#!/usr/bin/env python
# coding: utf-8

# # Plotting routines in `touchsim`
# The `touchsim` package uses `holoviews` for plotting.

# In[ ]:


import touchsim as ts
from touchsim.plotting import plot, figsave
import numpy as np
import holoviews as hv
hv.notebook_extension()
get_ipython().run_line_magic('output', "holomap='scrubber' # animate holomaps")

import warnings
warnings.simplefilter("ignore")


# ## Hand model

# In[ ]:


plot() # short for plot(ts.hand_surface)


# Regions labels and the coordinate system can be overlaid on the plot.

# In[ ]:


get_ipython().run_cell_magic('output', 'size=250', 'plot(tags=True,coord=10) # coord sets the lengths of the coordinate axes in mm\n')


# ## Visualising `AfferentPopulation` objects

# In[ ]:


a = ts.affpop_hand(region='D2d')
plot(a,size=10)


# Plots can be overlaid using the `*` operator.
# 
# Plots are `holoviews` objects and can be indexed to only show, say, a specific afferent population. The `*` operator overlays different plots (e.g. the hand outline and the afferent locations, as shown below).

# In[ ]:


plot(region='D2d') * plot(a,size=10)['PC']


# New subpanels can be added using the `+` operator.

# In[ ]:


(plot(region='D2d') * plot(a,size=10)['SA1']) + (plot(region='D2d') * plot(a,size=10)['RA']) + (plot(region='D2d') * plot(a,size=10)['PC'])


# The first index of an `AfferentPopulation` plot object is the afferent type, while the next two indices are pixel coordinates.

# In[ ]:


# zoom into fingertip
plot(a,size=10)[:,120:140,450:475]


# ## Visualising `Stimulus` objects
# Plotting a `Stimulus` object shows the trace of all pins by default.

# In[ ]:


s = ts.stim_ramp(len=0.25,amp=.1,ramp_len=0.05)
plot(s)


# In[ ]:


s += ts.stim_sine(freq=25.,len=.25,loc=[1.,1.])
plot(s)


# Pin traces can also be shown in a grid view.

# In[ ]:


plot(s,grid=True)


# Pin positions can also be shown spatially.

# In[ ]:


s = ts.stim_indent_shape(ts.shape_circle(hdiff=0.5,pins_per_mm=2,radius=3),ts.stim_ramp(len=0.1))
plot(s,spatial=True)


# The spatial view can be animated with pin depths indicated by color.

# In[ ]:


plot(region='D2d') * plot(s,spatial=True,bin=10)


# ## Visualising `Response` objects
# Plotting a `Response` object shows the spike trains of all included neurons

# In[ ]:


a = ts.affpop_hand(region='D2')
s = ts.stim_sine(freq=50.,amp=0.1,len=0.5)
r = a.response(s)
plot(r)


# The second index is the time index.

# In[ ]:


plot(r)[:,0:0.2]


# Responses can also be plotted spatially, in which case the size of each dot scaled with the neuron's firing rate.

# In[ ]:


get_ipython().run_cell_magic('output', 'size=150 # increase size of plot by 150%', "plot(region='D2') * plot(r,spatial=True,scaling_factor=.1)\n")


# Alternatively, firing rate can be indicated by color instead:

# In[ ]:


plot(region='D2') * plot(r,spatial=True,scale=False)[:,'RA'] + plot(region='D2') * plot(r,spatial=True,scale=False)[:,'PC']


# Responses in spatial view can also be shown animated if a bin size (in ms) is given.

# In[ ]:


plot(region='D2') * plot(r,spatial=True,bin=10)


# ### Advanced example (might be slow to compute)

# In[ ]:


contact_locs = np.zeros((2,2))
contact_locs[0] = np.array([0.,0.])
contact_locs[1] = np.array([150.,0])
    
a = ts.affpop_hand(noisy=False)
s = ts.stim_indent_shape(contact_locs,ts.stim_ramp(amp=0.75,len=.2,ramp_len=0.05,ramp_type='lin',pin_radius=5.,pad_len=0.025))
r = a.response(s)
plot(r)


# In[ ]:


plot() * plot(r,spatial=True,bin=10)


# ## Saving figures and animations
# Figures can be saved using the `figsave` function; figure size and resolution can be controlled using the `size` and `dpi` parameters, respectively.

# In[ ]:


figsave(plot(),'test_fig',size=150,dpi=50)


# When the `gif` format is selected, animations will be saved as animated gifs; their framerate can be controlled with the `fps` parameter

# In[ ]:


bin = 50
fig = plot(s,bin=bin) + plot()*plot(r,spatial=True,bin=bin)
figsave(fig,'test_gif',size=150,fps=5,fmt='gif')