What is the roadmap for this InfluxDB library ...

[michealroberts]

Having reviewed this client library, I found it to be poorly structured and lacking adequate type support.

The documentation is also out of date.

The dependencies on pandas and pyarrow also seem unnecessary.

Are there any plans to improve this client library, streamline the dependencies, and give it some overall love?

If not, I'm happy to help contribute to this effort.

[bednar]

@michealroberts Every PR is welcome, and we’ll be happy to review it. Feel free to contribute to our codebase! ❤️

[karel-rehor]

@michealroberts
We have an item in our backlog to refactor this library. However customer related items take precedent. I've seen there is a lot to be fixed, and a product management decision about introducing breaking changes during the refactoring process will need to be made.

Removing pandas support could break the library for other users. The query API however is bound tightly with Arrow Flight.

That said as Jakub stated above, all contributions are useful and welcome.

[michealroberts]

@karel-rehor We're in 0.y.z territory (initial, unstable development where the public API is not yet defined).

I think we can break things if needed, if it is a concern, move to v1.

The pandas dependency is bizarre, let the user decide on what data format they want to use.

You need to have the notion of models, nothing is modelled in or out of the db. I'll have a think and present the ideas.

[karel-rehor]

@michealroberts While the releases for this library are still at 0.x.y, the InfluxCommunity is offering client libraries as a support to all users of the Influx databases. This includes some paying customers, who already use them as dependencies in their pipelines. I agree that the technical debt in this library is high and unfortunately growing. Still, I think product management at Influx would prefer caution regarding breaking anything, that could negatively influence their subscribers' processes.