The segment symbol does not take effect

[chaijl]

For Word documents, they are divided into blocks in a general manner and a custom block symbol '||||' is set. However, by default, they will still be divided into blocks using either Word's block symbol or my custom block symbol, and there is no mandatory use of my custom block symbol.May I ask how to force me to customize the block symbol for partitioning

[dosubot[bot]]

To force RAGFlow to use your custom block symbol (like '||||') for partitioning Word documents, you need to avoid the 'General' slicing method, as it always splits by line breaks regardless of custom delimiter settings. Instead, use the 'Manual' or 'Paper' chunk methods, which respect custom delimiters for Word documents, including those with tables and images source.

When specifying your custom delimiter, wrap it in backticks—for example, use |||| instead of just ||||—to ensure it is recognized correctly for docx files source. Also, set the chunk token number to a very low value (like 2) so segmentation relies only on your delimiter source, source.

This combination should force RAGFlow to partition your Word document exclusively at your custom block symbol.

To reply, just mention @dosu.

---

^{How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other}