How it works
In order to track improvements over time, we use a comprehensive versioning system. It's important to understand that instead of simply versioning the software, our approach encompasses the combined version of the entire pipeline, including the data production process, the data itself, and any API layer changes that affect the data output. Each version of the Damage Classifications AI is represented in the following format.
impacttriageai
NOTE: This format will be displayed as a column in the CSV output.
The version is comprised of three components, separated by hyphens:
- Product name
- Data version
- API calculation version
Using the example above impacttriageai-v1-2, here's a description of the components of the version.
Component | Name | Description |
|---|
impacttriageai
| Product name | Product name in lowercase without spaces. |
1
| Data version | Numerically ordered set of improvements that represent major enhancements in the pipeline that produces the data and/or the data itself. |
2
| API calculation version | Numerically ordered set of improvements that represent major enhancements on the API layer that impacts the data output. |
NOTE: In the context of version changes for both the Data Version and API Calculation Version, it is important to consider the impact on backwards compatibility. When comparing two versions, it's crucial to exercise caution when detecting changes, as the enhancements in the newer version may overshadow the actual on-the-ground modifications when searching for differences.
CSV Format Version
This version represents the design and formatting version in which the data is structured and presented. A "CSV Format Version" will be displayed on the endpoint as shown below:
https://api.nearmap.com/ai/damage/v1/tx/…