Add CharsetDetector

[viktorbeck98]

Detects variables (non-numeric) that contain characters that were not seen in training before as anomalies

[ernstleierzopf]

also add files to
detectors/init.py
docs/detectors/detectorname.md
docs/detectors.md
config/pipeline_config_default.yaml

[ernstleierzopf]

@viktorbeck98 a few questions regarding this detector:

1. How to treat values where multiple new characters are seen: overall_score = 1.0 or overall_score = +1.0 for each new character -> potentially leading to overall_score values of for example 15.0. I assume the former option is the correct one, but please confirm.
2. Storing the full values and then checking if any of the values contains the character is incredibly inefficient. We should store only characters or at least add a known set of characters to the persistence. This requires changes to the persistence. Should they be implemented directly with this detector or should this be a new issue that will be worked on when we have more detectors?
3. Does the current setup store a set of values for each variable in all event types? Does this not lead to memory overflow, if you do this with all of the detectors?

[ernstleierzopf]

addition to 1.: what I meant is overall_score is changed when one of the variables has anomalies. If there are two variables with anomalies, then the score is 2.0, etc. This is what I implement currently.

[ernstleierzopf]

Done in #148