compatability update

this patch brings monte-carlo-vt up to date with more modern versions of gw-distributions and gw-detectors. Specifically

• gw-detectors : e72dddcd96b7884425c3c9f8be686452ed01ef46
• gw-distributions : fd9753c2004c2ea863016c4625b6a1869c16ae71

changes include

• significant refactoring of compute-horizon to save on waveform calls and use modern parsing/transform logic. Also updated output format to HDF
• update concatenate-sample to support both scalar and vector attributes, as well as to allow for concatenation of sequential files in addition to exactly-overlapping files
• moved workflow in condor-sample to a module and added the option to include a summarize-sample job within the DAG. This also included updating what is written to SUB files to comply with new condor rules (must specify request_disk, for example)
• updated mcvt-sample to make use of modern parsing, etc. Also moved the regularly-spaced sampling logic into EventGenerator so this is now a simple call to a classmethod.
• updates summarize sample to use modern file formats from gw-distributions. Moved much of the functionality into a module.
  • NOTE: there are still issues with the estimate of the rate of events, primarily associated with importance sampling the detected events and removing the time-distribution from logprob. However, these are not thought to be mission-critical at the moment, so I'm merging without fixing them (they currently raise an exception)
  • we may also want to support the ability to reweigh detected distribution to look like different populations. At the moment, it only plots the drawn distribution conditioned on user-specified selection criteria.
• renamed all executables to shorten their names
• created a few modules (condor.py, summary.py)
• moved example INI file into etc rather than installing it from within the module itself.
• changed some of the requirements and package_data declared in setup.py