oberix / star Goto Github PK

Bag class is just (as the name suggests) a box to keep together a DataFrame and some printing metadata.

Some plotters (bars mainly) doesn't rescale right when a secondary y axes is inserted. The issue is most evident with cumulative bar graphs since the bars overlaps.

ReMida should make use of gettext to produce reports in different languages

By now config.cfg files in report's folders may be empty but they must be there, or else remida.py does not identify the folder to be a report's folder.
when issuing the command

$ remida my_report_folder/

If a config.cfg file is not found, a file named main.tex or main.html should be searched for.

logging system seems to fallback to default level (warning); it was managed by config.py, for some reason it's now broken.

A more solid implementation should be provided

in star.remida.table.HTMLTable.out method escape result of _make_header method.

This result in a wrong thead.

star.remida.utils.escape should be used in _make_header and _make_body method on the contents of html tags and not around all methods returned values.

in star.remida.utils in function escape call re.sub that expect a raw/reges string as r"expression". HTML_ESCAPE is a dictionary with re.compiled (_SRE_Pattern) as keys.

resulting in a non iterable object.

graph.py should handle xticks as once for every plotted series, not one by one as it's doing by now.

The issue could be more complicated than it's seems, since not every plotter need the same transformation to set up axis tiks.

star/share/default_meta.py has two placeholders to define default values for automated text generation. Fill them up

in star/share/default_meta.py there are a lot of validation functions that needs an implementations

Des class (to make automatic text generation) has been written on branch 0.2, master branch has many differences, so it should be make compatible

Graph miss to find scatter plotter

in star.remida.utils variable HTML_ESCAPE is a dict with re.compiled as keys.
in star.remida.utils function escape use re.sub that expect r"expression" string instead _SRE_Pattern (compiled regex).

HTML_ESCAPE result in an not iterable object.

A Stark merge should regenerate expressions of elab columns, for some reason it's not doing so for logit()

when issuing for a calculated columns, also those columns from which it was evaluated must be in the list; this is not the desired behaviour: if a columns is present in the DataFrame, it must be printable in within STAR