add option to not initiate the calibration fit with a simple fit of t…

…heta circles

bin/fit_posparams

@@ -53,6 +53,7 @@ parser.add_argument('-v', '--verbose', action='store_true',
                     help='extra verbosity of print outs at terminal')
 parser.add_argument('--log-note-selection', type=str, default=None, required=False,
                     help="required keywords in LOG_NOTE selection, seperated by '&', like ''arc calibration & use_desimeter=True'")
+parser.add_argument('--no-circle-fit', action='store_true', help="do not initiate the fit with a fit of theta circles")
 
 args = parser.parse_args()
 
@@ -108,7 +109,8 @@ if __name__ == '__main__':
                       'savedir': args.outdir,
                       'static_quantile': args.static_quantile,
                       'printf': printf,
-                      'log_note_selection': args.log_note_selection
+                      'log_note_selection': args.log_note_selection,
+                      'no_circle_fit': args.no_circle_fit
                       }
             if args.n_processes_max == 1:
                 logstr = fithandler.run_best_fits(**kwargs)

py/desimeter/posparams/fithandler.py

@@ -31,7 +31,7 @@
 output_keys.update({key:True for key in fitter.all_keys})
 
 def run_best_fits(posid, path, period_days, data_window, savedir,
-                  static_quantile, printf=print, min_window=10, log_note_selection=None):
+                  static_quantile, printf=print, min_window=10, log_note_selection=None, no_circle_fit = False):
     '''Define best-fit analysis cases for one positioner.
 
     INPUTS:
@@ -44,6 +44,7 @@ def run_best_fits(posid, path, period_days, data_window, savedir,
         printf ... print function (so you can control how this module spits any messages)
         min_window ... (optional) minimum number of data points to attempt a fit
         log_note_selection ... (optional) required keywords in LOG_NOTE selection, seperated by '&', like ''arc calibration & use_desimeter=True'
+        no_circle_fit ... (bool, optional) if true, do not initiate the fit with a fit of theta circles
 
     OUTPUTS:
         logstr ... string describing (very briefly) what was done. suitable for
@@ -65,7 +66,7 @@ def run_best_fits(posid, path, period_days, data_window, savedir,
     # FIRST-PASS:  STATIC PARAMETERS
     static_out = _process_cases(table, cases, printf=printf, mode='static',
                                 param_nominals=fitter.default_values.copy(),
-                                )
+                                no_circle_fit=no_circle_fit)
 
     # DECIDE ON BEST STATIC PARAMS
     best_static = fitter.default_values.copy()
@@ -247,7 +248,7 @@ def _define_cases(table, datum_dates, data_window, printf=print):
     printf(f'{posid}: {len(cases):5d} analysis cases defined')
     return cases
 
-def _process_cases(table, cases, mode, param_nominals, printf=print):
+def _process_cases(table, cases, mode, param_nominals, printf=print, no_circle_fit = False):
     '''Feed analysis cases for a positioner through the best-fit function.
 
     INPUTS:
@@ -256,6 +257,7 @@ def _process_cases(table, cases, mode, param_nominals, printf=print):
         mode ... 'static' or 'dynamic'
         param_nominals ... Dict of nominal parameter values, structured like fitter.default_values
         printf ... print function
+        no_circle_fit ... (bool, optional) if true, do not initiate the fit with a fit of theta circles
 
     OUTPUT:
         output ... Astropy table of results, headers given by output_keys.
@@ -281,15 +283,16 @@ def _process_cases(table, cases, mode, param_nominals, printf=print):
         printf(f'  final idx = {n:5d}, date = {subtable["DATE"][-1]}')
         printf(f'  num points = {n-m+1:5d}')
         params, covariance_dict, rms_of_residuals = fitter.fit_params(posintT=xytp_data['posintT'],
-                                                                 posintP=xytp_data['posintP'],
-                                                                 ptlX=xytp_data['ptlX'],
-                                                                 ptlY=xytp_data['ptlY'],
-                                                                 gearT=xytp_data['gearT'],
-                                                                 gearP=xytp_data['gearP'],
-                                                                 mode=mode,
-                                                                 nominals=param_nominals,
-                                                                 bounds=fitter.default_bounds,
-                                                                 keep_fixed=[])
+                                                                      posintP=xytp_data['posintP'],
+                                                                      ptlX=xytp_data['ptlX'],
+                                                                      ptlY=xytp_data['ptlY'],
+                                                                      gearT=xytp_data['gearT'],
+                                                                      gearP=xytp_data['gearP'],
+                                                                      mode=mode,
+                                                                      nominals=param_nominals,
+                                                                      bounds=fitter.default_bounds,
+                                                                      keep_fixed=[],
+                                                                      no_circle_fit=no_circle_fit)
         output['ANALYSIS_DATE'].append(Time.now().iso)
         output['POS_ID'].append(posid)
         for suffix in {'', '_SEC'}:

py/desimeter/posparams/fitter.py

@@ -65,6 +65,7 @@ def fit_params(posintT, posintP, ptlX, ptlY, gearT, gearP,
                keep_fixed=None,
                ptlXerr=None,
                ptlYerr=None,
+               no_circle_fit=False
                ):
     '''Best-fit function for parameters used in the transformation between
     internally-tracked (theta,phi) and externally measured (x,y).
@@ -238,7 +239,7 @@ def compute_chi2(params):
         x_exp, y_exp = expected_xy(params)
         return np.sum(((x_exp - x_flat)/xerr_flat)**2 + ((y_exp - y_flat)/yerr_flat)**2)
 
-    if mode=='static' :
+    if (not no_circle_fit) and (mode=='static') :
         # we initialize OFFSET_X and OFFSET_Y by fitting circles
         # for all unique values of p_int
         unique_p_int = np.unique(p_int)