What the data for this category looks like
Oil and gas exploration and development data primarily comes from real-time data collected by on-site drilling sensors and logging equipment, plus historical materials archived by work area exploration teams. Data update schedules fall into two categories: real-time on-site data pushed during drilling operations, and archived materials compiled once per quarter. Document structures include structured single-well data tables, as well as unstructured seismic profiles and core scan images. Single-well data fields include well ID, drilling depth, resistivity value, porosity value, with units of none, meters, ohm·meters, and dimensionless respectively.
What constraints these characteristics impose on tool calling and plugins
Real-time on-site data requires tool calling to support low-latency local data source reading, to avoid delayed data impacting operational decisions. Mixed structured and unstructured document types require plugins to support both table parsing and image recognition capabilities, to adapt to different formats of exploration data. The special nature of fields and units requires tool calling to strictly retain the original data units and formats, to avoid operational risks caused by parsing errors. Frequently updated real-time data requires plugins to support incremental pull functions, to reduce resource consumption from full data transfers. Large-volume seismic profiles and core images require the tool calling upload file limit to adapt to actual needs; the default small-volume limit cannot meet these requirements.
How to set configurations
| Configuration Item | Recommended Setting | Rationale |
|---|---|---|
toolChoice | Set to auto | Supports automatic selection of appropriate tools, adapts to mixed structured and unstructured exploration data processing |
functionCall | Enable forced mode | Ensures tool calling parameters strictly match the field formats of exploration data, avoids parsing errors |
| Upload file size limit | 1000–2000 MB | Files such as seismic profiles and core scan images typically have large volumes; the default 2 MB limit cannot meet requirements, adjust based on actual measurements |
| Custom workflow parameter reference | Enable global variable reference | Parameters such as well IDs and work area names in exploration operations require dynamic replacement; supporting variable reference simplifies the configuration process |
| minio storage configuration | Set storage bucket lifecycle to 30 days | Temporary data from exploration sites requires regular cleanup to avoid excessive storage resource usage |
| Plugin marketplace template calling | Prioritize templates marked for oil and gas exploration use | Specialized templates already adapt to the field formats and units of exploration data, reducing additional adjustment work |
Three common misconfigurations
- When calling local MinerU to process exploration data, a 413 status code error is returned indicating file size limit exceeded. Cause: The upload file size configuration during container startup was not adjusted; the default limit is 2 MB, which cannot adapt to large-volume exploration files such as seismic profiles.
- After configuring upload file parameters in a custom workflow, the tool call cannot read dynamically generated file path variables, and only fixed links are supported. Cause: The variable reference switch for the custom workflow was not enabled; parameter binding only supports static link configuration.
- Exploration data stored in minio cannot be read normally by plugins, and logs show connection timeout. Cause: The minio access key and storage bucket permissions were not configured correctly, or network policies restricted communication ports between the FastGPT container and minio.
How to confirm the configuration is correct
- Upload files matching the actual scale of exploration files, check that tool calls can receive and process them normally with no error returns.
- Configure upload parameters including well ID variables in a custom workflow, run the workflow, and check if the tool call input parameters correctly replace the variable values.
- Log in to the minio console, check if new exploration data storage files have been generated, and confirm that the storage path and permission configurations are correct in the FastGPT plugin logs.
- Call the tool calling interface configured with
toolChoiceset toautoin the FastGPT v4.17.0 environment, check if it can automatically select appropriate tools to process mixed-format exploration data.
The values provided here are common starting points, and should be evaluated against the reader's own samples.
Question material comes from public community discussions. Configuration values are common starting points and should be measured against your own samples. Verified on 2026-09-28.