Faster and memory efficient way for cPeak mapping

[nilsmechtel]

• Use pandas to read fragments file in chunks with progress bar and faster runtime.
• Allow filtering of minimal and maximal fragment length.
• Allow to specify a minimum of basepair overlap between fragment and cPeak region.
• Use binary search with numpy to faster map fragments to cPeak regions.
• Save sparse count matrix as compressed .mtx.gz file.

[XinzeWu]

wow, thank you so much. It is such a great improvement!