What data looks like for this category
Data for photovoltaic module encapsulation scenarios comes from cell grading records, glass adhesive film and backsheet batch ledgers, workshop MES systems, string welding and lamination equipment logs, and EL/IV tester output reports. Updates are created for each manufactured module, and batch information is synchronized for each batch of raw materials. Daily documents include layout drawings, material batch records, lamination process records, EL defect images, IV test reports, insulation withstand voltage records, and packing lists. Fields include cell current gear, lamination temperature and pressure parameters, EL crack location coordinates, IV power values, insulation withstand voltage values, and other relevant fields. Units include degrees Celsius, pascals, watts, millimeters, and other standard units.
What constraints these characteristics impose on tool calling and plugins
Multi-source heterogeneous data sources require tool calling to adapt to pulling and parsing different format data sources. For example, pull batch ledgers from MES systems, and export images and numerical reports from EL/IV testers. High-frequency updated single-module data requires tool calling to have real-time capabilities, to avoid data inconsistency caused by cached expiration. Diverse document types and professional fields require plugins to support image parsing, structured table extraction, and retention of professional fields, to ensure the accuracy of process documents and exception reports. Closely associated data characteristics require tool calling to support variable binding, to automatically associate fields from different data sources to report templates.
This page configuration applies to FastGPT v4.17.0.
How to set configurations
| Configuration Item | Recommended Setting | Rationale |
|---|---|---|
| Workflow/API tool selection | auto | Adapts to parallel calling of multiple plugins, covering processing of multiple types of documents including drawings, test reports, and equipment logs |
| Upload file size limit | 1000 MB | EL defect images and IV test reports for photovoltaic modules typically have large individual file sizes, this setting accommodates storage and parsing needs for standard single documents |
| Workflow/API function-call configuration | enabled | Enables precise tool calling, ensuring that parsing and extraction of professional fields align with photovoltaic encapsulation process standards |
| Tool call timeout period | 600 seconds | Parsing large EL images and batch test reports requires a long processing cycle, this setting prevents task interruption due to timeout |
| Variable reference switch | Enabled | Supports binding dynamic variables such as MES batch numbers and cell grading data, to automatically fill process documents and exception reports |
| MinIO storage path | /fastgpt/plugins/minio/photovoltaic | Separates file storage directories for photovoltaic encapsulation scenarios, facilitating subsequent archiving and retrieval |
Three common mistakes
- An error "file size limit exceeded" occurs when calling a locally deployed MinerU. The cause is that the upload file size limit parameter was not configured when starting the container, and the default 2 MB threshold was used instead.
- When uploading files in a custom workflow, only external links can be used, and battery cell batch variables cannot be referenced. The cause is that the variable reference switch for tool calling was not enabled, or the variables were not associated with the MES data source.
- EL image parsing results returned after plugin calling lack crack location fields. The cause is that a parsing plugin supporting image formats was not configured, or
toolChoicedid not specify the corresponding parsing tool.
How to confirm the configuration is complete
- Upload a single EL test report with a volume of approximately 1000 MB, check that tool calling completes normally, and no errors related to file size limits occur.
- Bind the battery cell batch number variable in the tool calling parameters, initiate a call, and check that the returned result automatically fills in the ledger data for the corresponding batch.
- Access the configured MinIO storage path, confirm that photovoltaic encapsulation process documents, exception reports, and test data have been stored in the corresponding directory according to the rules.
- Call the EL image parsing plugin specified by
functionCall, check that the returned result includes professional fields for photovoltaic encapsulation scenarios such as crack location and power correction value.
The values provided here are common starting points, and should be evaluated against the reader's own samples.