{"id":13994047,"url":"https://github.com/juanmc2005/diart","last_synced_at":"2025-05-14T01:11:11.020Z","repository":{"id":36954216,"uuid":"394292135","full_name":"juanmc2005/diart","owner":"juanmc2005","description":"A python package to build AI-powered real-time audio applications","archived":false,"fork":false,"pushed_at":"2025-02-12T13:25:47.000Z","size":36494,"stargazers_count":1270,"open_issues_count":52,"forks_count":99,"subscribers_count":18,"default_branch":"main","last_synced_at":"2025-05-03T05:34:06.998Z","etag":null,"topics":["deep-learning","real-time","speaker-diarization","speaker-embedding","streaming-audio","transcription","voice-activity-detection"],"latest_commit_sha":null,"homepage":"https://diart.readthedocs.io","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/juanmc2005.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":"CONTRIBUTING.md","funding":".github/FUNDING.yml","license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null},"funding":{"github":["juanmc2005"]}},"created_at":"2021-08-09T12:59:01.000Z","updated_at":"2025-05-02T23:35:14.000Z","dependencies_parsed_at":"2023-10-11T17:58:11.090Z","dependency_job_id":"467bfd8b-32a3-4cf6-b715-affcb43849f2","html_url":"https://github.com/juanmc2005/diart","commit_stats":null,"previous_names":["juanmc2005/streamingspeakerdiarization"],"tags_count":13,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/juanmc2005%2Fdiart","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/juanmc2005%2Fdiart/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/juanmc2005%2Fdiart/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/juanmc2005%2Fdiart/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/juanmc2005","download_url":"https://codeload.github.com/juanmc2005/diart/tar.gz/refs/heads/main","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":254049368,"owners_count":22006040,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["deep-learning","real-time","speaker-diarization","speaker-embedding","streaming-audio","transcription","voice-activity-detection"],"created_at":"2024-08-09T14:02:40.877Z","updated_at":"2025-05-14T01:11:06.002Z","avatar_url":"https://github.com/juanmc2005.png","language":"Python","funding_links":["https://github.com/sponsors/juanmc2005"],"categories":["Python"],"sub_categories":[],"readme":"\u003cp align=\"center\"\u003e\n\u003cimg width=\"100%\" src=\"https://github.com/juanmc2005/diart/blob/main/logo.jpg?raw=true\" title=\"diart logo\" /\u003e\n\u003c/p\u003e\n\n\u003cp align=\"center\"\u003e\n\u003ci\u003e🌿 Build AI-powered real-time audio applications in a breeze 🌿\u003c/i\u003e\n\u003c/p\u003e\n\n\u003cp align=\"center\"\u003e\n\u003cimg alt=\"PyPI Version\" src=\"https://img.shields.io/pypi/v/diart?color=g\"\u003e\n\u003cimg alt=\"PyPI Downloads\" src=\"https://static.pepy.tech/personalized-badge/diart?period=total\u0026units=international_system\u0026left_color=grey\u0026right_color=brightgreen\u0026left_text=downloads\"\u003e\n\u003cimg alt=\"Python Versions\" src=\"https://img.shields.io/badge/python-3.10%20%7C%203.11%20%7C%203.12-dark_green\"\u003e\n\u003cimg alt=\"Code size in bytes\" src=\"https://img.shields.io/github/languages/code-size/juanmc2005/StreamingSpeakerDiarization?color=g\"\u003e\n\u003cimg alt=\"License\" src=\"https://img.shields.io/github/license/juanmc2005/StreamingSpeakerDiarization?color=g\"\u003e\n\u003ca href=\"https://joss.theoj.org/papers/cc9807c6de75ea4c29025c7bd0d31996\"\u003e\u003cimg src=\"https://joss.theoj.org/papers/cc9807c6de75ea4c29025c7bd0d31996/status.svg\"\u003e\u003c/a\u003e\n\u003c/p\u003e\n\n\u003cdiv align=\"center\"\u003e\n  \u003ch4\u003e\n    \u003ca href=\"#-installation\"\u003e\n      💾 Installation\n    \u003c/a\u003e\n    \u003cspan\u003e | \u003c/span\u003e\n    \u003ca href=\"#%EF%B8%8F-stream-audio\"\u003e\n      🎙️ Stream audio\n    \u003c/a\u003e\n    \u003cspan\u003e | \u003c/span\u003e\n    \u003ca href=\"#-models\"\u003e\n      🧠 Models\n    \u003c/a\u003e\n    \u003cbr /\u003e\n    \u003ca href=\"#-tune-hyper-parameters\"\u003e\n      📈 Tuning\n    \u003c/a\u003e\n    \u003cspan\u003e | \u003c/span\u003e\n    \u003ca href=\"#-build-pipelines\"\u003e\n      🧠🔗 Pipelines\n    \u003c/a\u003e\n    \u003cspan\u003e | \u003c/span\u003e\n    \u003ca href=\"#-websockets\"\u003e\n      🌐 WebSockets\n    \u003c/a\u003e\n    \u003cspan\u003e | \u003c/span\u003e\n    \u003ca href=\"#-powered-by-research\"\u003e\n      🔬 Research\n    \u003c/a\u003e\n  \u003c/h4\u003e\n\u003c/div\u003e\n\n\u003cbr/\u003e\n\n\u003cp align=\"center\"\u003e\n\u003cimg width=\"100%\" src=\"https://github.com/juanmc2005/diart/blob/main/demo.gif?raw=true\" title=\"Real-time diarization example\" /\u003e\n\u003c/p\u003e\n\n## ⚡ Quick introduction\n\nDiart is a python framework to build AI-powered real-time audio applications.\nIts key feature is the ability to recognize different speakers in real time with state-of-the-art performance,\na task commonly known as \"speaker diarization\".\n\nThe pipeline `diart.SpeakerDiarization` combines a speaker segmentation and a speaker embedding model\nto power an incremental clustering algorithm that gets more accurate as the conversation progresses:\n\n\u003cp align=\"center\"\u003e\n\u003cimg width=\"100%\" src=\"https://github.com/juanmc2005/diart/blob/main/pipeline.gif?raw=true\" title=\"Real-time speaker diarization pipeline\" /\u003e\n\u003c/p\u003e\n\nWith diart you can also create your own custom AI pipeline, benchmark it,\ntune its hyper-parameters, and even serve it on the web using websockets.\n\n**We provide pre-trained pipelines for:**\n\n- Speaker Diarization\n- Voice Activity Detection\n- Transcription ([coming soon](https://github.com/juanmc2005/diart/pull/144))\n- [Speaker-Aware Transcription](https://betterprogramming.pub/color-your-captions-streamlining-live-transcriptions-with-diart-and-openais-whisper-6203350234ef) ([coming soon](https://github.com/juanmc2005/diart/pull/147))\n\n## 💾 Installation\n\n**1) Make sure your system has the following dependencies:**\n\n```\nffmpeg \u003c 4.4\nportaudio == 19.6.X\nlibsndfile \u003e= 1.2.2\n```\n\nAlternatively, we provide an `environment.yml` file for a pre-configured conda environment:\n\n```shell\nconda env create -f diart/environment.yml\nconda activate diart\n```\n\n**2) Install the package:**\n```shell\npip install diart\n```\n\n### Get access to 🎹 pyannote models\n\nBy default, diart is based on [pyannote.audio](https://github.com/pyannote/pyannote-audio) models from the [huggingface](https://huggingface.co/) hub.\nIn order to use them, please follow these steps:\n\n1) [Accept user conditions](https://huggingface.co/pyannote/segmentation) for the `pyannote/segmentation` model\n2) [Accept user conditions](https://huggingface.co/pyannote/segmentation-3.0) for the newest `pyannote/segmentation-3.0` model\n3) [Accept user conditions](https://huggingface.co/pyannote/embedding) for the `pyannote/embedding` model\n4) Install [huggingface-cli](https://huggingface.co/docs/huggingface_hub/quick-start#install-the-hub-library) and [log in](https://huggingface.co/docs/huggingface_hub/quick-start#login) with your user access token (or provide it manually in diart CLI or API).\n\n## 🎙️ Stream audio\n\n### From the command line\n\nA recorded conversation:\n\n```shell\ndiart.stream /path/to/audio.wav\n```\n\nA live conversation:\n\n```shell\n# Use \"microphone:ID\" to select a non-default device\n# See `python -m sounddevice` for available devices\ndiart.stream microphone\n```\n\nBy default, diart runs a speaker diarization pipeline, equivalent to setting `--pipeline SpeakerDiarization`,\nbut you can also set it to `--pipeline VoiceActivityDetection`. See `diart.stream -h` for more options.