What the data for this use case looks like
Business data for photovoltaic module encapsulation scenarios comes from multiple sources: cell sorting systems, glass adhesive film backplane batch ledgers, MES production systems, string welding and lamination equipment, EL/IV testers, and others. Daily processed documents include layout diagrams, material batch records, lamination process records, EL images, IV test reports, insulation pressure resistance records, and packing lists. Corresponding records are generated for each module and each batch of materials, which defines the data update frequency. Each document includes structured and unstructured fields such as material batch numbers, current gear, lamination parameters, and test results.
Constraints on deployment and upgrade workflows
The need to sync multi-source heterogeneous data requires the deployment phase to support integration with multiple data source interfaces, and adapt to document upload and parsing for different formats. High-frequency per-module data updates require the upgrade phase to ensure uninterrupted service or fast switching, to avoid disrupting real-time review in production workflows. Scenarios with concurrent access from multiple departments require configuration options to adjust the maximum concurrent connections. Document types include graphics and structured tables, so deployment must adapt parsing rules for different file types. Errors directly impact product power output and customer claims, so post-upgrade review logic must be validated before full deployment.
How to configure settings
This document is written based on FastGPT v4.17.0.
| Configuration Item | Recommended Setting | Rationale |
|---|---|---|
PG_URL | postgresql://fastgpt:securepass@10.0.0.5:5432/pv_doc_review | Structured batch records and test reports for photovoltaic module encapsulation are suitable for stable storage and querying with PostgreSQL |
REDIS_URL | redis://:redispass@10.0.0.6:6379/1 | Cache real-time updated module data and session information to adapt to high-frequency access scenarios |
FILE_TOKEN_KEY | 32-bit random hexadecimal string | Protect document upload and authentication to prevent unauthorized access to production documents |
DB_MAX_LINK | 20–30 | Adapt to concurrent access requirements from multiple departments including string welding, lamination, and testing |
FE_DOMAIN | https://pv-doc-review.manufacturing.example.com | Bind a dedicated deployment domain name to distinguish FastGPT services for other use cases |
ROOT_KEY | 64-bit random secure key | Provide highest privilege authentication for system management, ensuring the security of configuration modifications |
Three common configuration errors
- Phenomenon: The identity authentication function cannot be enabled after deploying the open-source version. Cause: The
FILE_TOKEN_KEYorROOT_KEYparameters are not configured correctly, and the corresponding authentication switch is not enabled. - Phenomenon: After upgrading FastGPT, workflow AI conversations cannot obtain document content output by code runs. Cause: The vector database index pointed to by
PG_URLwas not updated synchronously, or the new document storage path rules were not adapted. - Phenomenon: The fastgpt-sandbox image cannot be pulled during Docker deployment. Cause: The correct image repository address was not configured, or network permissions for image pulling were not enabled.
How to confirm proper configuration
- Run
curl ${FE_DOMAIN}/api/health; a 200 status code indicates the base service is operating normally. - Upload an EL image test file for a photovoltaic module, verify that the authentication process passes and the file is parsed correctly.
- Check the database corresponding to
PG_URL, confirm that batch records and test data have been written synchronously according to preset rules. - Trigger a document cross-review task, verify that the returned result matching logic conforms to preset verification rules.
The values provided are common starting points and should be measured against the reader's own samples.