Add new handled content_types for processable_file validator (CSV, XML, JSON, txt, md, epub, open office, zip, ...) (POTENTIALLY BREAKING)

[Mth0158]

Expected behaviour

As we make the change mentioned in #404, we would like to add new supported content_types for deep analysis. Right now we handle:

• image (vips/minimagick)
• video (ffmpeg)
• audio (ffmpeg)
• pdf (poppler)

The gem code is clearly designed to handle new content_types, we could in theory add without too much work the following content_types:

• csv (ruby csv with binary detection, encoding check, parsing 1 or 2 lines) => text/csv + application/csv + application/vnd.ms-excel + text/tab-separated-values
• json (ruby json)
• ndjson (line by line ruby json, with a sanity cap)
• xml (nokogiri XML)
• yaml (ruby yaml with safe_load)
• txt (internal check with binary detection)
• xlsx (ruby zip (>= 3.1) should reject zip bombs + nokogiri XML)
• docx (ruby zip (>= 3.1) should reject zip bombs + nokogiri XML)
• zip (ruby zip (>= 3.1) should reject zip bombs)
• gzip (ruby zlib should reject zip bombs)
• epub (ruby zip (>= 3.1) should reject zip bombs)
• html, xhtml (nokogiri HTML)
• svg (nokogiri XML with extra checks)
• markdown (internal check with binary detection)

To not forget

• Update readme accordingly