What this category's data looks like
Feed data sources primarily include formula management systems of feed manufacturers, feeding logs of livestock farms, and inbound documents for raw material trade. Update rhythms vary:
- Formula documents are updated quarterly to align with raw material market changes
- Feeding logs are generated daily per breeding batch
- Raw material ledgers are updated in real time upon purchase inbound
Each document typically includes raw material lists, individual raw material addition amounts, nutritional indicators such as metabolic energy, applicable livestock and poultry categories, production batch numbers, and shelf life durations. Field units include kilograms per 100 kilograms of feed, megajoules per kilogram, days, and no unified single-scenario format exists.
What constraints do these characteristics impose on tool calling and plugins
The multi-source, heterogeneous nature of feed data requires tool calling plugins to support connections to different types of data source interfaces, with separate adaptation for the interface formats of formula systems, feeding logs, and raw material ledgers.
Data sources with different update rhythms require differentiated synchronization strategies: real-time feeding records need short-cycle polling, while static formula libraries can use long-cycle incremental synchronization.
The multi-layer nested structure of documents requires plugins to support structured parameter parsing, which cannot be handled solely with plain text rules.
The diversity of field units requires plugins to include built-in unit conversion logic, to avoid parameter errors caused by unit mismatches during calls.
The presence of associated fields such as batch numbers requires that dynamic filter parameters be included during tool calls, to ensure returned data corresponds to the specified breeding scenario.
Configuration Settings
| Config Item | Recommended Value | Rationale |
|---|---|---|
UPLOAD_FILE_MAX_SIZE | 4096 MB | Feed marketing content often includes multi-ingredient ratio tables and batch traceability documents, and single file size usually does not exceed 4 GB |
PARSE_FILE_TIMEOUT_SECONDS | 300 seconds | Feed documents contain multi-layer nested raw material lists and nutritional parameters, which take longer to parse structurally |
WORKFLOW_UPLOAD_VAR_ENABLE | Enabled | Feed marketing scenarios require dynamic association with variables such as breeding batches and livestock and poultry categories. Supporting variable reference uploads adapts to dynamic requirements |
MINIO_STORAGE_PATH | /fastgpt/plugins/feed-marketing | Uniformly store marketing materials generated by plugins and parsed feed documents, to facilitate subsequent traceability and calls |
PLUGIN_REQUEST_TIMEOUT | 120 seconds | Sufficient time for data return must be reserved when calling external feed raw material databases and livestock breeding data interfaces |
The parameter values provided on this page are common starting points for configuration. Actual values are affected by material form, data volume and business rules. Specific issues require specific analysis, and it is recommended to test on your own samples before finalizing.
Three Common Configuration Mistakes
- Phenomenon: When the local MinerU plugin is called, uploading a file triggers a 413 Request Entity Too Large error, or parsing fails after upload. Cause: The
UPLOAD_FILE_MAX_SIZEconfiguration was not modified during container startup. The default limit is 2 MB, which cannot adapt to the size of feed documents. - Phenomenon: When an upload file node is configured in a custom workflow, predefined breeding batch or livestock and poultry category variables cannot be selected, and only static public links can be entered. Cause: The
WORKFLOW_UPLOAD_VAR_ENABLEconfiguration was not enabled. By default, only static link file uploads are supported. - Phenomenon: A large number of invalid files appear in the MinIO storage bucket, or parsed feed documents cannot be retrieved. Cause: The
MINIO_STORAGE_PATHconfiguration was not specified, leading to chaotic file storage paths, or storage bucket read/write permissions were not configured, causing write failures.
How to Confirm Configuration is Successfully Applied
- Run the local MinerU container startup command, check if the startup log contains the
UPLOAD_FILE_MAX_SIZEconfiguration parameter, and confirm that the parameter value matches the preset value. - Create a new upload file node in a custom workflow, attempt to bind a created business variable, confirm that the node supports variable reference and static link input.
- Upload a test feed formula document to the plugin upload node, log in to the MinIO management backend, confirm that the document has been stored to the preset plugin-specific path.
- Call the external feed raw material database plugin, initiate a test request, check if the interface return result includes correct nutritional parameters and unit conversion results.
Question material comes from public community discussions. Configuration values are common starting points and should be measured against your own samples. Verified on 2026-09-14.