Repository navigation
Releases: Panda-of-Axin/Freepaper
Release list
v2.0.5
Freepaper v2.0.5 — Release Notes
Release title: Freepaper v2.0.5 — CNKI / CHNDOI DOI-resolution fixes
Tag: v2.0.5
English
Freepaper v2.0.5 improves the CNKI workflow for DOI-driven tasks and keeps authentication recoverable instead of treating it as a terminal failure.
Highlights
- Recoverable CNKI login/authentication: when CNKI requires login or institutional authentication, the current paper now waits for the user and resumes detection after authentication instead of immediately being counted as failed.
- Safer DOI handoff: DOI tasks wait until the resolver has actually reached the publisher/article page before PDF detection begins.
- CHNDOI multi-target resolver support: pages under
chndoi.org/Resolution/Handlerare recognized as resolver pages rather than article pages. Freepaper extracts the available HURL targets, prefers the domesticlink.cnki.netroute for CNKI records, and continues to the actual paper page. - Regression case added: DOI
10.13250/j.cnki.wndz.25110501is covered across thedoi.org → chndoi.org → link.cnki.net / CNKInavigation path. - Existing v2.0.2 PDF safety behavior is preserved: dynamic IEEE, Wiley, ScienceDirect, and CNKI PDF endpoints are handled in authenticated page context where possible, and HTML challenge pages are not intentionally saved as PDFs.
Upgrade notes
- Browser permissions are unchanged from v2.0.2; only the manifest version changes to 2.0.5.
- CNKI support remains experimental because page structure and institutional authentication flows vary by institution.
- Freepaper does not bypass paywalls, CAPTCHAs, institutional permissions, or publisher access controls.
Verification
npm run verify passed for the packaged source:
- extension checks passed;
- targeted regression tests passed.
中文
Freepaper v2.0.5 主要修复 知网 DOI 输入、登录等待以及 CHNDOI 多重解析页面 的处理流程。
本版重点
- 知网登录/机构认证改为可恢复等待状态: 遇到需要登录时不再直接判定失败,也不会立刻结束当前批量任务;用户完成登录后可以继续检测当前论文。
- 修复 DOI 过早扫描: DOI 输入会等待解析器真正跳转到出版商/论文页面后,再开始 PDF 检测。
- 支持 CHNDOI 多重解析地址选择页:
chndoi.org/Resolution/Handler不再被误判为论文页。Freepaper 会读取页面提供的 HURL,并对知网记录优先选择境内link.cnki.net路线,再继续进入真实论文页面。 - 加入针对性回归用例: DOI
10.13250/j.cnki.wndz.25110501已覆盖doi.org → chndoi.org → link.cnki.net / CNKI跳转流程。 - 保留 v2.0.2 的 PDF 安全策略: IEEE、Wiley、ScienceDirect、知网等动态 PDF 地址尽量在已认证的论文页面上下文中获取,并避免把 HTML 验证页误保存成 PDF。
升级说明
- 与 v2.0.2 相比,浏览器权限没有新增,仅版本号更新为 2.0.5。
- 知网支持仍属于实验性功能,不同学校、机构账号和知网页面结构可能存在差异。
- Freepaper 不绕过付费墙、验证码、机构权限或出版商访问控制。
验证结果
打包前已执行 npm run verify:
- extension checks passed;
- targeted regression tests passed。
v2.0.2
Freepaper v2.0.2
中文
Freepaper 是一款开源浏览器扩展,用于辅助自动化处理学术论文检索与下载流程。
本版本是 Freepaper 相较 v1.0 的一次重要升级,重新设计了批量下载流程、多阶段认证处理机制,并改进了出版商网站兼容性、下载状态监控、中英文界面以及首次使用引导系统。
主要更新
- 重新设计批量文献下载流程
- 支持 DOI / URL 批量导入
- 自动提取论文标题并去重
- 持久化下载任务监控
- 多阶段登录、认证与验证处理
- 用户完成登录或人工验证后自动恢复任务
- 改进 IEEE Xplore 兼容性
- 改进 Wiley Online Library 兼容性
- 改进 ScienceDirect 兼容性
- 修复部分动态 PDF 地址被错误下载为 HTML 文件的问题
- 实验性支持 CNKI
- 自动检测 PDF 并跟踪实际下载状态
- 支持自定义下载子文件夹
- 中英文界面
- 新增首次使用引导与帮助中心
- 改进错误诊断与隐私处理
认证与访问权限
Freepaper 仅适用于用户本身已经拥有访问权限的资源。
部分出版商可能要求:
- 学校或机构登录
- 出版商个人账号登录
- 人工验证
- 访问 PDF 前完成额外认证
当需要用户操作时,Freepaper 会暂停当前任务,并在用户完成认证后继续处理。
Freepaper 不绕过付费墙、机构访问限制、CAPTCHA 或其他安全验证机制。
下载行为
Freepaper 默认按照文献顺序逐篇处理,而不是将同一出版商的大量请求集中发送。
这是有意设计的。
连续向同一个学术数据库发送大量自动化请求,可能触发出版商限流机制或机构 IP 访问限制。
用户应遵守所在机构的电子资源使用政策以及相应出版商的服务条款。
支持的输入格式
- CSV
- TSV
- TXT
- DOI
- 论文 URL
目前暂不支持直接导入 XLSX 文件。
说明
出版商网站和认证机制可能随时间发生变化,因此不同数据库的实际兼容性可能受到网络环境、机构访问权限、账号状态以及出版商网站更新的影响。
欢迎通过 GitHub Issues 提交 Bug、兼容性问题和使用反馈。
English
Freepaper is an open-source browser extension for automating academic paper retrieval and download workflows.
This release represents a major evolution from Freepaper v1.0, with redesigned batch-download workflows, multi-stage authentication handling, improved publisher compatibility, download monitoring, bilingual UI, and a new onboarding system.
Highlights
- Redesigned batch paper download workflow
- DOI / URL batch import
- Automatic paper title extraction and deduplication
- Persistent download task monitor
- Multi-stage authentication and verification handling
- Automatic recovery after user login or human verification
- Improved IEEE Xplore compatibility
- Improved Wiley Online Library compatibility
- Improved ScienceDirect compatibility
- Fixed cases where context-bound PDF URLs could be downloaded as HTML instead of PDF
- Experimental CNKI support
- Automatic PDF detection and download tracking
- Custom download subfolder support
- Chinese / English interface
- First-run onboarding and help center
- Improved error diagnosis and privacy handling
Authentication and access
Freepaper works only with resources that the user is already authorized to access.
Some publishers may require:
- Institutional sign-in
- Personal account login
- Human verification
- Additional authentication before PDF access
Freepaper pauses the task when user interaction is required and resumes after authentication is completed.
Freepaper does not bypass paywalls, institutional access controls, CAPTCHA, or security verification.
Download behavior
Freepaper processes papers sequentially rather than grouping large numbers of requests by publisher.
This design is intentional. Consecutive automated requests to the same academic database may trigger publisher rate limits or institutional IP restrictions.
Users should comply with the electronic-resource policies of their institution and the terms of the corresponding publisher.
Supported input formats
- CSV
- TSV
- TXT
- DOI
- Article URL
Direct XLSX import is not currently supported.
Notes
Publisher websites and authentication mechanisms change over time. Compatibility with specific databases may therefore vary depending on network environment, institutional access, account status, and publisher-side changes.
Bug reports and compatibility feedback are welcome through GitHub Issues.