The Problem
When i run bddgen export i get all steps. But for my LLM Workflows or documentation i would need a bit more context then just the step names. For this we need maintanable comments to my steps.
Proposed Solution
The solution idea is as follows:
bddgen export does also look for JSDoc information above the Decorator.
/**
* Assert that the value of the given property on the queried track event is a valid UUID.
*
* @param property - The exact property name to check (as it appears in ClickHouse).
*/
@Then("the value of property {string} is an UUID")
async assertPropertyValueIsUUID(property: string) {
...
}
Then it creates a markdown table (as an example) and matches the Steps to the JSDoc information.
| Step Definition |
Description |
Then the value of property {string} is an UUID |
Assert that the value of the given property on the queried track event is a valid UUID. @param property - The exact property name to check (as it appears in ClickHouse). |
Now we can generate a step catalog that is always up to date. 👍
As a solution I prototyped and vibe prompted/coded a proof of concept straight into the node_modules of my installed playwright-bdd🗡️ 😆 - Although not really worth showing it did somewhat work though.
If I have time i will look into creating a propper MR. But wanted to share the idea first.
Additional Context
No response
Priority Level
Medium
The Problem
When i run bddgen export i get all steps. But for my LLM Workflows or documentation i would need a bit more context then just the step names. For this we need maintanable comments to my steps.
Proposed Solution
The solution idea is as follows:
bddgen export does also look for JSDoc information above the Decorator.
Then it creates a markdown table (as an example) and matches the Steps to the JSDoc information.
Then the value of property {string} is an UUIDNow we can generate a step catalog that is always up to date. 👍
As a solution I prototyped and vibe prompted/coded a proof of concept straight into the node_modules of my installed playwright-bdd🗡️ 😆 - Although not really worth showing it did somewhat work though.
If I have time i will look into creating a propper MR. But wanted to share the idea first.
Additional Context
No response
Priority Level
Medium