Visually exploring data / rules (charts)

My workflow for trading ideas and research is identifying a pattern visually, then writing the code to explore it for edges and finally a trading system. Very typically this means iterating on the code itself and visually confirming the signals are correctly matching the intended pattern.

When working outside of QuantRocket, I’ve done this by creating interactive UIs (typically using DearPyGui or similar) to explore signals alongside price and volume charts. As I understand it, I won’t be able to launch these types of interactive visual GUIs from within the QuantRocket code running environments (either Jupyter or VSCode with remote extensions).

What options do I have for being able to write code that can leverage the data / libraries in QuantRocket, but runs in my Windows context so I can continue to use my UIs?

That’s possible, per the docs:

https://www.quantrocket.com/docs/#deploy-connect-from-other-applications

Thanks Brian - yep I’ve used the external client, but this doesn’t seem to address being able to use the actual QuantRocket research APIs outside of the Docker context. For example, calling get_prices and getting the dataframe, or using the Pipeline tool, or using get_data from Zipline. Is it correct that these are unsupported outside of the container?

FWIW I’m very technical and happy to get under the covers. I can see where all the data mounts are etc for the zipline bundles, and so on, and I know how to expose Docker services to my local machine. What I don’t know is whether the actual Zipline / Moonshot libraries will allow themselves to be reconfigured in this way.

You can conda list and conda activate zipline && conda list the jupyter environments and replicate them as needed in your local environment. There is nothing proprietary in the jupyter container. Or you can install your additional libraries in the jupyter container and try that approach.

I get the sense you may be asking prematurely. I don’t have a blueprint for your situation but based on other users’ experiences, you should just poke around and see what works for you and I think you’ll find QuantRocket is pretty malleable.

1 Like

Thanks Brian that does help. I was thrown off by the fact that zipline==2.5.0.0 isn’t actually available in PyPi or any conda channel, but it seems like I can just directly install it from https://github.com/quantrocket-llc/zipline if need be.

Yes I was asking prematurely so appreciate the guidance so far.