Skip to content

Commit

Permalink
preview curve fit changes
Browse files Browse the repository at this point in the history
  • Loading branch information
@roasterdave
roasterdave committed Oct 28, 2019
1 parent 4966dd8 commit 0a9f5e1
Showing 1 changed file with 104 additions and 51 deletions.
155 changes: 104 additions & 51 deletions src/artisanlib/main.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -2088,6 +2088,8 @@ def __init__(self,parent,dpi):
self.analysisresultsloc = [.5,.5]
self.analysispickflag = False
self.analysisresultsstr = ""
self.analysis_iteration = 0 #dave
self.analysis_iterate = True #dave


#NOTE: empty Figure is initialy drawn at the end of aw.settingsload()
Expand Down Expand Up @@ -9601,53 +9603,43 @@ def polyfit(self,xarray,yarray,deg,startindex,endindex,_=False):
return None

#ln() regression. ln() will be used when power does not equal 2 (quadratic) or 3 (cubic).
def lnRegression(self,power=0, timeoffset=0):
def lnRegression(self,power=0, timeoffset=0, plot=True):
res = ""
try:
if self.timeindex[0] > -1: # only if CHARGE is set
from scipy.optimize import curve_fit
from scipy.optimize import curve_fit
if self.timeindex[0] > -1 and self.timeindex[6] > -1: #CHARGE and DROP events exist
charge = self.timex[self.timeindex[0]]
if timeoffset != None and timeoffset > 0:
timeidx = aw.time2index(timeoffset)
a = [self.timex[timeidx]-charge]
n = [self.temp2[timeidx]]
elif aw.qmc.ambientTemp != None and aw.qmc.ambientTemp > 0:
timeidx = 0
a = [0]
n = [aw.qmc.ambientTemp]
else:
timeidx = 0
a = [0]
if aw.qmc.mode == "F":
roomTemp = 70.0
if timeoffset != None and timeoffset > charge:
begin = aw.time2index(timeoffset)
time_l = []
temp_l = []
else:
a = [charge]
# find the DRY END point
if self.timeindex[1]: # take DRY if available
begin = self.timeindex[1]
else: # take DRY as specificed in phases
pi = aw.findDryEnd(phasesindex=1)
begin = aw.time2index(self.timex[pi])
# intial bean temp set to ambient
if aw.qmc.ambientTemp != None and aw.qmc.ambientTemp > 0:
time_l = [charge]
temp_l = [aw.qmc.ambientTemp]
else:
roomTemp = 21.0
n = [roomTemp]
if self.timeindex[1]: # take DRY if available
x = self.timex[self.timeindex[1]] - charge
a = a + [x]
n = n + [self.temp2[self.timeindex[1]]]
else: # take DRY as specificed in phases
i = aw.findDryEnd(phasesindex=1)
x = self.timex[i] - charge
a = a + [x]
n = n + [self.temp2[i]]
if self.timeindex[2]: # take FCs if available
x = self.timex[self.timeindex[2]] - charge
a = a + [x]
n = n + [self.temp2[self.timeindex[2]]]
else: # take FCs as defined in phases
i = aw.findDryEnd(phasesindex=2)
x = self.timex[i] - charge
a = a + [x]
n = n + [self.temp2[i]]
if self.timeindex[6]:
x = self.timex[self.timeindex[6]] - charge
a = a + [x]
n = n + [self.temp2[self.timeindex[6]]]
time_l = [charge]
if aw.qmc.mode == "F":
roomTemp = 70.0
else:
roomTemp = 21.0
temp_l = [roomTemp]
end = self.timeindex[6]
time_l = time_l + self.timex[begin:end]
temp_l = temp_l + self.temp2[begin:end]

xa = numpy.array(a)
yn = numpy.array(n)
xa = numpy.array(time_l) - charge
yn = numpy.array(temp_l)
# print(xa) #dave
# print(yn) #dave
if power == 2:
func = lambda x,a,b,c: a*x*x + b*x + c
elif power == 3:
Expand All @@ -9658,12 +9650,13 @@ def lnRegression(self,power=0, timeoffset=0):
warnings.simplefilter("ignore")
popt,_ = curve_fit(func, xa, yn)
#perr = numpy.sqrt(numpy.diag(pcov))
xb = numpy.array(self.timex)
xxb = xb + charge
xxa = xa + charge
self.ax.plot(xxb, func(xb, *popt), color="black", linestyle = '-.', linewidth=3)
self.ax.plot(xxa, yn, "ro")
self.fig.canvas.draw()
if plot:
xb = numpy.array(self.timex)
xxb = xb + charge
xxa = xa + charge
self.ax.plot(xxb, func(xb, *popt), color="black", linestyle = '-.', linewidth=3)
self.ax.plot(xxa, yn, "ro")
self.fig.canvas.draw()
if len(popt)>2:
if power == 2:
res = "%.8f * t*t %s %.8f * t %s %.8f" % (popt[0],("+" if popt[1] > 0 else ""),popt[1],("+" if popt[2] > 0 else ""),popt[2])
Expand Down Expand Up @@ -16249,6 +16242,20 @@ def curveSimilarity2(self,analysis_starttime=0,analysis_endtime=0):
diff_array = np_dbt - np_dbtb
zerocrosses = ((diff_array[:-1] * diff_array[1:]) < 0).sum()

crossings = numpy.sign(diff_array)
starts = numpy.r_[0, numpy.flatnonzero(~numpy.isclose(crossings[1:], crossings[:-1])) + 1]
lengths = numpy.diff(numpy.r_[starts, len(crossings)])
signs = crossings[starts]