\n\n### From python\n\nUse `StreamingInference` to run a pipeline on an audio source and write the results to disk:\n\n```python\nfrom diart import SpeakerDiarization\nfrom diart.sources import MicrophoneAudioSource\nfrom diart.inference import StreamingInference\nfrom diart.sinks import RTTMWriter\n\npipeline = SpeakerDiarization()\nmic = MicrophoneAudioSource()\ninference = StreamingInference(pipeline, mic, do_plot=True)\ninference.attach_observers(RTTMWriter(mic.uri, \"/output/file.rttm\"))\nprediction = inference()\n```\n\nFor inference and evaluation on a dataset we recommend to use `Benchmark` (see notes on [reproducibility](#reproducibility)).\n\n## 🧠 Models\n\nYou can use other models with the `--segmentation` and `--embedding` arguments.\nOr in python:\n\n```python\nimport diart.models as m\n\nsegmentation = m.SegmentationModel.from_pretrained(\"model_name\")\nembedding = m.EmbeddingModel.from_pretrained(\"model_name\")\n```\n\n### Pre-trained models\n\nBelow is a list of all the models currently supported by diart:\n\n| Model Name                                                                                                                | Model Type   | CPU Time* | GPU Time* |\n|---------------------------------------------------------------------------------------------------------------------------|--------------|-----------|-----------|\n| [🤗](https://huggingface.co/pyannote/segmentation) `pyannote/segmentation` (default)                                      | segmentation | 12ms      | 8ms       |\n| [🤗](https://huggingface.co/pyannote/segmentation-3.0) `pyannote/segmentation-3.0`                                        | segmentation | 11ms      | 8ms       |\n| [🤗](https://huggingface.co/pyannote/embedding) `pyannote/embedding` (default)                                            | embedding | 26ms      | 12ms      |\n| [🤗](https://huggingface.co/hbredin/wespeaker-voxceleb-resnet34-LM) `hbredin/wespeaker-voxceleb-resnet34-LM` (ONNX)       | embedding | 48ms      | 15ms      |\n| [🤗](https://huggingface.co/pyannote/wespeaker-voxceleb-resnet34-LM) `pyannote/wespeaker-voxceleb-resnet34-LM` (PyTorch)  | embedding | 150ms     | 29ms      |\n| [🤗](https://huggingface.co/speechbrain/spkrec-xvect-voxceleb) `speechbrain/spkrec-xvect-voxceleb`                        | embedding | 41ms      | 15ms      |\n| [🤗](https://huggingface.co/speechbrain/spkrec-ecapa-voxceleb) `speechbrain/spkrec-ecapa-voxceleb`                        | embedding | 41ms      | 14ms      |\n| [🤗](https://huggingface.co/speechbrain/spkrec-ecapa-voxceleb-mel-spec) `speechbrain/spkrec-ecapa-voxceleb-mel-spec`      | embedding | 42ms      | 14ms      |\n| [🤗](https://huggingface.co/speechbrain/spkrec-resnet-voxceleb) `speechbrain/spkrec-resnet-voxceleb`                      | embedding | 41ms      | 16ms      |\n| [🤗](https://huggingface.co/nvidia/speakerverification_en_titanet_large) `nvidia/speakerverification_en_titanet_large`    | embedding | 91ms      | 16ms      |\n\nThe latency of segmentation models is measured in a VAD pipeline (5s chunks).\n\nThe latency of embedding models is measured in a diarization pipeline using `pyannote/segmentation` (also 5s chunks).