What data for this category looks like
Data for this category primarily comes from Git repositories, CI pipelines, static scanning reports, and defect databases. Daily generated documents include merge requests, code review comments, commit records, and branch strategy descriptions. These documents contain fields such as merge request ID, commit hash, review comment number, and static scanning result identifier. Data updates occur daily or in real time alongside code commits. Each data entry is associated with specific code change scope and review node information. FastGPT v4.17.0 supports the standard data format for this category.
What constraints apply to citation source and traceability
Since data sources cover multiple tool types including code repositories and pipeline reports, citation traceability must link unique identifier fields such as commit hash and branch identifier. This ensures change sources can be traced. As data updates in real time with commits, a pull mechanism matching the update rhythm must be configured. This avoids referencing outdated review comments or merge statuses. Field formats vary across different document types. Field mapping rules must be established to match citation identifiers from different sources. This category includes manual review workflows. Traceability must distinguish between automated static scanning results and manual review comments. This ensures semantic consistency of citation sources.
How to configure the settings
| Configuration Item | Recommended Approach | Basis for This Approach |
|---|---|---|
quoteMaxToken | 800–1200 token | Matches the average content length of merge requests and review comments, to avoid truncating key citation identifiers |
responseData | Return full fields including merge request ID and commit hash | Core data identifiers for this category are merge requests and commit records. These must be fully returned for traceability |
detail | Enable full review comment traceability | Core code review information is scattered across multiple review comments. Full enabling covers all citation sources |
| Real-time pull trigger condition | Triggered per commit | Matches the data update rhythm tied to commits for this category, ensuring citation sources stay up to date |
| Traceability matching threshold | Calibrated based on actual testing | Adapts to the complexity of code review content across different teams, avoiding false matches or missed matches |
Three common configuration errors
- When calling the code review assistance interface, the returned result does not include the citation reference ID field. This occurs because the
responseDataconfiguration does not return the corresponding identifier fields, or the configured fields omit core traceability fields such as merge request ID and commit hash. - When viewing code review citation sources, a parameter error prompt appears. This occurs because the value of
quoteMaxTokenexceeds the maximum single-line length limit of the current code content, leading to parsing failure. - Only partial merge request citations are displayed in the code review assistance results. This occurs because the
detailfull traceability switch is not enabled, and only partial review comment data is pulled.
How to confirm correct configuration
Access the FastGPT configuration page to confirm the value range of quoteMaxToken matches the average length of current code review content. The content of a single merge request can be exported for verification. Call the test interface to check whether the returned result includes core identifier fields such as merge request ID and commit hash, confirming the responseData configuration is correct. Initiate a code commit to trigger a review, then check whether the traceability result covers all commit records and review comments to confirm the detail switch is enabled. Review system logs to confirm the real-time pull trigger condition matches the code commit rhythm, and no outdated data pull records exist.
The values provided are common starting points and should be measured 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 2026-10-19.