From 6efac9556cfc427fb9b5adfe897b99ed8ea79960 Mon Sep 17 00:00:00 2001
From: Nicolas Unger <nicounger94@gmail.com>
Date: Fri, 14 Sep 2018 17:40:35 +0200
Subject: [PATCH] Added Truncated Rayleigh distribution

---
 run/priors.py | 26 ++++++++++++++++++++++++++
 1 file changed, 26 insertions(+)

diff --git a/run/priors.py b/run/priors.py
index 7930dac..1e997dd 100644
--- a/run/priors.py
+++ b/run/priors.py
@@ -242,6 +242,30 @@ def _ppf(self, q, mu, sigma, xmin, xmax):
         # Interpolate the _inverse_ CDF
         return interpolate.interp1d(cdf, x)(q)
 
+class truncrayleigh_gen(rv_continuous):
+    def _argcheck(self, sigma, xmax):
+        return (sigma > 0)
+
+    def _pdf(self, x, sigma, xmax):
+        A1 = 1 - n.exp(-xmax**2/(2*sigma**2))
+        n1 = ((x/sigma**2) * n.exp(-x**2/(2*sigma**2)))/A1
+        return n.where((x >= 0) & (x < xmax), n1, 0.0)
+
+    def _cdf(self, x, sigma, xmax):
+        A1 = 1 - n.exp(-xmax**2/(2*sigma**2))
+        cdf = 1 - n.exp(-x**2/(2*sigma**2))
+        cdf = cdf/A1
+        # Consider the limits of the uniform
+        cdf = n.where((x >= 0), cdf, 0.0)
+        cdf = n.where((x < xmax), cdf, 1.0)
+        return cdf
+
+    def ppf(self, q, sigma, xmax):
+        A = 1 - n.exp(-xmax**2/(2*sigma**2))
+        ppf = n.sqrt(-2*sigma**2*n.log(1-(q*A)))
+        return ppf
+
+
     
 class truncnormJ_gen(rv_continuous):
     def _argcheck(self, mu, sigma, xmin, xmax):
@@ -434,6 +458,8 @@ def _ppf(self, q, alpha, beta):
                                       shapes='mu, sigma1, sigma2')
 TruncatedUNormal = truncnormU_gen(name='Truncated normal distribution',
                                   shapes='mu, sigma, xmin, xmax')
+TruncatedRayleigh = truncrayleigh_gen(name='Truncated Rayleigh Distribution',
+                                     shapes='sigma, xmax')
 PowerLaw = powerlaw_gen(name='Power law distribution',
                         shapes='alpha, xmin, xmax')
 DoublePowerLaw = doublepowerlaw_gen(name='Double Power law distribution',