Transfer-Encoding and Content-Encoding handling

[nvanva]

warc2text does not seem to handle Transfer-Encoding and Content-Encoding

[nvanva]

For ArchiveBot:

1. 70% of docs, 80% of chars come from chunked HTTP responses, HTMLs will contain traces of chunking (chunk size tokens - hexadecimal numbers, in the beginning and inside HTML code)
2. After html2text+Monotexting many of them survive (eyeballing the HPLT2.0 samples ArchiveBot: ~10-15% of docs start with chunk size tokens, more docs may have them in the middle)
3. Gzip-ed reponses won’t give much more text (upper bound: +3% more chars)

[nvanva]

@ZJaume Jaume, could you please take a look at this pull request:
#76