Fixes issue 22

[thibautjombart]

This should help address issue https://github.com/whocov/trend_analysis_public/issues/22 adding the following to thedf_dynamics data.frame:

• $tpr: TPR for the last week (average)
• $r_tpr: slope of the regression of TPR over time based on the last three weeks
• $r_tpr: slope of the regression of TPR over time based on the last three weeks
• $lower_r_tpr: lower bound of the conf int for r_tpr
• $upper_r_tpr: upper bound of the conf int for r_tpr
• $change_r_tpr: change to tpr as a factor (increasing, constant or decreasing)

Work left over will be to add this information to the ELR review script.

[henryls1]

Suggesting using the test_positivity_rate field in all_dat$testing_data2 (in assemble_data.Rmd) for outputting tpr - this is to deal with the issue of inconsistent reporting windows for testing data (weekly vs daily for some countries)

[thibautjombart]

Thanks for the reviews. @henryls1 can you check if the current solution is satisfying? Not using all_dat$testing_data2 as it is not currently exported due to the different logics daily /vs/ weekly data.