# print(starts,lengths,signs)
idelta = (np_dbt - np_dbtb) * crossings
# print(idelta)
ideltamax = []
for i in range(len(starts)):
ideltamax.append(numpy.amax(idelta[starts[i]:starts[i]+lengths[i]]))
# print(ideltamax)


else:
RoR_FCs_act = 0
RoR_FCs_templ = 0
Expand Down Expand Up @@ -29386,8 +29393,8 @@ def analysisfitCurves(self, exp=-1):
i = self.findDryEnd(phasesindex=2)
fcstime = self.qmc.timex[i]

# set the interval of interest from analysis_starttime to analysis_endtime
analysis_starttime = fcstime - 120
#analysis_endtime = fcstime + 30
analysis_endtime = aw.qmc.timex[aw.qmc.timeindex[6]]

#natural log needs a curve fit point sometime earlier than drytime. Pick one after TP if it exists. Otherwise after DROP.
Expand All @@ -29414,7 +29421,7 @@ def analysisfitCurves(self, exp=-1):
progress.setValue(1)
# cubic or all
if exp == 3 or exp == -1:
res = self.analysisGetResults(exp=3,timeoffset=drytime, analysis_starttime=analysis_starttime, analysis_endtime=analysis_endtime)
res = self.analysisGetResults(exp=3,timeoffset=analysis_starttime, analysis_starttime=analysis_starttime, analysis_endtime=analysis_endtime)
self.cfr["equ_cubic"] = res["equ"]
self.cfr["dbt_cubic"] = res["rmse_BT"]
self.cfr["dbdbt_cubic"] = res["rmse_deltaBT"]
Expand All @@ -29424,9 +29431,33 @@ def analysisfitCurves(self, exp=-1):
self.cfr['ror_min_delta_cubic'] = res['ror_min_delta']
self.cfr['ror_maxmin_delta_cubic'] = ("%4.1f%s%4.1f") % (res['ror_max_delta'], "/", res['ror_min_delta'])
progress.setValue(2)
# # cubic or all
if exp == 3 or exp == -1:
res = self.analysisGetResults(exp=3,timeoffset=drytime, analysis_starttime=analysis_starttime, analysis_endtime=analysis_endtime)
self.cfr["equ_cubicDRY"] = res["equ"]
self.cfr["dbt_cubicDRY"] = res["rmse_BT"]
self.cfr["dbdbt_cubicDRY"] = res["rmse_deltaBT"]
self.cfr["r2_deltabt_cubicDRY"] = res["r2_deltaBT"]
self.cfr['ror_fcs_delta_cubicDRY'] = res['ror_fcs_delta']
self.cfr['ror_max_delta_cubicDRY'] = res['ror_max_delta']
self.cfr['ror_min_delta_cubicDRY'] = res['ror_min_delta']
self.cfr['ror_maxmin_delta_cubicDRY'] = ("%4.1f%s%4.1f") % (res['ror_max_delta'], "/", res['ror_min_delta'])
progress.setValue(2)
# quadratic or all
if exp == 2 or exp == -1:
res = self.analysisGetResults(exp=2,timeoffset=drytime, analysis_starttime=analysis_starttime, analysis_endtime=analysis_endtime)
self.cfr["equ_quadraticDRY"] = res["equ"]
self.cfr["dbt_quadraticDRY"] = res["rmse_BT"]
self.cfr["dbdbt_quadraticDRY"] = res["rmse_deltaBT"]
self.cfr["r2_deltabt_quadraticDRY"] = res["r2_deltaBT"]
self.cfr['ror_fcs_delta_quadraticDRY'] = res['ror_fcs_delta']
self.cfr['ror_max_delta_quadraticDRY'] = res['ror_max_delta']
self.cfr['ror_min_delta_quadraticDRY'] = res['ror_min_delta']
self.cfr['ror_maxmin_delta_quadraticDRY'] = ("%4.1f%s%4.1f") % (res['ror_max_delta'], "/", res['ror_min_delta'])
progress.setValue(3)
# quadratic or all
if exp == 2 or exp == -1:
res = self.analysisGetResults(exp=2,timeoffset=analysis_starttime, analysis_starttime=analysis_starttime, analysis_endtime=analysis_endtime)
self.cfr["equ_quadratic"] = res["equ"]
self.cfr["dbt_quadratic"] = res["rmse_BT"]
self.cfr["dbdbt_quadratic"] = res["rmse_deltaBT"]
Expand All @@ -29436,6 +29467,24 @@ def analysisfitCurves(self, exp=-1):
self.cfr['ror_min_delta_quadratic'] = res['ror_min_delta']
self.cfr['ror_maxmin_delta_quadratic'] = ("%4.1f%s%4.1f") % (res['ror_max_delta'], "/", res['ror_min_delta'])
progress.setValue(3)

