streamline and simplify logic tree in metagenomics.krona regarding the handling of --inputType. In particular, the previous inclusion of logic for handling krakenuniq and kaiju summary reports failed to pass the root_name variable like the previous tsv-based input, making it harder to tell which krona plot corresponded to which sample.
pass the filename base name to krona as a "dataset name". This embeds the sample's name in the krona html which gets picked up in later merge operations to distinguish which plots correspond to which sample in a multi-sample html file. This is required for PR40 in viral-pipelines.
fix a regex string
rename all the test/unit/integration.py test files to test_integration.py -- pytest was previously ignoring these entirely. Re-enables the integration tests in Travis. Comment out all the ones that were broken. Specifically all of the kraken (v1 non-uniq) ones that never worked (because kraken and krakenuniq conda packages cannot coexist on the same docker image). Comment out some other bits of tests that apparently never worked (e.g. the ones that parse krona html output).
root_namevariable like the previous tsv-based input, making it harder to tell which krona plot corresponded to which sample.