What the data for this category looks like
Core data for this category comes from code repositories, test framework reports, requirement specifications, and interface definitions. The update frequency is daily or after each code commit. Daily produced documents include test cases, coverage reports, Mock instructions, and defect tickets. Test cases contain fields for case ID, preconditions, execution steps, and expected results. Coverage reports include fields for file path, coverage value, and associated test cases. Mock instructions contain fields for interface address, request parameters, and return examples. Most data uses structured fields, and some documents include binary-format test log attachments.
What constraints do these characteristics impose on source citation and traceability workflows
Data sources are scattered across code repositories and test frameworks. Traceability requires linking to specific commit hashes or build records to avoid confusion across data sources. The high update frequency requires the traceability system to support near-real-time data source synchronization, ensuring that cited test data matches the current code state. Daily documents have many structured fields, so citations must accurately match unique identifiers such as case IDs and file paths to avoid traceability errors caused by generalized matching. The test data construction process takes significant time. Traceability must reduce repeated lookup steps, and errors may lead to missed local logic checks. As a result, traceability accuracy directly impacts test quality.
How to configure settings
| Configuration Item | Recommended Setting | Rationale |
|---|---|---|
| Maximum Context | Top 8000 tokens | Matches the average length of documents such as test cases and coverage reports, avoids truncating critical test steps and interface definition content |
| Single Search Limit | Top 10 entries | Controls the total volume of recalled test data, balances traceability accuracy and interface response speed |
| Reranking | Enabled | Prioritizes documents more relevant to the current test task, optimizes the relevance of traceability results |
| Query Optimization | Enabled | Performs semantic optimization on test-related queries, improves recognition accuracy for identifiers such as test case IDs and interface paths |
quoteMaxToken | 500–800 tokens | Limits the length of cited document chunks, ensures traceability information is concise and does not exceed the context window limit |
responseData | Return only associated fields | Returns only necessary traceability fields such as test case IDs and file paths, reduces interference from redundant information on queries |
detail | Full | Retains complete execution steps and expected results for test cases, ensures traceability information can be fully traced back |
Three common configuration mistakes
- Phenomenon: The returned traceability information does not link to the corresponding test case variables. Cause: The data source binding for variable references was not configured correctly, preventing structured fields for test cases from being retrieved.
- Phenomenon: The cited document chunk exceeds the expected length. Cause: The
quoteMaxTokenparameter was not set correctly, resulting in recalled test data that is too long and exceeds the context window. - Phenomenon: Interface Q&A requests time out. Cause: The single search limit was not configured, resulting in the recall of too many irrelevant test framework reports and defect tickets, leading to context loading timeouts.
How to confirm configuration is complete
- Run an interface Q&A request that includes a test case ID, check if the returned results include the corresponding traceability document link or unique identifier.
- View the FastGPT v4.17.0 traceability log to confirm that the cited document source points to the correct code repository path or test framework report.
- Adjust the
quoteMaxTokenparameter to different ranges, verify that the length of the returned cited document chunks matches the set value. - Disable the reranking configuration, compare the relevance of the two recall results to confirm the reranking configuration has taken effect.
The values provided 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 2026-10-19.