\n\n\\* CPU: AMD Ryzen 9 - GPU: RTX 4060 Max-Q\n\n### Custom models\n\nThird-party models can be integrated by providing a loader function:\n\n```python\nfrom diart import SpeakerDiarization, SpeakerDiarizationConfig\nfrom diart.models import EmbeddingModel, SegmentationModel\n\ndef segmentation_loader():\n    # It should take a waveform and return a segmentation tensor\n    return load_pretrained_model(\"my_model.ckpt\")\n\ndef embedding_loader():\n    # It should take (waveform, weights) and return per-speaker embeddings\n    return load_pretrained_model(\"my_other_model.ckpt\")\n\nsegmentation = SegmentationModel(segmentation_loader)\nembedding = EmbeddingModel(embedding_loader)\nconfig = SpeakerDiarizationConfig(\n    segmentation=segmentation,\n    embedding=embedding,\n)\npipeline = SpeakerDiarization(config)\n```\n\nIf you have an ONNX model, you can use `from_onnx()`:\n\n```python\nfrom diart.models import EmbeddingModel\n\nembedding = EmbeddingModel.from_onnx(\n    model_path=\"my_model.ckpt\",\n    input_names=[\"x\", \"w\"],  # defaults to [\"waveform\", \"weights\"]\n    output_name=\"output\",  # defaults to \"embedding\"\n)\n```\n\n## 📈 Tune hyper-parameters\n\nDiart implements an optimizer based on [optuna](https://optuna.readthedocs.io/en/stable/index.html) that allows you to tune pipeline hyper-parameters to your needs.\n\n### From the command line\n\n```shell\ndiart.tune /wav/dir --reference /rttm/dir --output /output/dir\n```\n\nSee `diart.tune -h` for more options.\n\n### From python\n\n```python\nfrom diart.optim import Optimizer\n\noptimizer = Optimizer(\"/wav/dir\", \"/rttm/dir\", \"/output/dir\")\noptimizer(num_iter=100)\n```\n\nThis will write results to an sqlite database in `/output/dir`.\n\n### Distributed tuning\n\nFor bigger datasets, it is sometimes more convenient to run multiple optimization processes in parallel.\nTo do this, create a study on a [recommended DBMS](https://optuna.readthedocs.io/en/stable/tutorial/10_key_features/004_distributed.html#sphx-glr-tutorial-10-key-features-004-distributed-py) (e.g. MySQL or PostgreSQL) making sure that the study and database names match:\n\n```shell\nmysql -u root -e \"CREATE DATABASE IF NOT EXISTS example\"\noptuna create-study --study-name \"example\" --storage \"mysql://root@localhost/example\"\n```\n\nYou can now run multiple identical optimizers pointing to this database:\n\n```shell\ndiart.tune /wav/dir --reference /rttm/dir --storage mysql://root@localhost/example\n```\n\nor in python:\n\n```python\nfrom diart.optim import Optimizer\nfrom optuna.samplers import TPESampler\nimport optuna\n\ndb = \"mysql://root@localhost/example\"\nstudy = optuna.load_study(\"example\", db, TPESampler())\noptimizer = Optimizer(\"/wav/dir\", \"/rttm/dir\", study)\noptimizer(num_iter=100)\n```\n\n## 🧠🔗 Build pipelines\n\nFor a more advanced usage, diart also provides building blocks that can be combined to create your own pipeline.\nStreaming is powered by [RxPY](https://github.com/ReactiveX/RxPY), but the `blocks` module is completely independent and can be used separately.\n\n### Example\n\nObtain overlap-aware speaker embeddings from a microphone stream:\n\n```python\nimport rx.operators as ops\nimport diart.operators as dops\nfrom diart.sources import MicrophoneAudioSource, FileAudioSource\nfrom diart.blocks import SpeakerSegmentation, OverlapAwareSpeakerEmbedding\n\nsegmentation = SpeakerSegmentation.from_pretrained(\"pyannote/segmentation\")\nembedding = OverlapAwareSpeakerEmbedding.from_pretrained(\"pyannote/embedding\")\n\nsource = MicrophoneAudioSource()\n# To take input from file:\n# source = FileAudioSource(\"\u003cfilename\u003e\", sample_rate=16000)\n\n# Make sure the models have been trained with this sample rate\nprint(source.sample_rate)\n\nstream = mic.stream.pipe(\n    # Reformat stream to 5s duration and 500ms shift\n    dops.rearrange_audio_stream(sample_rate=source.sample_rate),\n    ops.map(lambda wav: (wav, segmentation(wav))),\n    ops.starmap(embedding)\n).subscribe(on_next=lambda emb: print(emb.shape))\n\nsource.read()\n```\n\nOutput:\n\n```\n# Shape is (batch_size, num_speakers, embedding_dim)\ntorch.Size([1, 3, 512])\ntorch.Size([1, 3, 512])\ntorch.Size([1, 3, 512])\n...