Disable writing data during learning process

[nick-harder]

-no more outputs written when learning
-adjusted learning dashboard

[nick-harder]

@maurerle I have also adjusted the learning dashboard, but I couldn't get the last two graphs to work - Bid price and bid volume, and also the noise. Could you pleae take a look at them? Ideally the bids ones should display all bids submitted by the learning agent. Thanks!

[codecov]

Codecov Report

Patch coverage: 93.54% and project coverage change: +0.07% 🎉

Comparison is base (5d9234d) 80.90% compared to head (157be69) 80.97%.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #205      +/-   ##
==========================================
+ Coverage   80.90%   80.97%   +0.07%     
==========================================
  Files          40       40              
  Lines        4058     4052       -6     
==========================================
- Hits         3283     3281       -2     
+ Misses        775      771       -4     

Flag	Coverage Δ
pytest	80.97% <93.54%> (+0.07%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Files Changed	Coverage Δ
assume/common/outputs.py	82.25% <87.50%> (+0.71%)	⬆️
assume/common/units_operator.py	91.35% <100.00%> (+0.54%)	⬆️
assume/markets/base_market.py	89.44% <100.00%> (ø)
assume/world.py	83.42% <100.00%> (+0.27%)	⬆️

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[maurerle]

I would find it much better, if we split this up by
output_agent_addr and
learning_output_agent_addr in the data_dict

As we are already checking if a output_agent_id and addr is given, we can use this to distinguish if we should write..?

then we can set this in world, that market and units operator do not have an output agent and units_operator only knows a learning_output_agent - therefore writes only learning..?
I can prepare a draft for this if this is unclear?

[maurerle]

Sorry, that's also great to not hardcode the learning_agent!
I just meant to provide the output agent twice with a different param, as it is used in a different context (writing learning output instead of other output)

[maurerle]

Looks good!
Thank you!