{"title":"biorecap:用本地 LLM 总结 bioRxiv 预印本的 R 软件包","authors":"Stephen D. Turner","doi":"arxiv-2408.11707","DOIUrl":null,"url":null,"abstract":"The establishment of bioRxiv facilitated the rapid adoption of preprints in\nthe life sciences, accelerating the dissemination of new research findings.\nHowever, the sheer volume of preprints published daily can be overwhelming,\nmaking it challenging for researchers to stay updated on the latest\ndevelopments. Here, I introduce biorecap, an R package that retrieves and\nsummarizes bioRxiv preprints using a large language model (LLM) running locally\non nearly any commodity laptop. biorecap leverages the ollamar package to\ninterface with the Ollama server and API endpoints, allowing users to prompt\nany local LLM available through Ollama. The package follows tidyverse\nconventions, enabling users to pipe the output of one function as input to\nanother. Additionally, biorecap provides a single wrapper function that\ngenerates a timestamped CSV file and HTML report containing short summaries of\nrecent preprints published in user-configurable subject areas. By combining the\nstrengths of LLMs with the flexibility and security of local execution,\nbiorecap represents an advancement in the tools available for managing the\ninformation overload in modern scientific research. The biorecap R package is\navailable on GitHub at https://github.com/stephenturner/biorecap under an\nopen-source (MIT) license.","PeriodicalId":501219,"journal":{"name":"arXiv - QuanBio - Other Quantitative Biology","volume":"13 1","pages":""},"PeriodicalIF":0.0000,"publicationDate":"2024-08-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"biorecap: an R package for summarizing bioRxiv preprints with a local LLM\",\"authors\":\"Stephen D. Turner\",\"doi\":\"arxiv-2408.11707\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"The establishment of bioRxiv facilitated the rapid adoption of preprints in\\nthe life sciences, accelerating the dissemination of new research findings.\\nHowever, the sheer volume of preprints published daily can be overwhelming,\\nmaking it challenging for researchers to stay updated on the latest\\ndevelopments. Here, I introduce biorecap, an R package that retrieves and\\nsummarizes bioRxiv preprints using a large language model (LLM) running locally\\non nearly any commodity laptop. biorecap leverages the ollamar package to\\ninterface with the Ollama server and API endpoints, allowing users to prompt\\nany local LLM available through Ollama. The package follows tidyverse\\nconventions, enabling users to pipe the output of one function as input to\\nanother. Additionally, biorecap provides a single wrapper function that\\ngenerates a timestamped CSV file and HTML report containing short summaries of\\nrecent preprints published in user-configurable subject areas. By combining the\\nstrengths of LLMs with the flexibility and security of local execution,\\nbiorecap represents an advancement in the tools available for managing the\\ninformation overload in modern scientific research. The biorecap R package is\\navailable on GitHub at https://github.com/stephenturner/biorecap under an\\nopen-source (MIT) license.\",\"PeriodicalId\":501219,\"journal\":{\"name\":\"arXiv - QuanBio - Other Quantitative Biology\",\"volume\":\"13 1\",\"pages\":\"\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2024-08-21\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"arXiv - QuanBio - Other Quantitative Biology\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/arxiv-2408.11707\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"arXiv - QuanBio - Other Quantitative Biology","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/arxiv-2408.11707","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0
摘要
bioRxiv的建立促进了预印本在生命科学领域的快速应用,加速了新研究成果的传播。然而,每天发表的预印本数量之大可能会让人应接不暇,这使得研究人员难以及时了解最新进展。在这里,我将介绍 biorecap,它是一个 R 软件包,可以使用在几乎所有商品笔记本电脑上本地运行的大型语言模型(LLM)检索和汇总 bioRxiv 预印本。biorecap 利用 ollamar 软件包与 Ollama 服务器和 API 端点接口,允许用户通过 Ollama 提示任何可用的本地 LLM。该软件包遵循 tidyverseconventions,使用户能够将一个函数的输出作为另一个函数的输入。此外,biorecap 还提供了一个封装函数,可生成带有时间戳的 CSV 文件和 HTML 报告,其中包含用户可配置的主题领域中近期发表的预印本的简短摘要。通过将 LLM 的优势与本地执行的灵活性和安全性相结合,biorecap 代表了现代科学研究中信息过载管理工具的一大进步。biorecap R 软件包可在 GitHub https://github.com/stephenturner/biorecap 上获取,采用开源(MIT)许可。
biorecap: an R package for summarizing bioRxiv preprints with a local LLM
The establishment of bioRxiv facilitated the rapid adoption of preprints in
the life sciences, accelerating the dissemination of new research findings.
However, the sheer volume of preprints published daily can be overwhelming,
making it challenging for researchers to stay updated on the latest
developments. Here, I introduce biorecap, an R package that retrieves and
summarizes bioRxiv preprints using a large language model (LLM) running locally
on nearly any commodity laptop. biorecap leverages the ollamar package to
interface with the Ollama server and API endpoints, allowing users to prompt
any local LLM available through Ollama. The package follows tidyverse
conventions, enabling users to pipe the output of one function as input to
another. Additionally, biorecap provides a single wrapper function that
generates a timestamped CSV file and HTML report containing short summaries of
recent preprints published in user-configurable subject areas. By combining the
strengths of LLMs with the flexibility and security of local execution,
biorecap represents an advancement in the tools available for managing the
information overload in modern scientific research. The biorecap R package is
available on GitHub at https://github.com/stephenturner/biorecap under an
open-source (MIT) license.