Fix fasta file generation and change fasta files to use 2bit as source

[Xewar313]

#19

[iromeo]

I see that fun Iterable<FastaRecord>.write(path: Path, width: Int = 80) was introduced earlier and wont covered by tests. Could you please cover it with some unit tests. In tests hg38, hg19 and other real genomes are not available. Instead, we use a generated Genome["to1"] genome with a couple of chromosomes.

• For this method seems you do not need to1 and it's ok to generate same fake FastaRecords that are configured using string sequence + validate that header and split by desired width works ok

Also, please add a test for Genome.writeAsFasta and verify:

• that the chromosomes order in resulting file is same as in Genome.chromSizesMap
• that chromosomes sequence start with same nucleotieds as in Chromosome.sequence() for correspoding chromosomes
• chomosome lenth in result file matched to Chromosome length in to1 genome.
  Keep in mind, that exact nucleotide sequence in the test chromosome is randomly generated, although seed is fixed and it should be same each time (see org.jetbrains.bio.genome.TestOrganismDataGenerator#generateSequence)

[iromeo]

Also AnnotationsConfig.fastaUrl isn't used any more, so seems we could clean it up

[iromeo]

@Xewar313 GenomeAnnotationEditor compilation is broken due to missing fasta url in epigenome repo, please fix there also

[Xewar313]

@iromeo Sorry, I have forgotten to push jbr changes