What this category of data looks like
The data for this category has four fixed sources: territorial spatial basic information platform, natural resource survey and monitoring data, population and economic statistics, and remote sensing images. The update frequency is annual. Daily produced documents include territorial spatial overall plans, detailed plans, special plans, planning implementation evaluation reports, public opinion summaries, and more. Document structures cover fields such as text descriptions, spatial coordinates, area indicators, and statistical values. Common units include hectares, decimal longitude and latitude, ten thousand people, and others. A single planning document can cover associated content from multiple data sources.
What constraints do these characteristics impose on the citation source and traceability link?
Annual updated data sources require the knowledge base to perform full synchronization on a regular basis. This ensures cited content aligns with the latest policies and monitoring data. Mixed citations of multi-source data require accurate marking of the original source platform and dataset for each entry. This avoids confusion of statistical results from different channels. Complex document structures and spatially associated fields require traceability information to include both text paragraphs and corresponding spatial ranges. This meets compliance requirements for planning approval. Errors can easily lead to delays in planning approval, so the traceability link must be fully traceable.
How to Configure
| Configuration Item | Recommended Setting | Rationale |
|---|---|---|
quoteMaxToken | 800–1200 token | Territorial spatial planning documents have long single-paragraph citation content, requiring coverage of complete data descriptions and spatial ranges |
responseData | Enabled | Full traceability of original data sources and entries is required to match compliance requirements for planning approval |
detail | Enabled | Specific fields of cited data such as coordinates and area must be displayed to meet the rigor of planning documents |
| Single search limit | Top 6 entries | Covers main source datasets, avoids redundant information interfering with planning text writing |
| Context window length | 15000–20000 characters | Planning documents have long length, requiring retention of contextual association for multiple rounds of citations |
Three Common Mistakes
- Phenomenon: Generated citation content only displays data IDs, without showing specific fields and source descriptions. Cause: The
detailconfiguration item is not enabled, and the detail display function is not activated. - Phenomenon: After setting the variable citation mode, the temperature setting button disappears, and generation parameters cannot be adjusted. Cause: In FastGPT v4.17.0, variable citation mode locks generation parameters by default. Temperature parameters must be configured via preset templates.
- Phenomenon: The number of cited data entries is insufficient, failing to cover multi-source data required for planning. Cause: The single search limit is set too low, not matching the coverage requirements of multi-source data.
How to Confirm Configuration is Complete
- Generate a citation snippet for a single planning document, check that it includes original data sources and specific field information.
- Test the variable citation mode, confirm that generation parameters can be adjusted via preset templates, and restore the temperature setting function.
- Verify that the number of cited data entries matches the expected coverage range, and can match the multi-source data required for planning.
- Check the token count of cited content, confirm it falls within the 800–1200 token range to meet document display requirements.
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 2026-09-28.