 \\\n --output_dir \"<path to outputdir>\" \\\n --max_steps 5\n```\n\n## References\n\nIf you use this repository, please cite the following related [paper](https://arxiv.org/abs/2310.10773#):\n\n```bib\n@misc{noutahi2023gotta,\n title={Gotta be SAFE: A New Framework for Molecular Design},\n author={Emmanuel Noutahi and Cristian Gabellini and Michael Craig and Jonathan S. C Lim and Prudencio Tossou},\n year={2023},\n eprint={2310.10773},\n archivePrefix={arXiv},\n primaryClass={cs.LG}\n}\n```\n\n## License\n\nThe training dataset is licensed under CC BY 4.0. See [DATA_LICENSE](DATA_LICENSE) for details. This code base is licensed under the Apache-2.0 license. See [LICENSE](LICENSE) for details. \n\nNote that the model weights of **SAFE-GPT** are exclusively licensed for research purposes (CC BY-NC 4.0).\n\n## Development lifecycle\n\n### Setup dev environment\n\n```bash\nmamba create -n safe -f env.yml\nmamba activate safe\n\npip install --no-deps -e .\n```\n\n### Tests\n\nYou can run tests locally with:\n\n```bash\npytest\n```\n",
"bugtrack_url": null,
"license": "Apache-2.0",
"summary": "Implementation of the 'Gotta be SAFE: a new framework for molecular design' paper",
"version": "0.1.13",
"project_urls": {
"Bug Tracker": "https://github.com/datamol-io/safe/issues",
"Documentation": "https://safe-docs.datamol.io/",
"Source Code": "https://github.com/datamol-io/safe"
},
"split_keywords": [
"safe",
" smiles",
" de novo",
" design",
" molecules"
],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "a1a575286e08d2ec1937b54e40a762da2b754e47b7eacf6647f23fb576b373e3",
"md5": "4eadd153a5a7a6e97745f139d4c4473a",
"sha256": "50d7e2338921b52d05181699e92859eb315dc0e22dc5cb5be03527ccec14e8fe"
},
"downloads": -1,
"filename": "safe_mol-0.1.13-py3-none-any.whl",
"has_sig": false,
"md5_digest": "4eadd153a5a7a6e97745f139d4c4473a",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": ">=3.9",
"size": 60722,
"upload_time": "2024-11-27T08:25:44",
"upload_time_iso_8601": "2024-11-27T08:25:44.999864Z",
"url": "https://files.pythonhosted.org/packages/a1/a5/75286e08d2ec1937b54e40a762da2b754e47b7eacf6647f23fb576b373e3/safe_mol-0.1.13-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": "",
"digests": {
"blake2b_256": "a9a938d14f78a4d15716a8cbe7bfe4348e14d78a44534ca066fe8df101970368",
"md5": "b1728457548219481b872afc32755247",
"sha256": "9c8ccf0f13e94a28dcf8c0d77f8c3954d2bc805f58b70d8d645055d3bc1b8f9f"
},
"downloads": -1,
"filename": "safe_mol-0.1.13.tar.gz",
"has_sig": false,
"md5_digest": "b1728457548219481b872afc32755247",
"packagetype": "sdist",
"python_version": "source",
"requires_python": ">=3.9",
"size": 507761,
"upload_time": "2024-11-27T08:25:46",
"upload_time_iso_8601": "2024-11-27T08:25:46.691244Z",
"url": "https://files.pythonhosted.org/packages/a9/a9/38d14f78a4d15716a8cbe7bfe4348e14d78a44534ca066fe8df101970368/safe_mol-0.1.13.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2024-11-27 08:25:46",
"github": true,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"github_user": "datamol-io",
"github_project": "safe",
"travis_ci": false,
"coveralls": false,
"github_actions": true,
"lcname": "safe-mol"
}
