What the data for this category looks like
Data for this category mainly comes from the national land spatial planning one-map, project approval and supervision platform, land use status database, and ecological protection red line data. Update frequency is irregular. Daily processed documents include land pre-examination and site selection opinions, construction land planning permits, construction engineering planning permits, red line drawings, and planning conditions. Fields cover project site selection scope, land scale, planning control requirements, prior approval conditions, and more. There are no unified fixed units. Some fields require matching corresponding values based on actual project scenarios.
What constraints these characteristics impose on model access and configuration
Data sources are scattered, and updates follow no fixed cycle. Model access must support dynamic pulling of multi-source heterogeneous data. This avoids data lag caused by reliance on static caching. Document types are diverse, and fields have no unified format. Parsing logic that adapts to multiple document structures must be configured. This prevents field matching deviations. Business results of land pre-examination and planning permission are directly linked to project approval and construction. Model output must ensure accuracy. Configuration links must strictly control parameters related to generation accuracy. A data verification link must also be reserved to meet business compliance requirements.
How to set configurations
| Configuration Item | Recommended Setting | Rationale |
|---|---|---|
maxContext | 8000–12000 token | Adapts to long-text context requirements of multi-source planning documents, covers all associated data across the project approval process |
maxTokens | 2000–3000 token | Meets the full output length required for business process guidance, avoids truncating key compliance explanations |
quoteMaxToken | 1500 token | Limits the length of referenced planning document fragments, ensures referenced content is strongly relevant to business questions |
toolChoice | auto | Automatically matches planning data query tools that need to be called, adapts to multi-source data pulling requirements |
temperature | 0.1–0.3 | Reduces generation randomness, ensures accuracy and compliance of business guidance content |
AIPROXY_API_ENDPOINT | Fill in according to the actual deployment address of the platform | Adapts to the interface path for internal data pulling, ensures access to data sources such as national land spatial planning |
AIPROXY_API_TOKEN | Fill in according to the authorization credential | Completes interface identity verification, ensures security of multi-source data access |
Reference only; verify against the target FastGPT release and deployment documentation.
Three common mistakes
- Symptom: Model stream response is empty. The interface prompts "Model stream response is empty, please check if the model stream output is normal". Cause:
AIPROXY_API_ENDPOINTorAIPROXY_API_TOKENis not configured correctly. This prevents the model from normally pulling planning data to generate a response. - Symptom: After migrating the knowledge base, the conversation prompts that the knowledge base is empty, but all content can be viewed on the knowledge base page. Cause: The
maxContextparameter is not set correctly. This causes the context window to be insufficient to load the full knowledge base content, so the model cannot retrieve associated data. - Symptom: Calling the MCP tool returns a 400 error, but normal operation occurs without calling the tool. Cause: The
toolChoiceparameter is configured incorrectly. This causes the model to attempt calling an incompatible tool interface, leading to abnormal request format.
How to confirm the configuration is complete
- Manually trigger a business process guidance test that includes project site selection. Check if the model output includes compliant handling steps and data verification requirements.
- View interface logs. Confirm that requests for
AIPROXY_API_ENDPOINTandAIPROXY_API_TOKENcan return valid content from data sources such as national land spatial planning and project approval. - After migrating a test knowledge base, initiate a query related to red line drawings and planning conditions. Confirm that the model can correctly reference corresponding document content in the knowledge base.
- Adjust the
temperatureparameter to around 0.2, generate the same guidance content three times. Check if output consistency meets business accuracy 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.