#dave analysis_iteration
# if aw.qmc.analysis_iteration == 0 and aw.qmc.analysis_iterate and exp == -1:
# res = self.analysisGetResults(exp=2,timeoffset=analysis_starttime, analysis_starttime=analysis_starttime, analysis_endtime=analysis_endtime)
# aw.sendmessage("This curve fit: x^2 FCs-120 to DROP")
# aw.qmc.analysis_iteration += 1
# elif aw.qmc.analysis_iteration == 1 and aw.qmc.analysis_iterate and exp == -1:
# res = self.analysisGetResults(exp=2,timeoffset=drytime, analysis_starttime=analysis_starttime, analysis_endtime=analysis_endtime)
# aw.sendmessage("This curve fit: x^2 DRY to DROP")
# aw.qmc.analysis_iteration += 1
# elif aw.qmc.analysis_iteration == 2 and aw.qmc.analysis_iterate and exp == -1:
# res = self.analysisGetResults(exp=3,timeoffset=analysis_starttime, analysis_starttime=analysis_starttime, analysis_endtime=analysis_endtime)
# aw.sendmessage("This curve fit: x^3 FCs-120 to DROP")
# aw.qmc.analysis_iteration += 1
# elif aw.qmc.analysis_iteration == 3 and aw.qmc.analysis_iterate and exp == -1:
# res = self.analysisGetResults(exp=3,timeoffset=drytime, analysis_starttime=analysis_starttime, analysis_endtime=analysis_endtime)
# aw.sendmessage("This curve fit: x^3 DRY to DROP")
# aw.qmc.analysis_iteration = -1

# find the curve with the best fit
try:
Expand All @@ -29449,8 +29498,12 @@ def analysisfitCurves(self, exp=-1):
x.float_format = "5.2"
if "equ_quadratic" in self.cfr:
x.add_row(["*" if self.cfr["dbdbt_quadratic"] == bestfit else "", "x\u00b2", self.cfr["dbt_quadratic"], self.cfr["dbdbt_quadratic"], self.cfr["r2_deltabt_quadratic"], self.cfr['ror_fcs_delta_quadratic'], self.cfr['ror_maxmin_delta_quadratic']])
if "equ_quadraticDRY" in self.cfr:
x.add_row(["*" if self.cfr["dbdbt_quadraticDRY"] == bestfit else "", "x\u00b2DRY", self.cfr["dbt_quadraticDRY"], self.cfr["dbdbt_quadraticDRY"], self.cfr["r2_deltabt_quadraticDRY"], self.cfr['ror_fcs_delta_quadraticDRY'], self.cfr['ror_maxmin_delta_quadraticDRY']])
if "equ_cubic" in self.cfr:
x.add_row(["*" if self.cfr["dbdbt_cubic"] == bestfit else "", "x\u00b3", self.cfr["dbt_cubic"], self.cfr["dbdbt_cubic"], self.cfr["r2_deltabt_cubic"], self.cfr['ror_fcs_delta_cubic'], self.cfr['ror_maxmin_delta_cubic']])
if "equ_cubicDRY" in self.cfr:
x.add_row(["*" if self.cfr["dbdbt_cubicDRY"] == bestfit else "", "x\u00b3DRY", self.cfr["dbt_cubicDRY"], self.cfr["dbdbt_cubicDRY"], self.cfr["r2_deltabt_cubicDRY"], self.cfr['ror_fcs_delta_cubicDRY'], self.cfr['ror_maxmin_delta_cubicDRY']])
if "equ_naturallog" in self.cfr:
x.add_row(["*" if self.cfr["dbdbt_naturallog"] == bestfit else "", "ln()", self.cfr["dbt_naturallog"], self.cfr["dbdbt_naturallog"], self.cfr["r2_deltabt_naturallog"], self.cfr['ror_fcs_delta_naturallog'], self.cfr['ror_maxmin_delta_naturallog']])

Expand Down Expand Up @@ -29528,7 +29581,7 @@ def analysisGetResults(self,exp=2,timeoffset=0, analysis_starttime=0, analysis_e
restoreF = False

res = {} #use dict to allow more flexible expansion in the future
res['equ'] = self.qmc.lnRegression(power=exp, timeoffset=timeoffset)
res['equ'] = self.qmc.lnRegression(power=exp, timeoffset=timeoffset, plot=False)
self.deleteBackground()
self.setbackgroundequ(EQU=["",res['equ']], silent=True)
#QApplication.processEvents() #occasionally the fit curve remains showing.
Expand Down

0 comments on commit 0a9f5e1

Please sign in to comment.