------------------------------------------------------------------
 A variational autoencoder for minimally supervised
pulse shape discrimination by Abdullah Abdulaziz(1), Jianxin Zhou(2), Ming Fang(2), Stephen McLaughlin(1), Angela Di Fulvio(2), Yoann
Altmann(1)
(1) School of Engineering and Physical Sciences, Heriot-Watt University, Edinburgh, U.K.
(2) Department of Nuclear, Plasma, and Radiological Engineering, University of Illinois Urbana-Champaign,
Urbana, U.S.A.

     ==================================================================
     Binary files of the deuterated siltbene pulses and a Python parser and a PSD parser
     ==================================================================

                            Binary files
                            -----------

There are 10 binary files of the detector pulses. Names are "case1" to "case10".
Each case contains neutron and gamma pulses with a different ratio. 
The acquisition methods are described in Section 3.2 Measured Data of the manuscript above.
The data sets are listed below, the ratio is the ratio between number of gamma-ray pulses and neutron pulses,
the classification is performed using the PSD CI method, also available in the attached PSD_measured_data.py script:

	1. "case1", gamma pulses 376180, neutron pulses 63597, ratio 6.9. 
	The detector measured the PuBe source at 120cm and the Cs-137 source at 30cm for 10 min.

	2. "case2", gamma pulses 253197, neutron pulses 59243, ratio 4.8. 
	The detector measured the PuBe source at 120cm and the Cs-137 source at 40cm for 10 min.

	3. "case3", gamma pulses 192813, neutron pulses 56852, ratio 3.8. 
	The detector measured the PuBe source at 120cm and the Cs-137 source at 50cm for 10 min.

	4. "case4", gamma pulses 156009, neutron pulses 55790, ratio 3.1. 
	The detector measured the PuBe source at 120cm and the Cs-137 source at 60cm for 10 min.

	5. "case5", gamma pulses 132483, neutron pulses 55250, ratio 2.6. 
	The detector measured the PuBe source at 120cm and the Cs-137 source at 70cm for 10 min.

	6. "case6", gamma pulses 116691, neutron pulses 55034, ratio 2.3 
	The detector measured the PuBe source at 120cm and the Cs-137 source at 80cm for 10 min.

	7. "case7", gamma pulses 104459, neutron pulses 54333, ratio 2.1. 
	The detector measured the PuBe source at 120cm and the Cs-137 source at 90cm for 10 min.

	8. "case8", gamma pulses 97022, neutron pulses 55031, ratio 1.9. 
	The detector measured the PuBe source at 120cm and the Cs-137 source at 100cm for 10 min. 

	9. "case9", gamma pulses 91329, neutron pulses 54921, ratio 1.8. 
	The detector measured the PuBe source at 120cm and the Cs-137 source at 110cm for 10 min.

	10. "case10", gamma pulses 58353, neutron pulses 54342, ratio 1.2. 
	The detector measured the PuBe source at 120cm for 10 min.

The neutron and gamma pulses are seperated with the traditional PSD technique (charge intergation method).



                            Python parser
                            -----------

The "Parser.py" file in the folder is a python script to read the binary files and write the pulses in the text files.
	1. Put the python script and the binary files in the same folder. 

	2. At line 76 in the python script, add the name of the file that you would like process in "fileName='xxxx.bin'"

	3. Execute the script, and you will get two .txt files. The "Pulse_waveform.txt" file contains the waveforms of every pulses.
	Each row represets a single pulse. The unit of the values are Volt(V). Each pulse has 296 time stamps and each time stamp represents 2 ns. 
	Thereofore the length of each waveform is 592 ns. The "Pulse_label.txt" file contains the label of the pulses. The value "0" means this 
	pulse is classified as a gamma-ray pulse by the traditional PSD. The value "1" means this pulse is classified as a neutron pulse by the PSD.


                            PSD parser
                            -----------

The "PSD_measured_data.py" file in the folder is a python script to read the pulses waveform in the text files and export the PSD result. The formate of the pulse waveform file must be: each row represent a single pulse. 

	1. Put the PSD script and the pulse waveform text file in the same folder. 

	2. At line 18 in the PSD script, put the name of the file that you would like process in "data=np.loadtxt('xxxx.txt')"

	3. Execute the script, and you will get the PSD result files, 'PSD_label.txt'. It contains the label of the pulses. The value "0" means this pulse is classified as a gamma pulse by the traditional PSD. The value "1" means this pulse is classified as a neutron pulse by the PSD.