What the data for this category looks like
The core data for this category comes from BOM, Gerber, and coordinate files provided by the client, paired with factory MES systems, solder paste and component batch ledgers, placement machines, SPI/AOI testing equipment, and reflow oven temperature collection data. Daily generated documents include changeover sheets, stencil archives, furnace temperature records, first article confirmation sheets, SPI/AOI defect reports, maintenance records, and traceability labels. Real-time data updates occur for every work order and production line changeover. Fields include part numbers, reference designators, component batches, furnace temperature parameters, testing coordinates, and some parameters carry physical units such as millimeters and degrees Celsius.
What constraints these characteristics impose during deployment and upgrade
Data sources for this category are scattered, and update frequency is high. During deployment, interface integration with multiple types of industrial equipment and systems must be supported, and storage parameters adapted to high-frequency, small-batch data writes must be configured. The need to parse multiple document types requires that parsing rules for corresponding file formats be configured in advance during deployment. Real-time updated work order data requires context isolation by production line and work order dimension to prevent cross-work order data confusion. Errors in the inspection link directly affect production line yield. During upgrades, a rollback path for the original verification logic must be retained to enable rapid recovery in case of exceptions. This page is written based on FastGPT v4.17.0.
How to Set Configurations
| Configuration Item | Recommended Value | Rationale |
|---|---|---|
DB_MAX_LINK | 100–150 connections | Adapt to concurrent writes and queries of high-frequency work order data, matching the load requirements of real-time production line data processing |
REDIS_URL | redis://:{custom password}@{redis host address}:6379/0 | Store real-time work order context cache, ensure low latency for high-frequency access; port 6379 is the standard Redis port |
STORAGE_S3_ENDPOINT | http://192.168.1.100:9000 (fill in according to actual private storage address) | Store multiple types of industrial documents and testing data, adapting to access requirements for local or private cloud object storage |
FILE_TOKEN_KEY | 32-character random string | Ensure authentication security for document upload and parsing interfaces, preventing unauthorized access |
PG_URL | postgresql://{username}:{password}@{pg host address}:5432/fastgpt_smt | Store structured work order and inspection data, ensuring stability of data queries; port 5432 is the standard PostgreSQL port |
ROOT_KEY | Set via actual testing | Manage system-level permissions, adapting to permission division requirements for multiple factory accounts |
Three Common Configuration Mistakes
- Symptom: When document parsing or model interfaces are called after startup, the response returns
{"detail":"Error message: ｜ "message": "403 This token does not have permission to use the model"}. Cause: The key and permission for the model service were not correctly associated in the deployment configuration, or the corresponding model was not added to the allowed use list. - Symptom: No parsing result is generated after uploading documents such as changeover sheets or SPI reports. Cause: Parsing rules for the corresponding file format were not configured, or the
FILE_TOKEN_KEYconfiguration was incorrect, causing authentication failure. - Symptom: Multiple accounts cannot log in normally, or permission division is abnormal. Cause: The
ROOT_KEYand multi-account permission association logic were not correctly configured, or database connection parameters were not adapted to multi-account data isolation requirements.
How to Verify Correct Configuration
- Run the
docker pscommand to confirm that all FastGPT-related containers are running, with no abnormal exits. - Upload a test BOM file, wait for parsing to complete, then check the field integrity of the parsing result to confirm it matches the preset parsing rules.
- Call the model interface for testing, verify that returned results meet expectations, and no 403 permission error occurs.
- Create multiple test accounts, verify that account login and permission division functions work normally, with no cross-account data access exceptions.
The values provided are common starting points and should be measured against the reader's own samples.