What the data for this category looks like
Data sources for this category include State-owned Assets Supervision and Administration Commission investment supervision systems, enterprise ERP and project management systems, industry market data, audit and financial statements. The update frequency is quarterly. Daily processed documents include annual investment plans, project proposals, feasibility study reports, due diligence reports, asset evaluation reports, and post-investment evaluation reports. The document structure covers approval, financial, risk, and post-evaluation fields across the full project approval workflow. Fields adjust based on document type, with no unified fixed units.
Constraints on the Sharing and Embedding Workflow
Data sources span multiple internal and external systems. The shared and embedded Q&A system must have permission to access the corresponding data sources. The quarterly update frequency requires the embedding component’s caching mechanism to match the update cycle, avoiding the return of outdated project data. The long document structure requires the Q&A system’s recall logic to accurately match project-related party building learning content, preventing interference from irrelevant information. Error consequences involve risks such as state-owned asset losses. The shared API must include a clear error feedback mechanism to facilitate troubleshooting.
Configuration Settings
Reference only; verify against the target FastGPT release and deployment documentation.
| Configuration Item | Recommended Setting | Rationale |
|---|---|---|
shareId | Exclusive ID copied from the FastGPT knowledge base configuration page | Used to uniquely bind the investment project approval and post-evaluation knowledge base for party building learning Q&A |
| Chunk length | 800–1200 characters | Adapts to the long text structure of feasibility study reports and post-evaluation reports, avoiding overly fragmented chunks or lost context |
| Similarity threshold | 0.75–0.85 | Filters low-relevance project document content, ensuring responses focus on party building requirements and project data related to investment |
| Single search limit | Top 6 results | Balances response information volume and system response speed, avoiding system overload |
| Cache duration | 7 days | Matches the quarterly update frequency, regularly refreshing knowledge base content to ensure data timeliness |
| Embedded domain whitelist | Fill in the domain names of internal enterprise business systems | Restricts embedding scope to prevent unauthorized external calls |
Three Common Configuration Errors
- Symptom: Some questions fail to recall knowledge base content when calling via API, but return results normally in the debug page. Cause: The
shareIdparameter is not included in the API request, or the knowledge base bound to the parameter does not contain party building learning documents related to investment projects. - Symptom: Images in the embedded chat window can only be viewed inside the iframe, and cannot be zoomed in or out externally. Cause: The external image pop-up configuration for the FastGPT embedding component is not enabled. By default, images are only displayed inside the iframe.
- Symptom: A cross-domain connection block error pops up when embedding a website. Cause: The allowed embedded domain whitelist is not configured in the FastGPT backend, so the browser blocks the cross-domain request.
How to Verify Successful Configuration
- Copy the
shareIdparameter, initiate a test request via the FastGPT-provided API debug tool, and verify that party building Q&A results related to investment projects can be recalled. - Open the web page containing the embedded chat window, send a test question with images, and verify that images can be zoomed in and out via external pop-ups.
- Access the embedded page from a non-configured domain, confirm that a cross-domain error appears, verifying the whitelist configuration is effective.
- Wait 7 days, refresh the knowledge base, access the embedded page again, and verify that responses are updated to the latest quarterly data.
The values provided are common starting points and should be measured against samples relevant to the specific deployment.
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.