Each backtest record has a copy of its respective code. The backtest can be re-run with its original code, and if permitted, Q can review the code (e.g. as part of a help request from a user). This is the code that is carried forward into live trading, as well. So one hang-up with
import my_library is that there would be no revision control of the code associated with a given backtest and live trading instance (although it is not clear how Q ties revisions of imported libraries and their API to backtests and live trading instances...my guess is that they do not, but instead do QC checks by running backtests, as changes are made).
One solution, although still not pretty, would be to have automatic copying and pasting of my_library into the IDE. Every time a backtest is launched (or maybe every time the code is saved), a special text section would be replaced with the latest version of the contents of my_library (an automatic copy & paste).
Or maybe import my_library would work as expected for backtesting (pulling in the current revision when code is executed), but for live trading, the current revision of my_libary would be captured and associated with the live trading instance?
Within the research platform, perhaps import my_library could be supported?
At one point, Fawce commented that github (or github-like) integration was on the roadmap, but maybe now that Q is in bed with Point72, they have another angle. I have to think that if they are gearing up to manage $250M, someone is thinking about revision control and custom library import (if anything, from a legal/compliance standpoint). But then, this may be a completely separate system for "manager's" code, versus users who are trading their own money.
Also, I have to think that as part of the roadmap, the browser-based IDE may be scrapped in favor of a more integrated platform, using the IPython notebooks that are implemented in the research platform. It is probably irrelevant to the topic of this thread, but I thought I'd mention it.