Batch 9: Obsidian插件/AI开源平替/Coze培训/TK面单/Ubuntu科学上网
- Sources: 5个新文档 - Concepts: ProxyChains, SOCKS5代理, Docker Daemon代理 - Index: 更新至 Batch 9 - 累计 sources: 108/182
This commit is contained in:
38
wiki/sources/arXiv-Paper-Reader.md
Normal file
38
wiki/sources/arXiv-Paper-Reader.md
Normal file
@@ -0,0 +1,38 @@
|
||||
---
|
||||
title: "arXiv Paper Reader"
|
||||
type: source
|
||||
tags: [research, arxiv, openclaw-skill]
|
||||
date: 2026-04-16
|
||||
---
|
||||
|
||||
## Source File
|
||||
- [[raw/Agent/usecases/arxiv-paper-reader.md]]
|
||||
|
||||
## Summary
|
||||
- 核心主题:基于 OpenClaw Agent 的 arXiv 论文阅读助手工作流
|
||||
- 问题域:PDF 下载后切换论文丢失上下文、LaTeX 公式难以解析
|
||||
- 方法/机制:Prismer AI 的 arxiv-reader skill(3 个工具:arxiv_fetch/arxiv_sections/arxiv_abstract);Node.js 原生实现无需 Docker;直接从 arXiv 下载并自动扁平化 LaTeX 源码
|
||||
- 结论/价值:对话式论文阅读,支持摘要浏览、多篇对比、章节定位、结果缓存
|
||||
|
||||
## Key Claims
|
||||
- arxiv-reader skill 无需 Docker 或 Python,通过 Node.js 内置模块独立运行
|
||||
- LaTeX 源码自动展平(flatten includes),消除公式解析障碍
|
||||
- 多篇论文可并行获取摘要并生成对比表格,按相关性排序
|
||||
- 结果本地缓存,回访论文秒级加载
|
||||
|
||||
## Key Quotes
|
||||
> "Reading arXiv papers means downloading PDFs, losing context when switching between papers, and struggling to parse dense LaTeX notation."
|
||||
|
||||
## Key Concepts
|
||||
- [[arXiv-API]]:论文元数据和 PDF 源码获取接口,支持 ID 检索
|
||||
- [[LaTeX-Flattening]]:自动解析并合并 LaTeX 源码中的 \include 语句,生成可读文本
|
||||
- [[arxiv-reader-skill]]:Prismer 项目开源 skill,包含 fetch/sections/abstract 三个工具
|
||||
|
||||
## Key Entities
|
||||
- [[Prismer-AI]]:arxiv-reader skill 的开发方,GitHub 仓库为 Prismer-AI/Prismer
|
||||
|
||||
## Connections
|
||||
- [[OpenClaw]] ← 平台 ← [[arXiv-Paper-Reader]]
|
||||
- [[Personal-Knowledge-Base-RAG]] ← 类似工作流 ← [[arXiv-Paper-Reader]]
|
||||
|
||||
## Contradictions
|
||||
Reference in New Issue
Block a user