TimeFrame has fallen out of use

[MichaelTiemannOSC]

The documentation describes short, medium, and long-term timeframes. All the test files use a timeframe of LONG and we've largely defaulted to that being the only thing. We need to prioritize and define whatever needs to be done to make SHORT and MEDIUM meaningful, or change the documentation to leave timeframe out (since target dates provide concrete meaning anyway).

See Time Frames section here: https://github.com/os-climate/ITR/blob/develop/docs/FunctionalOverview.rst

@LeylaJavadova @ImkeHorten

[MichaelTiemannOSC]

Adding @FannySternfeld

I'm going to put the following text into the docs for now.

The tool collects all the interim targets that each company provides
(which almost always include a 2030 date or a "50% reduction by X"
date and/or a 2050 date or a "100% reduction by Y" date). A minority
of companies also provide a short-term term (2-3 years out). The tool
uses all the stated targets for shaping the totality of target
projections from the year after the last reported emissions data
through 2050. Interpolation from reported data through successive
targets is based on a hybrid CAGR/Linear model (explained below).

A company that supplies short, medium, and long-term targets would
have a target reduction curve that interpolates between the three
points. A company that provides only a single long-term target would
have a target reduction curve that intercepts that single target.

Then, instead of selecting SHORT, MEDIUM or LONG, we allow the user to
set the end-date of the analysis (to one of 2025, 2030, 2035, 2040,
2045, or 2050). We then score the temperature based on the
overshoot/undershoot of the benchmark's temperature as of that
date. Thus, if the benchmark is presumed 1.2˚C in the current era and
1.5˚C in 2050 and the end-date selected is 2030, the temperature will
be scored against 1.35˚C (if that's what the carbon budget/ITR of the
benchmark is at 2030).

The target interpolation method is based on a hybrid CAGR/Linear
model. Because no CAGR model can ever reach zero (Xeno's paradox), a
pure CAGR-based model is inappropriate for describing an emissions
reduction curve that reaches zero. To solve this problem, when the
target reduction is > 90% we create a weighed average of a CAGR-based
reduction and a linear interpolation. Thus, if the present analysis
begins 2025 with a 50% reduction target for 2030, calculate the CAGR
for a 50% reduction across the 6 years (2025, 2026, 2027, 2028, 2029,
2030). If that company has a 100% reduction target for 2050, we
create a CAGR curve targeting a 90% reduction over that 20 year period
(2031-2050) and a linear curve targeting a 100% reduction from
2031-2050, and then compute a date-weighted average of the two, with
all the reduction coming from the CAGR curve in 2031 and all the
reduction coming from the linear curve in 2050.