\n```\n\n## 🌐 WebSockets\n\nDiart is also compatible with the WebSocket protocol to serve pipelines on the web.\n\n### From the command line\n\n```shell\ndiart.serve --host 0.0.0.0 --port 7007\ndiart.client microphone --host \u003cserver-address\u003e --port 7007\n```\n\n**Note:** make sure that the client uses the same `step` and `sample_rate` than the server with `--step` and `-sr`.\n\nSee `-h` for more options.\n\n### From python\n\nFor customized solutions, a server can also be created in python using the `WebSocketAudioSource`:\n\n```python\nfrom diart import SpeakerDiarization\nfrom diart.sources import WebSocketAudioSource\nfrom diart.inference import StreamingInference\n\npipeline = SpeakerDiarization()\nsource = WebSocketAudioSource(pipeline.config.sample_rate, \"localhost\", 7007)\ninference = StreamingInference(pipeline, source)\ninference.attach_hooks(lambda ann_wav: source.send(ann_wav[0].to_rttm()))\nprediction = inference()\n```\n\n## 🔬 Powered by research\n\nDiart is the official implementation of the paper\n[Overlap-aware low-latency online speaker diarization based on end-to-end local segmentation](https://github.com/juanmc2005/diart/blob/main/paper.pdf)\nby [Juan Manuel Coria](https://juanmc2005.github.io/),\n[Hervé Bredin](https://herve.niderb.fr),\n[Sahar Ghannay](https://saharghannay.github.io/)\nand [Sophie Rosset](https://perso.limsi.fr/rosset/).\n\n\n\u003e We propose to address online speaker diarization as a combination of incremental clustering and local diarization applied to a rolling buffer updated every 500ms. Every single step of the proposed pipeline is designed to take full advantage of the strong ability of a recently proposed end-to-end overlap-aware segmentation to detect and separate overlapping speakers. In particular, we propose a modified version of the statistics pooling layer (initially introduced in the x-vector architecture) to give less weight to frames where the segmentation model predicts simultaneous speakers. Furthermore, we derive cannot-link constraints from the initial segmentation step to prevent two local speakers from being wrongfully merged during the incremental clustering step. Finally, we show how the latency of the proposed approach can be adjusted between 500ms and 5s to match the requirements of a particular use case, and we provide a systematic analysis of the influence of latency on the overall performance (on AMI, DIHARD and VoxConverse).\n\n\u003cp align=\"center\"\u003e\n\u003cimg height=\"400\" src=\"https://github.com/juanmc2005/diart/blob/main/figure1.png?raw=true\" title=\"Visual explanation of the system\" width=\"325\" /\u003e\n\u003c/p\u003e\n\n### Citation\n\nIf you found diart useful, please make sure to cite our paper:\n\n```bibtex\n@inproceedings{diart,  \n  author={Coria, Juan M. and Bredin, Hervé and Ghannay, Sahar and Rosset, Sophie},  \n  booktitle={2021 IEEE Automatic Speech Recognition and Understanding Workshop (ASRU)},   \n  title={Overlap-Aware Low-Latency Online Speaker Diarization Based on End-to-End Local Segmentation}, \n  year={2021},\n  pages={1139-1146},\n  doi={10.1109/ASRU51503.2021.9688044},\n}\n```\n\n### Reproducibility\n\n![Results table](https://github.com/juanmc2005/diart/blob/main/table1.png?raw=true)\n\n**Important:** We highly recommend installing `pyannote.audio\u003c3.1` to reproduce these results.\nFor more information, see [this issue](https://github.com/juanmc2005/diart/issues/214).\n\nDiart aims to be lightweight and capable of real-time streaming in practical scenarios.\nIts performance is very close to what is reported in the paper (and sometimes even a bit better).