From 26810b1ea1d5b052b8a9d12d06f31e11d556ce8a Mon Sep 17 00:00:00 2001
From: Jiao Lin <linjiao@ornl.gov>
Date: Wed, 17 Apr 2019 15:29:10 -0400
Subject: [PATCH] Refs #16. added methods to calculate 2D and 4D PDF

---
 dgsres/singlextal/violini.py | 57 ++++++++++++++++++++++++++++++++++++
 1 file changed, 57 insertions(+)

diff --git a/dgsres/singlextal/violini.py b/dgsres/singlextal/violini.py
index 5fad115..cab8a5e 100644
--- a/dgsres/singlextal/violini.py
+++ b/dgsres/singlextal/violini.py
@@ -46,3 +46,60 @@ class dynamics:
         InvCov4D = cm_res['hklE_inv_cov']
         return np.linalg.inv(InvCov4D)/2.355
     
+    def pdf4D(self, hkl, E):
+        "return a pdf function pdf((h,k,l,E)) gives value of pdf at that point"
+        sigma2 = self.computeCovMat(hkl, E)
+        return create_pdf(sigma2)
+
+    def pdf2D(self, hkl, E, eu, ev, ew):
+        """return a 2D pdf function along uaxis and Eaxis at the point h,k,l,E
+        the pdf is 4D pdf integrated over v and w direction
+
+        eu, ev, ew are unit vectors along uvw
+
+        see https://jupyter.sns.gov/user/{UID}/notebooks/data/SNS/SEQ/IPTS-21411/shared/resolution/violini-04162019.ipynb
+        for an example
+        """
+        # compute covmat
+        cm = self.computeCovMat(hkl, E)
+        cm = np.matrix(cm)
+        # inverse
+        ic = cm.I
+        # transform the coordinate system
+        A = np.array( (eu,ev,ew) ).T
+        A = np.block( [[A, np.zeros((3,1))],
+                       [np.zeros((1,3)), 1]])
+        A = np.matrix(A)
+        cm0 = cm
+        # print cm0
+        cm = A.T * cm * A # u,v,w,E
+        # print cm
+        # integration along v and w
+        cm2d = [[cm[0,0] - (cm[0,1]+cm[1,0])**2/4./cm[1,1] - (cm[0,2]+cm[2,0])**2/4./cm[2,2],
+                 cm[0,3]],
+                [cm[3,0],
+                 cm[3,3] - (cm[3,1]+cm[1,3])**2/4./cm[1,1] - (cm[3,2]+cm[2,3])**2/4./cm[2,2]
+                ]]
+        # print cm2d
+        return create_pdf(cm2d)
+
+
+def create_pdf(covmat):
+    "create probability distribution function out of covariance matrix"
+    sigma2 = np.matrix(covmat)
+    det = np.linalg.det(sigma2)
+    # det = np.abs(det)
+    if det == 0:
+        raise ValueError("The covariance matrix can't be singular")
+    size, size2 = sigma2.shape
+    assert size == size2
+    # print "det", det
+    norm_const = 1.0/ ( np.power((2*np.pi),float(size)/2) * np.power(det,1.0/2) )
+    # print "norm_const", norm_const
+    inv = sigma2.I
+    # print "inv", inv
+    def _(x):
+        x = np.matrix(x)
+        result = np.exp( -0.5 * (x * inv * x.T) )
+        return norm_const * result
+    return _