Improve CePO capability

[ohpauleez]

Thanks for all the great work on this project. I've been collecting together small changes I've been making locally as I've been using the proxy (and the various approaches).

This PR focuses on improvements to CePO.

• Fixes the variable renaming bugs
• Adds logging to bring it inline with the other modules
• Adds a new option to turn off thinking level (for the models that don't support it); Updates made to the config files
• Robustly handles errors where thinking level wasn't supported (and will disable it)
• CePO collects responses as Strings (with the think-block removed) -- updated response-logging to accept that format
• When debug log level is enabled, deep exceptions will be logged in the main server loop

[CLAassistant]

All committers have signed the CLA.

[CLAassistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}

[ohpauleez]

Apologies for the whitespace edits (I have my editor set to clean up lines of whitespace and trailing whitespace).

It seems the tests are failing due to access to models. I'm happy to help here if you provide direction (perhaps you already know how to fix this).

There are some other items that probably should be cleaned up in the code:

• server.py imports traceback, but I don't think it's needed (I commonly use it while debugging).
• Near the bottom of cepo.py conditional statements check to see if "print_output" is set in the config - these can probably be removed since this is controlled by logging statements and setting the log level
• In cepo.py, on line 325, the code checks to see if the error occurred because of invalid effort levels. This check works but is very brittle. Perhaps you know a better way to make this check.
• In cepo.py, there may be a missing continue on line 327. It works without the continue there, but to match the old logic (even with the fallback in place), the continue should probably be there. A quick test can confirm the behavior.