Improve the robustness of police.hu crawler

[snorbi07]

This is a larger issue and serves as an aggregation of know issues or missing features. During implementation, we need to split this up into multiple separate issues or PRs.

Primarily there are 3 missing features:

1. currently if we have an unsupported format in the parsing phase, it will lead to either an empty string or a crash. This needs to be more consistent and reliable. In case of an unsupported format, parts should fall back to a default value. For unsupported corner cases, the event should be logged with a warning level containing the string that led to a fallback value so we can provide support for it in the future.
2. the fetched and parsed HTML should be stored in a raw format as well, so we don't loose data if the parsing fails for whatever reason. The storage should happen in such a way that we can easily pair the parsed and raw HTML formats together.
3. the crawler messes up and losses the timezone information when storing the parsed data. It only supports UTC, without taking into consideration the timezone info of HU. The stored and parsed data should clearly reflect when it was captured (CEST, CET).

Definition of done:

• extend the crawler to store the fetched HTML page and not just the result of the parsing to avoid data loss.
• extend the various parsing steps to handle malformed inputs, by returning a default value and logging the input that it failed to handle.
• cover the malformed input parsing cases with unit tests.