PoC: PdfCapabilities flags for selective content extraction

[EdwardBlair]

Proof of concept for the capabilities enum discussed in #980. Builds on top of #1281 (lazy stream loading).

Adds a PdfCapabilities flags enum to ParsingOptions so you can tell PdfPig what you actually need:

var options = new ParsingOptions
{
    Capabilities = PdfCapabilities.Text,
    LazyLoading = true
};

Flags are Text, Images, Paths, and All (the default, so nothing changes for existing users). Could add more (Colors, AcroForms, etc) if this direction makes sense.

With LazyLoading = false (the default), the capabilities flags skip processing of unwanted content which saves CPU but the raw stream bytes are still read into memory by the scanner. Combined with LazyLoading = true, skipped streams are never accessed so the bytes stay on disk - thats where the real memory savings come from.

TODO:

• Thread ParsingOptions through to Page so we can throw if someone tries to access data they didn't extract (e.g. calling GetImages() when Images wasn't in capabilities)
• More capability flags (Colors, AcroForms, etc)

Depends on #1281.