The genome is available from e.g. UCSC and Ensembl and only a google search away plus there are plenty of questions here already on that matter, please use the search function. For the binning, take a look at bedtools makewindows. Not python but hey, why reinventing the wheel if a superfast option already exists.

You can use Kent tools and BEDOPS bedops --chop with standard Unix streams to do this quickly and efficiently:

$ fetchChromSizes hg19 | awk '{print $1"\t0\t"$2}' | sort-bed - | bedops --chop 50000 - > hg19.50k.bed

Note that the bounds of chromosomes will almost certainly not split evenly by 50kb units. In other words, chopping up a chromosome will leave some chunk at the end that is less than the desired length.

If you're doing calculations that involve aggregation, you may want to leave out that last < 50kb bin, by adding the -x operator to bedops --chop:

$ fetchChromSizes hg19 | awk '{print $1"\t0\t"$2}' | sort-bed - | bedops --chop 50000 -x - > hg19.50k.chunkless.bed

As a bonus, the output is sorted and ready to use for fast set operations with BEDOPS.

for hg19 chromosome sizes:

https://genome.ucsc.edu/goldenPath/help/hg19.chrom.sizes

for making windows of 10kb:

$ wget -np -r -nd https://genome.ucsc.edu/goldenPath/help/hg19.chrom.sizes
$ bedtools makewindows -g hg19.chrom.sizes -w 10000

try using pybedtools. https://daler.github.io/pybedtools/