Review Next Design Models by Review Points
Overview
This sample demonstrates how to perform reviews on Next Design models using a workflow that follows specified review points.
It reads the target design model and the list of review points, then runs an AI agent iteratively for each combination of review point × model.
The extracted review findings are output to Excel by the AI agent.
This sample executes the following flow as a workflow:
- Retrieve the target for review
- Retrieve the review points
- Execute the reviews iteratively
- Aggregate the review results
- Output to Excel
This lets you verify the process of reviewing a Next Design model by review point while combining tool integrations and AI agents.
What you can do with this sample
- Perform reviews on Next Design models
- Check model contents according to review points
- Use a workflow to repeatedly run an agent per review point × model unit
- Output the review results in Excel format
Intended users
This sample is primarily intended for users who:
- Are considering integration between Next Design and AI
- Want to streamline design model reviews
- Want to experience repeated execution using workflows
Included contents
This sample includes the following example data.
You can run the sample as is to verify its behavior, and then replace the input/output files with your own project artifacts to try it in a setting closer to your actual work.
If you replace files, follow the steps described in "Steps to replace with your own work" and then perform the procedure in the "Procedure" section.
| Data | File name | Steps to replace with your own work |
|---|---|---|
| Workflow definition | NDモデルと観点ごとにレビューしてExcelに指摘を出力.yaml | Do not change |
| Agent definition | ND設計モデルレビュー_指摘JSON生成.yaml | Do not change |
| Input: design model | inputData/先進運転システムソフト開発.nproj Next Design model | Do not change |
| Input: review points | inputData/reviewPointList.json | 1. Place your own review-points file in JSON format into the inputData folder. 2. In step 6 of the "Procedure" section, specify the file path. |
| Document | This README | Do not change |
Using the sample data
Prerequisites
- The meta-model of the Next Design design model to be reviewed must be the same as in this sample.
- The data structure of the review points must match the structure used in this sample.
- The target Next Design project must be opened.
- The Excel file to receive the findings must be open and active.
Inputs and outputs
- Input: Next Design design model, review points file
- Output: Excel file listing review findings
Procedure
- Open the target Next Design project.
- In the Next Design model, select the top element of the review target scope.
Example:先進運転システムソフト開発/ソフト設計/ADAS ECUソフト構造/APP/車速偏差/目標車速演算 - Open the Excel file for outputting review results and make it the active workbook.
- In the DC Agentiqs workflow screen, click "NDモデルと観点ごとにレビューしてExcelに指摘を出力".
- In the "NDモデルと観点ごとにレビューしてExcelに指摘を出力" dialog, click "Run".
- When a dialog appears, enter the file path for the review points list and click "Run".
- Review findings will be written to the Excel file for each finding.
Targets
- Workflow: NDモデルと観点ごとにレビューしてExcelに指摘を出力
- Agent: ND設計モデルレビュー_指摘JSON生成
Limitations
- This sample processes combinations of review points and target models. As the number of review points and target models increases, the number of executions will also increase.
- Review points must be usable in the format expected by the workflow and agent. If the structure or contents differ from the expectation, the workflow will need to be modified.
- This sample targets Next Design design models. If the structure differs, the workflow will need to be modified.
Reference: Workflow execution flow
This sample decomposes processing into stages using a workflow so you can verify a review process that combines tools and AI agents.
- Read the review points list file and shape it into a structure that can be iterated per review point.
- Read the Next Design design model.
- For each design model and each review point, input to the review agent in order and extract review findings.
- Output the review findings to Excel.