Add support for non-regex processing

Currently, pywhat only uses regular expressions to identify stuff. However, we may often want to update the description based on a match or to filter out false positives.

Some examples:

- Pinging URLs to check if it is responding or if a certain resource exists, etc.
- Parsing timestamps (https://github.com/bee-san/pyWhat/issues/234) (`1637093119.558717` to `November 16, 2021 8:05:19 PM`)
- Changing rarity based on a match
- And many more

How it should be implemented?

I think the best idea would be to write a class for every regex pattern (so that it can save the state between several matches). That class should have a method like `process()` that gets a match object and can alter it and return it or return None to filter it out.

Example:

```py
class URLProcessor:
def process(match):
if "http://trashurl.it" in match["url"]: # Or maybe we should create an actual Match class and do match.url
return None
```

No files, tests, or entry points are named. Start by tracing how pywhat currently applies regular expressions and represents matches, then compare that flow with the proposed stateful processor and Match interface; done should support URL checks, timestamp parsing, metadata changes, and filtering false positives without breaking existing regex identification.

Written by the indexing model from the issue text.