What Data for This Category Looks Like
Data sources include engineering orders, sand, gravel, cement and admixture inspection records, aggregate moisture content measurement data, mixing plant control system logs, and laboratory test reports. Update frequency follows per-vehicle production batches, per raw material deliveries, and daily summaries. Daily documents include mix proportion notices, opening identification forms, production and transportation orders, raw material inspection records, test block retention records, and strength test reports. Fields include mix proportion parameters, aggregate moisture content values, slump indicators, and test block strength results. Each parameter has a clearly defined unit of measurement: mix proportion uses kilograms per cubic meter, slump uses millimeters, and test block strength uses megapascals.
What Constraints Do These Characteristics Impose on the "Tool Calling and Plugins" Workflow
High data update frequency and multi-source coverage require tool calling to support real-time pulling of the latest production and inspection data. This prevents use of expired mix proportion or moisture content parameters. Multi-source data originates from laboratories and mixing plant systems. Plugins must connect to the interfaces of both data source types separately. Data must be unified and organized to generate process documents that meet format requirements. Exception reports must be linked to test block strength results. Tool calling must trigger the strength report query interface to ensure accurate exception determination bases. Each process document must correspond to a single vehicle’s production parameters. Tool calling must support precise data matching via batch ID to avoid mixing production records from different vehicle trips.
Configuration Settings
This configuration is based on FastGPT v4.17.0.
| Configuration Item | Recommended Setting | Rationale |
|---|---|---|
| Workflow/API tool selection | auto | Supports automatic selection of tool calling mode, adapting to multi-data-source process document generation needs |
| Workflow/API function-call configuration | enabled | Enables explicit function calling, ensuring that tool calling parameters accurately match the field requirements of concrete production |
| Plugin call timeout | 60–90 seconds | Single-vehicle production data query and report generation must be completed within a reasonable time frame, avoiding timeouts that disrupt production scheduling |
| Data source matching rule | Match by batch ID | Each vehicle's production data corresponds to a unique batch ID, ensuring that parameters associated with process documents are accurate and error-free |
| Number of returns per single tool call | Top 3 entries | Process documents and exception reports only require the latest 3 sets of relevant data to complete generation, no redundant information needed |
| Exception report trigger condition | Determine based on measured deviation values | Trigger tool calling by combining the deviation between measured values of slump and moisture content and standard values to generate exception reports |
Three Common Misconfigurations
- Symptom: Content extraction nodes return empty fields, or tool calling fails to trigger. Cause:
toolChoiceandfunctionCallparameters are not configured correctly, or their values are set incorrectly. - Symptom: Plugins cannot pull online data, with error prompts indicating connection timeout or resource unavailability. Cause: A domestically available network proxy plugin is not configured, and reliance on Google-related plugins prevents access to target data sources.
- Symptom: Tool calling returns a 401 status code. Cause: Plugin secret key parameters are not configured correctly, or secret key permissions do not match FastGPT calling requirements.
How to Verify Successful Configuration
- Review FastGPT tool calling logs to confirm that
toolChoiceandfunctionCallconfiguration values match preset settings. - Input a single vehicle’s batch ID in a simulation to trigger process document generation, and verify that returned results include the batch’s corresponding data such as mix proportion and moisture content.
- Input abnormal parameters (such as slump deviation exceeding standard ranges) in a simulation, and confirm that the plugin triggers the exception report generation workflow.
- Check plugin call response times to confirm they do not exceed preset timeout thresholds.
The values provided are common starting points and should be evaluated against local samples.