dosubot[bot]
dosubot[bot]
Hi @wanglxi1! I'm [Dosu](https://go.dosu.dev/dosubot) and I’m helping the ant-design-charts team. 你的问题和 ant-design-charts 及其底层 @antv/g-lite 的缓存（memoize）实现有关，社区中已经有类似的内存泄漏反馈。主要原因是 memoize 的缓存没有释放，导致内存不断增长，尤其是在高频率数据更新或不断传入新数据时更明显。这种问题在 chart update 时不会自动清理缓存，只有在销毁组件时才会释放资源，导致内存溢出[相关讨论1](https://github.com/ant-design/ant-design-charts/issues/910) [相关讨论2](https://github.com/ant-design/ant-design-charts/issues/3015)。 近期 ant-design-charts 有修复部分内存溢出问题（如事件监听器未清理），但 memoize 缓存未释放的问题在 g-lite 层面依然存在[修复PR](https://github.com/ant-design/ant-design-charts/pull/3018)。目前社区常见的临时解决方案是：定期销毁并重建图表实例，避免长时间运行导致缓存无限增长。 你的代码中 setInterval...
memoize 在 g-lite 里确实是内部实现，没有导出，所以无法直接引用或清理缓存。目前没有官方 API 或方法可以手动清除 memoize 的缓存[相关讨论](https://github.com/ant-design/ant-design-charts/issues/3015)。 社区推荐的规避方式是：通过销毁并重建图表实例（比如卸载组件或调用 destroy 方法），这样会释放相关资源和缓存，但无法只清理 memoize 缓存。如果你的场景是高频率数据更新，建议定期销毁并重建图表，或者优化数据流，减少不必要的缓存增长[修复PR说明](https://github.com/ant-design/ant-design-charts/pull/3018)。 另外，记得在 useEffect 的 return 里加上 clearInterval，避免额外的内存泄漏。当前没有更直接的解决办法，建议关注后续 g-lite 或 ant-design-charts 的版本更新。 *To reply, just mention [@dosu](https://go.dosu.dev/dosubot).* --- How...
Hi, @axiomofjoy. I'm [Dosu](https://dosu.dev), and I'm helping the Arize Phoenix team manage their backlog. I'm marking this issue as stale. **Issue Summary** - Original concern about the complexity and fragmentation...
PR #1703 adds a new "AI Tools in Development" section to Development.md, describing Coderabbit and Dosu, their purpose, usage via GitHub automation, and clarifying that no local setup is needed....
Based on everything found, this timeline mismatch error during major version upgrades from PostgreSQL 16 to 17 in CloudNativePG 1.27 does not appear to be a known or previously reported...
This behavior is a known limitation in CloudNativePG 1.26.x and earlier: updates to `.spec.serviceTemplate.metadata.annotations` in a Pooler resource are not automatically reconciled to the underlying Service after creation. The annotations...
You’re seeing three main errors: 1. **Missing resource file (`huqie.txt`)**: The benchmark expects `/usr/share/infinity/resource/rag/huqie.txt`, but recent config changes removed the `resource_dir` setting, so the file isn’t being found. You’ll need...
There isn’t official end-to-end documentation for running benchmark.py with bge-m3 in a self-hosted MySQL/OSS/Redis/Elasticsearch setup, but here’s a concise checklist and references to help you get a successful run: 1....
You don’t need to set the tenant manually—benchmark.py uses the kb_id you provide to look up the knowledge base, which includes the tenant_id. So, the tenant context is set automatically...
The official download link for the MS MARCO v1.1 dataset is: https://huggingface.co/datasets/microsoft/ms_marco This is the source used by RAGFlow maintainers for benchmarking [source](https://github.com/infiniflow/ragflow/issues/3504#issuecomment-2502620420). RAGFlow’s benchmark.py expects the dataset as one...