srb update, presentation prep

Author: alTeska

build.sh

@@ -5,5 +5,6 @@ python scripts/relationalDatabaseEvent.py
 
 python scripts/filteredDatabase.py
 python scripts/velocityDatabase.py
+python scripts/SRBautoDatabase.py
 
 # jupyter nbconvert --to html reports/*.ipynb --output-dir reports/html

notebooks/Presentation_plots_1005.ipynb

@@ -0,0 +1,90 @@
+import cfg
+import h5py
+import numpy as np
+import pandas as pd
+from tqdm import tqdm
+
+pd.options.mode.chained_assignment = None  # to be fixed
+
+
+rat_auto = pd.DataFrame()
+rat  = pd.read_hdf(cfg.velocity_fname, 'Rat_Behavior').set_index('index')
+stim = pd.read_hdf(cfg.relational_fname, 'Events').set_index('index')
+stim.drop(labels=['MotiveExpTimeSecs'], axis=1, inplace=True)
+
+rat_s = pd.merge(rat, stim, on=['Frame', 'Time', 'session_id'])
+rat_sf = rat_s[rat_s.Filtered == 1]
+
+limits = pd.DataFrame(cfg.LIMITS)
+
+
+
+# create limit columns
+rat_sf['Dmin'], rat_sf['Dmax'] = 0, 0
+rat_sf['change'] = False
+for index, row in tqdm(limits.iterrows()):
+    rat_sf['change'] = (rat_sf['Dmin']==0) & (rat_sf['speed']==row['speed'])
+    rat_sf['Dmin'] = np.where(rat_sf['change']==True, row['min'], rat_sf['Dmin'])
+    rat_sf['Dmax'] = np.where(rat_sf['change']==True, row['max'], rat_sf['Dmax'])
+
+rat_sf['SRB'] = (rat_sf['U'] > rat_sf['Dmin']) & (rat_sf['U'] < rat_sf['Dmax'])
+
+# dropping unused columns
+rat_sf.drop(['Filtered', 'change'], axis=1, inplace=True)
+
+
+# MERGING SRB THAT HAVE REALLY SHORT BREAKS
+rat_up2 = pd.DataFrame()
+for name, dd in rat_sf.groupby('session_id'):
+    # calculating dtime - time difference between srb
+    dfSRBtemp = dd[dd['SRB']==True]
+    dfSRBtemp['dtime'] = dfSRBtemp['Time'].diff(1)
+    df2 = pd.merge(dd, dfSRBtemp[['dtime', 'Frame', 'session_id','Time']], on=['Frame', 'session_id','Time'], how='outer')
+    df2.fillna(0, inplace=True)
+
+    #special time cases
+    dftemp = dfSRBtemp[(dfSRBtemp['dtime']>0.005) & (dfSRBtemp['dtime']<0.1)]
+    dftemp = dftemp.reset_index(drop=True)
+
+    maxF, minF = {}, {}
+    minF={}
+    df2['SRBall'] = df2['SRB']
+    df2['SRBtt'] = False
+    # merge super close events
+    for i in np.arange(0,len(dftemp)):
+        maxF[i] = dftemp.Time[i]
+        minF[i] = maxF[i]-dftemp.dtime[i]
+
+        df2['SRBtt'] = (df2['Time']>minF[i]) & (df2['Time']<maxF[i]) | (df2['SRBtt']==True)
+        df2['SRBall'] = (df2['SRBtt']==True) | (df2['SRBall']==True)
+
+
+    df2['dtimeA'] = df2['dtime']
+    df2.drop('dtime', inplace=True, axis=1)
+
+    # DELETING SHORT SRB
+    # calculating dtime - time of srb
+    dfSRBtemp2 = df2[df2['SRBall']==False]
+    dfSRBtemp2['dtime'] = dfSRBtemp2['Time'].diff(1)
+    df22 = pd.merge(df2, dfSRBtemp2[['dtime', 'Frame', 'session_id','Time']], on=['Frame', 'session_id','Time'], how='outer')
+    df22.fillna(0, inplace=True)
+
+    #special time cases
+    dftemp2 = dfSRBtemp2[(dfSRBtemp2['dtime']<0.4) & (dfSRBtemp2['dtime']>0.0049)]
+    dftemp2 = dftemp2.reset_index(drop=True)
+    maxF2, minF2 = {}, {}
+    df22['SRBall2'] = df22['SRBall']
+    df22['SRBtt2'] = False
+
+    # delete short events
+    for i in np.arange(0,len(dftemp2)):
+        maxF2[i] = dftemp2.Time[i]
+        minF2[i] = dftemp2.Time[i] - dftemp2.dtime[i]
+        df22['SRBtt2'] = ((df22['Time']>=minF2[i]) & (df22['Time']<=maxF2[i])) | (df22['SRBtt2']==True)
+        df22['SRBall2'] = (df22['SRBtt2']==False) & (df22['SRBall2'] == True)
+
+    rat_auto = pd.concat([rat_auto, df22], axis=0, ignore_index=True)
+
+
+with h5py.File(cfg.SRBauto_fname, 'w') as f:
+    f.create_dataset('Rat_Behavior', data=rat_auto.to_records())

notebooks/SRB_auto_1105.ipynb

@@ -1,7 +1,8 @@
-3# File names
+# File names
 relational_fname = 'datasets/preprocessed/relationalDatabase.h5'
-velocity_fname = 'datasets/preprocessed/velocityDatabase.h5'
-filtered_fname = 'datasets/preprocessed/filteredDatabase.h5'
+velocity_fname   = 'datasets/preprocessed/velocityDatabase.h5'
+filtered_fname   = 'datasets/preprocessed/filteredDatabase.h5'
+SRBauto_fname    = 'datasets/preprocessed/SRBautoDatabase.h5'
 
 # Rolling windows
 WIDNOW_DATA = 40
@@ -13,3 +14,11 @@
 FILT_CLEANING_YPOS = 0.07
 FILT_CLEANING_YORI = -0.75
 FILT_DTIME = 0.005
+
+# Automatic
+LIMITS = ({'speed':[ 7, 14,  28, -7, -14  , -28],
+           'min'  :[ 1,  5,  13, -9, -21.5, -30],
+           'max'  :[ 9, 18,  30, -1, -5   , -13]})
+
+CLOSE_SRB_LIM = [0.005 , 0.2]
+SHORT_SRB_LIM = [0.0049, 0.4]