Pyani

- fix for issue 53 (--maxmatch has no effect)
- fix to `genbank_get_genomes_by_taxon.py` to account for NCBI FTP location changes
- fixed issue 52 (local variable bug)
- fixed issued 49 (TETRA failure) and 51 (matplotlib bug)
- add several tests and support for `codecov.io`, `landscape.io` and `Travis-CI`
- removed requirement for `rpy2`
- moved scripts to `bin/` subdirectory

- `pyani` now requires `rpy2` v2.8.0 in order to satisfy running under Anaconda (see issue 26)
- `pyani` now checks for presence of `rpy2` and - when run from source - if `rpy2` is not available, `pyani` doesn't throw an error until R graphical output is requested. If installed -via- `pip`, then `pyani` still raises `pkg_resources.DistributionNotFound` if `rpy2` is missing.
- Updated `genbank_get_genomes_by_taxon.py` script to use the new FTP locations at NCBI for each assembly.
- Fixed bug where `ANIb` would not go to completion if empty BLASTN files were generated (see issue 27)
- Fixed bug where `ANIm` would not finish under `multiprocessing` if input sequences were highly divergent.
- Added Hadamard product of percentage identity and alignment coverage as output.
- Fixed bug where label/classes are out of sync with new NCBI downloaded filenames
- Added --rerender option to draw (new) graphics from old output, without recalculation
- Corrected matplotlib row dendrogram orientation
- Seaborn output no longer dumps core on large (ca. 500 genome) datasets
- `genbank_get_genomes_by_taxon.py` attempts to identify cause for failed downloads and correct, where nomenclature/versions are at fault
- graceful replacement of classes that are not present in `classes.txt`
- add `pyani` version to log file

- Merged [pull request](https://github.com/widdowquinn/pyani/pull/17) from peterjc to make printing from tests Python3-friendly.
- Merged [pull request](https://github.com/widdowquinn/pyani/pull/21) from peterjc to use `open()` for opening files.
- Merged [pull request](https://github.com/widdowquinn/pyani/pull/24) from peterjc to cope with missing labels/classes more gracefully
- Fixed `-s`/`--fragsize` option in `average_nucleotide_identity.py` (thanks to Joseph Adelskov for hte report).
- BLAST and `nucmer` results are now written to a subdirectory of the output folder. By default, these sequence search output files are compressed, but this behaviour can be suppressed using the `--nocompress` option.
- Added `genbank_get_genomes_by_taxon.py` as an aid to downloading publicly-available genome files from GenBank, for analysis.