dosubot[bot]

Results 14268 comments of dosubot[bot]

[Question]: I downloaded the MS MACROW CREATE V1.1 dataset and encountered an error when testing the retrieval performance using the benchmark.py file.

Here’s a complete Python script to download the MS MARCO Passage Ranking dataset from HuggingFace and convert it to the Parquet format required by RAGFlow’s benchmark.py. This script builds the...

[Question]: I downloaded the MS MACROW CREATE V1.1 dataset and encountered an error when testing the retrieval performance using the benchmark.py file.

You need to use the `"v1.1"` config for the MS MARCO dataset on HuggingFace, not `"passage-ranking"`—that’s why you saw the ValueError. For the KeyError on `"relevance"`, the `"v1.1"` split does...

[Question]: I downloaded the MS MACROW CREATE V1.1 dataset and encountered an error when testing the retrieval performance using the benchmark.py file.

The HuggingFace MS MARCO v1.1 dataset only provides the splits "train", "validation", and "test"—there is no "validation_qrels" or similar split, which is why you see that error. To get the...

[Question]: I downloaded the MS MACROW CREATE V1.1 dataset and encountered an error when testing the retrieval performance using the benchmark.py file.

The error `ValueError: max() arg is an empty sequence` happens because the evaluation step in `benchmark.py` is receiving empty or malformed data—specifically, the qrels or run structures are empty. This...

[Question]: I downloaded the MS MACROW CREATE V1.1 dataset and encountered an error when testing the retrieval performance using the benchmark.py file.

The ValueError is caused by empty or malformed dataset files, not by your embedding model. The cleaning script removed all rows because your Parquet file(s) were already missing valid data—likely...

[🐛 Bug]: Service Topology screen does not show related alerts

The issue you're experiencing, where enriched alerts are not appearing in the Service Topology Map, could be related to the configuration or the criteria used by the Topology Processor in...

[Bug]: build_TOC failed.

Hi @wcbzcy! I'm [Dosu](https://go.dosu.dev/dosubot) and I’m helping the ragflow team. The error happens because the TOC generation code expects every item to be a dictionary, but sometimes receives a string...

作者有没有想过用langgraph重写一下

Hi, @TryAwesome. I'm [Dosu](https://dosu.dev), and I'm helping the 666ghj/BettaFish team manage their backlog and am marking this issue as stale. **Issue Summary** - You proposed rewriting the project using LangGraph...

[Knowledge] KB Meatadata don't stay enabled in empty KB

The issue you're experiencing, where the built-in metadata toggle does not stay enabled in an empty Knowledge Base (KB), might be related to the default behavior or a specific condition...

[Bug]: itemMarkerSize 设置为函数时，会导致高度坍塌

Hi @Song-aff! I'm [Dosu](https://go.dosu.dev/dosubot) and I’m helping the ant-design-charts team. 根据代码和相关 issue 的调查，legend.color.itemMarkerSize 目前只支持数值类型，不支持函数类型。如果设置为函数，图例的高度计算会出错，导致“高度坍塌”现象。建议将 itemMarkerSize 设置为具体数值（如 12 或 16），这样可以正常显示和调整图例 marker 的大小 [相关讨论](https://github.com/ant-design/ant-design-charts/issues/2631)。 目前没有发现官方文档或代码支持 itemMarkerSize 为函数，也没有相关修复的 PR。如果需要动态调整 marker 大小，建议在配置前自行处理逻辑，最终传递数值类型给 itemMarkerSize。 *To...