Clean the pymorize logs

[mandresm]

Make a user-friendly log. At the moment the current log is too verbose. We could have a more complete log and a simpler one.

[pgierz]

Before I switch this from "Task" to "Feature", we need to clarify a few things.

Technical:

• The logger mechanism needs to be reworked/unified. At the moment, logs end up in several different files, sometimes via prefect's logger, sometimes via loguru, and sometimes with click's interface to loguru. In the end, I would like to always have this:

from pymor.core.logging import logger

logger.info("My message for users")
logger.debug("My message for completeness and hard-core debugging")
logger.success("Something worked and is complete")
logger.exception("Something went wrong. A custom message is displayed and traceback is dumped")
logger.error("Something went wrong. Just a message")

Implementing this logger function to cleverly push around the messages to the correct places is something loguru sells itself as being good at. Adding in Prefect makes this slightly more complicated, it pollutes the logs with its own (sometimes helpful) things.

• I'll need to write a few examples and add them to the developer section of the handbook. Someone else needs to test them for understandability.

User Experience

• What needs to be logged? What is too much?
• How to control verbosity (quiet mode)? There are already a couple of FIXME comments about this...
• Environment variable control?

[siligam]

Recently, I am using markdown style in the logging for each feature used in the core or std_lib. here are some examples

[Temporal Resolution Check]
  → Inferred Frequency     : M (Δ ≈ 30.0000 days)
  → Target Approx Interval : 30.0000 days (~M)
  → Comparison Status      : equal
  → Valid for Resampling   : ✅

[Freq Check] Test Series
  → Inferred Frequency : None
  → Step Multiple      : None
  → Median Δ (days)    : None
  → Regular Spacing    : ❌
  → Strict Mode        : ✅
  → Status             : too_short

[Time bounds] unnamed_dataset 
  → is dataset:   ✅             
  → time label  : time           
  → bounds label: bnds      
  → has time bounds: ❌      
  → time values  : 5 points from  2000-01-01 to 2000-01-05                     
  → time step    : 86400000000000  nanoseconds                                        
  → set time bounds:   time_bnds(5, 2)                                    
  → bounds range   :    2000-01-01 to 2000-01-06                      
--------------------------------

Maybe such a thing is sufficient for user side logging. Every feature has its own section and minimal information that users may be interested to know.