Codex/release v0.5.2#51
Conversation
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: de33305c89
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| let model_plan_valid = planned.len() >= 2; | ||
| let queries = merge_queries(planned, fallback); |
There was a problem hiding this comment.
When the LLM returns four valid queries, normalize_queries caps the concatenated list at four, so every locally generated fallback query is discarded. In that case an LLM response that omits a selected author, venue, category, or keyword causes the Semantic Scholar requests to run without that user-specified constraint; this is especially likely for the large venue lists produced by the domain filter. Reserve a fallback/constraint-bearing query or merge the structured terms into each planned query before applying the cap.
Useful? React with 👍 / 👎.
| } catch (nextError) { | ||
| setResult(null); | ||
| setError(formatErrorMessage(nextError)); |
There was a problem hiding this comment.
If a user changes the draft and the subsequent academic search fails (for example, due to a network error or an invalid persisted cutoff date), this catch only records the error and leaves session.result and session.webSupplement untouched. The panel therefore shows the prior query's papers and web sources beneath the new draft, and those stale results remain persisted across navigation; clear them at search start or in this failure path so the displayed data cannot be mistaken for the failed search's output.
Useful? React with 👍 / 👎.
这个 Pull Request 推出了小妍桌面应用的 0.5.2 版本，重点改进了论文阅读与搜索体验、增强了缓存机制，并提升了错误处理的健壮性。本次更新还包含若干 Bug 修复，以及针对学术检索和专利检索工作流的优化。以下按主题分组列出最重要的变更：
功能增强：
论文搜索改进：
缓存与性能：
错误处理与用户反馈：
Bug 修复：
其他更新：
package.json、Cargo.toml、用户代理字符串和文档中的版本号更新至 0.5.2。这些变更整体上提升了小妍桌面应用在学术研究场景下的易用性、效率和可靠性。