shimazaki / adaptivekde Goto Github PK

using your function : sskernel :
while anlyzing multiple data sessions I encountered the following problem :

Traceback (most recent call last):

  File "/home/azare/groups/PrimNeu/Aryo/analysis/fit/automate_mix_psth_fit_nmz_10_re_sig_8_unit_14_opt.py", line 458, in <module>
    kv , t , okbw = KDE(l_f_8_soi_ms[i])

  File "/home/azare/groups/PrimNeu/Aryo/analysis/opt/KDE.py", line 133, in KDE
    y = y * len(spike_times)

UnboundLocalError: local variable 'y' referenced before assignment

The problematic input array :
np.array([425.36666667, 230.6 , 35.86666667])

While an array of only 3 elemtns is not realistic to be considered as data, this was running through a loop of hundreds of sessions, hence such input data is also imaginable.
The program stuck in the middle of a loop with the above error.

It seems to me that in this case, your while statement was never satisfied (not 'True') :

The code needs to cast from a float to an int when making the call to np.ceil(). Right now it wont run unless the code is normalized to integer values.

I would like to mention about adding an option to the current sskernel or ssvkernel functions :
It seems to me that your functions normalize the plot so that the area under the curve would be 1.
The similar function in your colleague's (Shinomoto) website returns similar plots but the y axis has a different scale (possibly not normalized).
https://s-shinomoto.com/histograms/python/KDE_rate_v4.py

While normalization has particular advantages, I would also like to get an estimated real (absolute) firing rate.
The reason is that I would like to compare the firing rate of different sessions.
How can this be achieved ?

First, many thanks for providing this package.

I would like to mention a small bug in this program :
The following line outputs a float :

The line thereafter ('range' function) accepts an integer :

This can simply be fixed by type-casting when defining 'N_MAX' :
N_MAX = int ( min(np.floor((x_max - x_min) / (2*dx)), max(N)) )

I would like to mention a small but important improvement to the program.

np.int & np.float are aliases of the python operator 'int' , 'float' respectively. They are deprecated in newer versions of numpy.
Usage of them does not make any error but since they are extensively used in your program, there are over 10 warnings per analysis session, which if used in a loop of many sessions, may cause crashing of the IDE and may decrease the speed of the program.

Just using 'int' or 'float' instead is safe & is preferred.
Below I put the program warnings :

N = sum(y_hist).astype(np.float)
/home/azare/anaconda3/envs/env_2/lib/python3.10/site-packages/adaptivekde/ssvkernel.py:99: DeprecationWarning: np.float is a deprecated alias for the builtin float. To silence this warning, use float by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use np.float64 here.
Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations

f = np.linspace(0, n-1, n.astype(np.int)) / n
/home/azare/anaconda3/envs/env_2/lib/python3.10/site-packages/adaptivekde/ssvkernel.py:268: DeprecationWarning: np.int is a deprecated alias for the builtin int. To silence this warning, use int by itself. Doing this will not modify any behavior and is safe. When replacing np.int, you may wish to use e.g. np.int64 or np.int32 to specify the precision. If you wish to review your current use, check the release note link for additional information.
Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations

hi i cannot find any documentation on how to use this?

Dear developers of AdaptiveKDE method,

I am trying to use this package for estimating the temporal firing rate profile from spike trains of several trials. I had a look a the function "ssvkernel" and I am having trouble understanding the input parameters of this function. The argument "x" says it requires "The one-dimensional samples drawn from the underlying density". I am not able to figure out how to transform information in the trial spike trains of my data before supplying it to "x". Could you provide an example use case, preferably, the code used to generate Fig. 4(b) or (c) of the reference?

Thank you.

Sincerely
Ajay