Move read_table() and Counter() to a top-level utils directory as they are used both by pcigale and pcigale-plots.

Some lines have been made slightly thicker in SED plots so the different components are more visible

The lines linking the different bands in the residual SED plot have been eliminated to improve the readability

The uncertainties on the SED plots now correspond only to 1σ rather than 3σ so they do not appear exceedingly large

Change the location of the logo so that it does not overlap with anything and replace it with a less pixelated version

Adding an option to change the output file format. The available formats depends of the host, but usually png, pdf, ps, eps and svg is supported.

Example:
--format svg

Use the --series <series1 series2...> option to indicate one o more of this series:
stellar_attenuated,
stellar_unattenuated,
nebular,
dust,
agn,
radio,
model.

If the --seriesdisabled flag is used, then all available series will be enabled, except the ones that are indicated in the --series option. If not, then only the series indicated in the --series option will be enabled.

SED: Adding --xrange and --yrange to set the plot axis ranges (optionally). The format are [<min>]:[<max>], where min and max are Python floats (e.g. 1E4).

Examples:
--xrange :1E3  # automatically X min
--yrange 1E2:1E4

* Include dependence of pcigale.analysis_modules.utils.Counter class, to perform the progress message.
* Fixed error in the way to pass "outdir" parameter to the pool in chi2, and pdf.

Now it is possible to choose the folder that contains the output files (the "out" folder). It is expected that the default configuration file pcigale.ini be sibling of the "out" folder.

Minor optimization in logo reading in workers.
Fix logo position.
Adding logging.

Fν was computed by calculating Fλ and then converting to Fν, which led to typical differences in fluxes of typically 1-2% and a bit more for a handful of pathological filters. Now Fν is computed directly and a bit faster. At the same time rather than computing an effective wavelength (which was more a mean wavelength), we compute the pivot wavelength, which is more appropriate when converting between Fν and Fλ.

For more reliability we now save the chi² by blocks. Otherwise it could lead to some crashes when there was more than 1 redshift.