What the data for this category looks like
Primary users include product designers, interaction designers, and front-end product teams. Daily processed documents include interaction prototypes, page flowcharts, component descriptions, and review records. Data sources typically include Figma or Axure files, internal team design systems, user feedback, and usability test records. Update frequency is usually weekly or during version releases. Document structures often contain nested component hierarchies, linear operation process descriptions, and structured review comments. There is no unified fixed field format. Some documents mix vector graphics and text content.
Constraints imposed on document parsing and chunking
The nested hierarchy of interaction prototypes and review records requires parsing to retain component-operation path associations, and avoid splitting key process nodes. Multi-source mixed document formats (exported PDFs, JSON, spreadsheet files) require parsing rules to adapt to multiple input types. High-frequency document updates require chunking logic to quickly adapt to new structures, without repeated basic configuration adjustments. Short, high-comment-density review and test records require chunking to retain individual record integrity, and avoid information ambiguity from cross-paragraph splicing.
How to set configurations
| Configuration Item | Recommended Setting | Rationale |
|---|---|---|
chunkSize | 800–1200 characters | Interaction prototype documents often contain continuous operation steps and component descriptions. This length supports RAG retrieval contextual coherence and avoids splitting key operation logic |
chunkSplitter | Custom delimiter + semantic chunking | Documents exported from Axure and Figma often contain fixed-format review records and flow chart descriptions. Custom delimiters retain paragraph integrity, and semantic chunking handles text fragments without clear delimiters |
customPdfParse | Enabled | PDFs exported from interaction prototypes often contain vector graphics and nested text. Enabling this option optimizes text extraction accuracy and reduces text loss in graphic areas |
UPLOAD_FILE_MAX_SIZE | 1000 MB | The total file set for a single version of interaction prototypes usually does not exceed this threshold, avoiding upload restrictions from oversized files |
PARSE_FILE_TIMEOUT_SECONDS | 300–600 seconds | Large design system documents require processing many nested components and text. This time range reserves sufficient processing duration and avoids parsing timeouts |
chunkSettingMode | Manual mode | Chunking logic varies widely between different interaction document types (such as flow charts vs. review records). Manual configuration allows adjusting parameters for specific document scenarios |
Three common configuration errors
- Uploading Excel-format review records exported from Axure triggers an upload failure error. The cause is that parsing adaptation for the corresponding format is not enabled, and default parsing rules do not cover spreadsheet files commonly used for design reviews.
- Uploading a 3MB interaction prototype PDF results in a parsing timeout. The cause is that the
PARSE_FILE_TIMEOUT_SECONDSconfiguration value is lower than actual parsing duration, and insufficient time is reserved for processing nested text and vector graphics. - After setting a custom delimiter, chunking results either merge multiple review record paragraphs or split a single operation step. The cause is insufficient matching between
chunkSizeand the custom delimiter, and chunking parameters are not adjusted based on the document's paragraph line break habits.
How to verify correct configuration
- Upload a single typical interaction prototype file, and verify on the FastGPT v4.17.0 platform that parsed text extraction completely retains component hierarchies and operation paths.
- Test design documents in different formats (such as JSON exported from Figma, PDF exported from Axure) to confirm that parsing success rates for each format meet expectations.
- Adjust the
chunkSizeparameter, then compare chunking result coherence to ensure key operation steps are not split and review record paragraphs are not incorrectly merged. - Verify that uploaded file sizes do not exceed the
UPLOAD_FILE_MAX_SIZEconfiguration value to avoid triggering upload restriction errors.
The values provided are common starting points and should be measured against the reader's own samples.
Question material comes from public community discussions. Configuration values are common starting points and should be measured against your own samples. Verified on 2026-10-19.