removed depreacted messages for type instead of mode, as well as gaus…

…sian smoothed changed behaviour.

Added residual return on full mode in spec fit.

oned/base.py

@@ -162,12 +162,7 @@ def from_fits(cls, filename, **kwargs):
         return cls(wave, flux, mode='waveflux')
 
     def __init__(self, *args, **kwargs):
-        #deprecate type soon
-        if kwargs.has_key('type') and not kwargs.has_key('mode'):
-            print "Usage of keyword type is deprecated. Please use mode instead"
-            kwargs['mode'] = kwargs['type']
-
-
+
         if kwargs.has_key('mode'):
             if kwargs['mode'] == 'ndarray':
 
@@ -433,7 +428,7 @@ def gaussian_smooth(self, kernel, **kwargs):
         of <kernel> pixels.
         
         """
-        print "Deprecated: method has changed. It now returns a new instance with the smoothing applied"
+
 
         newFlux = ndimage.gaussian_filter1d(self.flux, kernel, **kwargs)
 

oned/util.py

@@ -587,22 +587,27 @@ def normalise(spectrum, function='spline',
         return (spectrum / continuum, continuum, continuum_regions, coeffs)
 
 def continuum2(spectrum, low_rej=2., high_rej=3., function='legendre', maxiter=3, order=5, mode='normal'):
-    def fitLegendre(x, y):
+    def fitLegendre(x, y, mode='fit'):
         p = np.polynomial.Legendre.fit(x, y, order)
-        return p(spectrum.wave)
+        if mode=='fit':
+            return p(spectrum.wave)
+        if mode=='func':
+            return p
 
-    def fitChebyshev(x, y):
+    def fitChebyshev(x, y, mode='fit'):
         p = np.polynomial.Chebyshev.fit(x, y, order)
-        return p(spectrum.wave)
-
+        if mode=='fit':
+            return p(spectrum.wave)
+        if mode=='func':
+            return p
     if function=='legendre':
         fitfunc = fitLegendre
 
     if function=='chebyshev':
         fitfunc = fitChebyshev
 
     contFlux = fitfunc(spectrum.wave, spectrum.flux)
-    mask = np.ones(spectrum.wave.shape).astype(bool)
+    mask = spectrum.dq
     high_rej_mask = np.zeros(spectrum.wave.shape).astype(bool)
     low_rej_mask = np.zeros(spectrum.wave.shape).astype(bool)
     for i in range(maxiter):
@@ -614,9 +619,16 @@ def fitChebyshev(x, y):
         contFlux = fitfunc(spectrum.wave[mask], spectrum.flux[mask])
     if mode == 'normal':
         return onedspec(spectrum.wave, contFlux, mode='waveflux')
+
     elif mode == 'rms':
         residual = (spectrum.flux[mask] - contFlux[mask])
         return onedspec(spectrum.wave, contFlux, mode='waveflux'), np.std(residual)
+
+    elif mode == 'full':
+        residual = (spectrum.flux[mask] - contFlux[mask])
+        return (onedspec(spectrum.wave, contFlux, mode='waveflux'),
+                onedspec(spectrum.wave, residual, mode='waveflux'),
+                fitfunc(spectrum.wave[mask], spectrum.flux[mask], mode='func'))