APIOfficial documentation10 min readAPI

Configure FastGPT Dataset Collection Creation Parameters

This document outlines the required and optional parameters for creating a dataset collection via the FastGPT OpenAPI, including complete request paramete…

Overview

This document outlines the required and optional parameters for creating a dataset collection via the FastGPT OpenAPI, including complete request parameter specifications and standard response fields. All details are pulled directly from the official FastGPT OpenAPI dataset documentation.

Request Parameter Reference

The following table lists all available request parameters for dataset collection creation, with their purpose, requirement status, and default behavior where applicable:

ParameterDescriptionRequired
datasetIdUnique identifier of the target dataset
parentIdParent directory ID. Defaults to the root dataset directory if not provided
trainingTypeData processing method. Accepts chunk (split by text length) or qa (Q&A extraction)
indexPrefixTitleToggles automatic generation of title-based indexes
customPdfParseToggles enhanced PDF parsing. Defaults to false (disabled) when unset, enables enhanced parsing when set to true
autoIndexesToggles automatic index generation. Only available for commercial FastGPT versions
imageIndexToggles automatic image indexing. Only available for commercial FastGPT versions
chunkSettingModeChunk parameter configuration mode. Accepts auto (system default) or custom (manual specification)
chunkSplitModeChunk splitting strategy. Accepts size (split by fixed length) or char (split by custom characters). Only active when chunkSettingMode=custom
chunkSizeTarget size for each text chunk, defaults to 1500. Only active when chunkSettingMode=custom
indexSizeTarget size for generated indexes, defaults to 512, must be less than the embedding model’s maximum token limit. Only active when chunkSettingMode=custom
chunkSplitterCustom highest-priority split symbol. Will not split further unless exceeding file processing context limits. Only active when chunkSettingMode=custom
qaPromptCustom prompt template for Q&A pair extraction, used when trainingType=qa
tagsArray of string tags to assign to the collection
createTimeFile creation timestamp, accepts Date object or string format

Standard Response Fields

Upon successful creation of the dataset collection, the API returns two core fields:

  • collectionId: Unique alphanumeric identifier for the newly created collection
  • insertLen: Integer count of successfully inserted text chunks

Key Usage Notes

All mandatory parameters must be included in the request to avoid validation errors. Optional parameters will use their default values if omitted. Commercial-exclusive parameters will not function on non-commercial FastGPT deployments. Chunk configuration parameters only apply when chunkSettingMode is set to custom.

Source: FastGPT official source

Applicability and version scope

Use this page for the documented API scenario. Confirm the FastGPT, dependency, API, and deployment versions in the official source before applying a change.

Safety guardrails

Use [REDACTED_CREDENTIAL] for credentials and private data. Confirm the documented environment and version before review.

Rollback guidance

Restore the prior technical-content authority snapshot. Restore saved configuration and data snapshots, then repeat the smallest verification scenario.