SummaryTable#

class ns_sdn.summary.SummaryTable(title, metadata=(), columns=(), rows=(), notes=())#

A rendered-on-demand estimation summary.

Holds the structure of a summary rather than a formatted string: a title, label/value metadata pairs rendered two per line, one table, and closing notes. Renders as fixed-width text (to_text(), which is also str and repr), as HTML in a notebook (_repr_html_), or as a dataframe (to_pandas()), and imports nothing optional until a frame is asked for. Every value is already a string: the producer decides the number formatting, the table decides the layout.

The text layout is the econometric convention: a centred title, a rule, the metadata in two columns of label: value pairs, a rule, the table with its first column left-aligned and the rest right-aligned, a rule, and the notes word-wrapped to the table width (at least 80 characters).

Variables:

  • title – Headline.

  • metadata – Label/value pairs describing the fit, laid out two per line in the order given.

  • columns – Column headers; the first names the row key and may be empty.

  • rows – Table rows; the first entry of each is the row key and the rest align with columns[1:].

  • notes – Closing lines stating what the numbers assume.

Example

>>> from ns_sdn.summary import SummaryTable
>>> table = SummaryTable(
...     title="Demo",
...     metadata=(("Observations", "99"), ("Parameters", "7")),
...     columns=("", "mean", "share"),
...     rows=(("component 1", "0.50", "0.800"), ("component 2", "0.10", "0.200")),
...     notes=("Shares sum to one.",),
... )
>>> print(table)
                                       Demo
================================================================================
Observations:   99  Parameters:    7
--------------------------------------------------------------------------------
               mean   share
component 1    0.50   0.800
component 2    0.10   0.200
================================================================================
Shares sum to one.
>>> table.to_pandas().loc["component 2", "share"]
'0.200'

Attributes#

title

metadata

columns

rows

notes

Methods#

to_pandas()

The table as a pandas.DataFrame, metadata in attrs.

to_text()

Render the summary as fixed-width text.