Feature request: Support headerless format in decompress and get_frame_parameters

[swehner]

We're using zstandard wihtout the magic header, but for decompression there is no support in decompress method. This requires to prepend the magic header before processing it. It's a minor issue but it's not very efficient having to concatenate the buffers unnecessarily.

Opened a pull request here: https://github.com/indygreg/python-zstandard/pull/217 with the following changes:

• Add optional paramater format to get_frame_paramaters and switch to using getFrameHeader_advanced internally
• Use getFrameHeader_advanced in decompress instead of getFrameContentSize