Bycon

* new `geneSymbols` parameter for plot labeling allows to add gene labels to
the position of a given gene on a plot
- [http://progenetix.org/services/intervalFrequencies/?geneSymbols=MYC,ERBB2,CDK1&filters=pgx:icdom-85003&output=histoplot](/services/intervalFrequencies/?geneSymbols=MYC,ERBB2,CDK1&filters=pgx:icdom-85003&output=histoplot)
- TODO: still needs a "no-overlap" shifting method for closely spaced labels (as in the Perl version...)

* refinement of `histoplot` options; now as the standard for standard Beacon
results by using the `&output=histoplot` pragma
- [/beacon/biosamples/?datasetIds=examplez,progenetix,cellz&referenceName=9&variantType=DEL&start=21500000&start=21975098&end=21967753&end=22500000&filters=NCIT:C3058&output=histoplot](http://progenetix.org/beacon/biosamples/?datasetIds=examplez,progenetix,cellz&referenceName=9&variantType=DEL&start=21500000&start=21975098&end=21967753&end=22500000&filters=NCIT:C3058&output=histoplot)
* change of `histoplot` handover (`id: 'pgx:handover:histoplot'`) to use
the plotting option instead of a redirect to the Perl `PGX` version

* basic implementation of plot labels for the collation frequency plots, using
the `plot_region_labels` parameter
* switch of the cancercelllines.org UI to use this plotter instead of the Perl
based PGX one (only for the pre-computed collations)

* added `plot_utils.py` and `plot_defaults.yaml`; now there is a first method
for plotting histograms
- so far limited to CNV histograms of many or few whole chromosomes from
pre-computed frequencymaps
- no separate sevice so far; can be invoked from the `intervalFrequencies`
service with added `&output=histoplots` pragma
- so far no marker addition etc.
- <http://progenetix.org/services/intervalFrequencies/?filters=pgx:icdom-85003,pgx:icdom-87003,pgx:icdom-81403&plotChros=7,8,9,13,17&plot_title=CNV+Comparison&output=histoplot&size_plotimage_w_px=800&plot_chro_height=18>

* fixed broken dataset selection (bug introduced w/ v1.0.28)
* moved dataset parsing to separate library `lib/dataset_parsing.py`

* fixed the filter processing where "correctly looking but not existing" filter
patterns were pruned from the query instead of being kept & leading to a mismatch
* added 2 types of warnings for such cases:
- undefined filter pattern
- correct pattern but value not in database