CSVVCurveGen attrs

#seed: 10
#this is the ouput of CSVVCurveGen

constant = {'tasmax': -41.739060383321529,
            'tasmax2': 2.19043060424606,
            'tasmax3': -0.036357886410900681,
            'tasmax4': 1.1182497539239723e-05
            }
predcovars = {'tasmax': ['loggdppc', 'hotdd_agg', 'coldd_agg'],
              'tasmax2': ['loggdppc', 'hotdd_agg', 'coldd_agg'],
              'tasmax3': ['loggdppc', 'hotdd_agg', 'coldd_agg'],
              'tasmax4': ['loggdppc', 'hotdd_agg', 'coldd_agg']
             }
predgammas = {'tasmax': array([ 3.11738478,  0.01543973,  0.01258509]),
              'tasmax2': array([-0.15492075, -0.00040419, -0.00080722]),
              'tasmax3': array([  1.91988729e-03,  -1.57687723e-06,   1.26788237e-05]),
              'tasmax4': array([  1.73683776e-05,  -2.12296873e-06,   1.65078588e-06])
              }

And lets say I have a datastructure from all IR regions that look like the following where every key is an IR with covariates for a given time period

{
IR_0: {'coldd_agg': array(0.10893644730466345),
  'hotdd_agg': array(0.7372790235956397),
  'loggdppc': array(0.01685680452545113)},
  
IR_1: {'coldd_agg': array(0.22338429591820586),
  'hotdd_agg': array(0.8162201398900606),
  'loggdppc': array(0.2871891762242994)},
 
 ...
 
 With the above specified inputs, and using line 46 of CSVVCurveGen we generate our IR-level coefficients 
 
{IR_0: {'tasmax': 0.3432, 'tasmax2': 0.6677, 'tasmax3': 0.234355, 'tasmax4': .1455}, 
 IR_1: {'tasmax': .056289103, 'tasmax2': .879740, 'tasmax3': .1434243, 'tasmax4': .108971},
 ...