Repatch (from linux) of OSX IRAF

1 files changed, 112 insertions, 0 deletions

diff --git a/pkg/xtools/ranges/rgwtbind.x b/pkg/xtools/ranges/rgwtbind.x
new file mode 100644
index 00000000..82adeba5
--- /dev/null
+++ b/pkg/xtools/ranges/rgwtbind.x
@@ -0,0 +1,112 @@
+# Copyright(c) 1986 Association of Universities for Research in Astronomy Inc.
+
+include	<pkg/rg.h>
+
+# RG_WTBIN -- Weighted average or median of data.
+#
+# The ranges are broken up into subranges of at most abs (nbin) points and a
+# minimum of max (3, (abs(nbin)+1)/2) (though always at least one bin).  The
+# subranges are weighted averaged if nbin > 1 and medianed if nbin < 1.
+# The output weights are the sum of the weights for each subrange.
+# The output array must be large enough to contain the desired points.
+# If the ranges are merged then the input and output arrays may be the same.
+
+procedure rg_wtbind (rg, nbin, in, wtin, nin, out, wtout, nout)
+
+pointer	rg				# Ranges
+int	nbin				# Maximum points in average or median
+double	in[nin]				# Input array
+double	wtin[nin]			# Input weights
+int	nin				# Number of input points
+double	out[ARB]			# Output array
+double	wtout[ARB]			# Output weights
+int	nout				# Number of output points
+
+int	i, j, k, l, n, npts, ntemp, nsample
+
+double	asumd(), amedd()
+
+errchk	rg_packd
+
+begin
+	# Check for a null set of ranges.
+
+	if (rg == NULL)
+	    call error (0, "Range descriptor undefined")
+
+	# If the bin size is exactly one then move the selected input points
+	# to the output array.
+
+	if (abs (nbin) < 2) {
+	    call rg_packd (rg, in, out)
+	    call rg_packd (rg, wtin, wtout)
+	    nout = RG_NPTS(rg)
+	    return
+	}
+
+	# Determine the subranges and take the median or average.
+
+	npts = abs (nbin)
+	ntemp = 0
+
+	do i = 1, RG_NRGS(rg) {
+	    nsample = 0
+	    if (RG_X1(rg, i) > RG_X2(rg, i)) {
+		j = min (nin, RG_X1(rg, i))
+		k = max (1, RG_X2(rg, i))
+		while (j >= k) {
+		    n = max (0, min (npts, j - k + 1))
+		    if (nsample > 0 && n < max (min (npts, 3), (npts+1)/2))
+			break
+		    k = k - n
+		    nsample = nsample + 1
+		    ntemp = ntemp + 1
+		    wtout[ntemp] = asumd (wtin[k + 1], n)
+		    if (nbin > 0) {
+		        if (wtout[ntemp] != 0.) {
+			    out[ntemp] = 0.
+			    do l = k + 1, k + n
+			        out[ntemp] = out[ntemp] + in[l] * wtin[l]
+		            out[ntemp] = out[ntemp] / wtout[ntemp]
+			} else {
+			    out[ntemp] = 0.
+			    do l = k + 1, k + n
+			        out[ntemp] = out[ntemp] + in[l]
+		            out[ntemp] = out[ntemp] / n
+			}
+		    } else {
+			out[ntemp] = amedd (in[k+1], n)
+		    }
+		}
+	    } else {
+		j = max (1, RG_X1(rg, i))
+		k = min (nin, RG_X2(rg, i))
+	        while (j <= k) {
+		    n = max (0, min (npts, k - j + 1))
+		    if (nsample > 0 && n < max (min (npts, 3), (npts+1)/2))
+			break
+		    nsample = nsample + 1
+		    ntemp = ntemp + 1
+		    wtout[ntemp] = asumd (wtin[j], n)
+		    if (nbin > 0) {
+			if (wtout[ntemp] != 0.) {
+			    out[ntemp] = 0.
+			    do l = j, j + n - 1
+		                out[ntemp] = out[ntemp] + in[l] * wtin[l]
+			    out[ntemp] = out[ntemp] / wtout[ntemp]
+			} else {
+			    out[ntemp] = 0.
+			    do l = j, j + n - 1
+		                out[ntemp] = out[ntemp] + in[l]
+			    out[ntemp] = out[ntemp] / n
+			}
+		    } else {
+			out[ntemp] = amedd (in[j], n)
+		    }
+		    j = j + n
+		}
+	    }
+	}
+
+	nout = ntemp
+end