\n\nTo obtain the best results, make sure to use the following hyper-parameters:\n\n| Dataset     | latency | tau    | rho    | delta |\n|-------------|---------|--------|--------|-------|\n| DIHARD III  | any     | 0.555  | 0.422  | 1.517 |\n| AMI         | any     | 0.507  | 0.006  | 1.057 |\n| VoxConverse | any     | 0.576  | 0.915  | 0.648 |\n| DIHARD II   | 1s      | 0.619  | 0.326  | 0.997 |\n| DIHARD II   | 5s      | 0.555  | 0.422  | 1.517 |\n\n`diart.benchmark` and `diart.inference.Benchmark` can run, evaluate and measure the real-time latency of the pipeline. For instance, for a DIHARD III configuration:\n\n```shell\ndiart.benchmark /wav/dir --reference /rttm/dir --tau-active=0.555 --rho-update=0.422 --delta-new=1.517 --segmentation pyannote/segmentation@Interspeech2021\n```\n\nor using the inference API:\n\n```python\nfrom diart.inference import Benchmark, Parallelize\nfrom diart import SpeakerDiarization, SpeakerDiarizationConfig\nfrom diart.models import SegmentationModel\n\nbenchmark = Benchmark(\"/wav/dir\", \"/rttm/dir\")\n\nmodel_name = \"pyannote/segmentation@Interspeech2021\"\nmodel = SegmentationModel.from_pretrained(model_name)\nconfig = SpeakerDiarizationConfig(\n    # Set the segmentation model used in the paper\n    segmentation=model,\n    step=0.5,\n    latency=0.5,\n    tau_active=0.555,\n    rho_update=0.422,\n    delta_new=1.517\n)\nbenchmark(SpeakerDiarization, config)\n\n# Run the same benchmark in parallel\np_benchmark = Parallelize(benchmark, num_workers=4)\nif __name__ == \"__main__\":  # Needed for multiprocessing\n    p_benchmark(SpeakerDiarization, config)\n```\n\nThis pre-calculates model outputs in batches, so it runs a lot faster.\nSee `diart.benchmark -h` for more options.\n\nFor convenience and to facilitate future comparisons, we also provide the\n\u003ca href=\"https://github.com/juanmc2005/diart/tree/main/expected_outputs\"\u003eexpected outputs\u003c/a\u003e\nof the paper implementation in RTTM format for every entry of Table 1 and Figure 5.\nThis includes the VBx offline topline as well as our proposed online approach with\nlatencies 500ms, 1s, 2s, 3s, 4s, and 5s.\n\n![Figure 5](https://github.com/juanmc2005/diart/blob/main/figure5.png?raw=true)\n\n## 📑 License\n\n```\nMIT License\n\nCopyright (c) 2021 Université Paris-Saclay\nCopyright (c) 2021 CNRS\n\nPermission is hereby granted, free of charge, to any person obtaining a copy\nof this software and associated documentation files (the \"Software\"), to deal\nin the Software without restriction, including without limitation the rights\nto use, copy, modify, merge, publish, distribute, sublicense, and/or sell\ncopies of the Software, and to permit persons to whom the Software is\nfurnished to do so, subject to the following conditions:\n\nThe above copyright notice and this permission notice shall be included in all\ncopies or substantial portions of the Software.\n\nTHE SOFTWARE IS PROVIDED \"AS IS\", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR\nIMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,\nFITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE\nAUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER\nLIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,\nOUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE\nSOFTWARE.\n```\n\n\u003cp style=\"color:grey;font-size:14px;\"\u003eLogo generated by \u003ca href=\"https://www.designevo.com/\" title=\"Free Online Logo Maker\"\u003eDesignEvo free logo designer\u003c/a\u003e\u003c/p\u003e\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fjuanmc2005%2Fdiart","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fjuanmc2005%2Fdiart","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fjuanmc2005%2Fdiart/lists"}