bug fix and doc updates

docs/get_started.rst

@@ -13,7 +13,7 @@ Second, install **SplitWavePy** with the following terminal command.
 .. tip::
 	Ensure you have the latest release.
 
-	``pip upgrade splitwavepy``
+	``pip install splitwavepy --upgrade``
 
 Now try the code.
 

docs/real_data.rst

@@ -91,7 +91,7 @@ It worked, kind of.  The maximum delay time in the grid search is a bit high.  W
 	:include-source:
 
 	measure = sw.EigenM(realdata, lags=(2,))
-	measure.plot(marker=True)
+	measure.plot(marker=True,conf95=True)
 
-That looks better.
+That looks better.  Notice I've added a marker at the max :math:`\lambda_2/\lambda_2` location using ``marker=True`` and a contour at the 95% confidence level using ``conf95=True``.
 

docs/tutorial.rst

@@ -109,12 +109,14 @@ It is quite common to plot the :math:`\lambda_2` surface.  To do this you use th
 
 .. _setgrid:
 
-Setting the lag time search range
+Setting the lag time grid search
 ``````````````````````````````````
-The code automatically sets the maximum lag time to be half the window length.  To set the max search time manually you use the ``lags`` keyword.  This accepts a tuple of length 1, 2, or 3, and will be interpreted differently depending on this length.  The rules are as follows: for a 1-tuple ``lags = (maxlag,)``, a 2-tuple ``lags = (maxlag, nlags)``, and finally a 3-tuple ``(minlag, maxlag, nlags)``.  In the previous example we can see that the solution has a delay time less than 2 seconds, so let's repeat the measurement using a more sensible grid search.
+The code automatically sets the maximum lag time to be half the window length.  To set the max search time manually you use the ``lags`` keyword.  This accepts a tuple of length 1, 2, or 3, and will be interpreted differently depending on this length.  The rules are as follows: for a 1-tuple ``lags = (maxlag,)``, a 2-tuple ``lags = (maxlag, nlags)``, and finally a 3-tuple ``(minlag, maxlag, nlags)``.  Alternatively will accept a numpy array containing all nodes to search.
 
-.. note::
-	The code automatically grid searches every 2 degrees along the fast direction axis.  That's ``ndegs = 90`` nodes in total (180/2).  You can change this number using the ``ndegs`` keyword.
+Setting the fast direction grid search
+````````````````````````````````````````
+
+The code automatically grid searches every 2 degrees along the fast direction axis.  That's ``degs = 90`` nodes in total (180/2).  You can change this number using the ``degs`` keyword and providing an integer.  Alternatively will accept a numpy array containing all nodes to search.
 
 
 Tabulating the result

release_notes.txt

@@ -9,12 +9,10 @@ checklist:
 	git tag        (shows a list of tags)
 	git push --tags origin master (to put it on github)
 
----OPTIONAL TEST SERVER
-• register package on PyPI Test server (as a check eveything's working)
-	python setup.py register -r pypitest
-• upload to test PyPI repository
+UPLOAD TO PYPI
+
+-- (optional test first)
 	python setup.py sdist upload -r pypitest
 
 ---THE REAL DEAL 
-	python setup.py register -r pypi
 	python setup.py sdist upload -r pypi

setup.py

@@ -9,9 +9,6 @@
     license = 'MIT',
     url = 'https://github.com/JackWalpole/splitwavepy',
     packages = find_packages(),
-    install_requires = [
-        'matplotlib',
-        'numpy',
-        'scipy']
-    keywords = ['shear', 'wave', 'splitting', 'seismic', 'anisotropy','seismology']
+    install_requires = ['matplotlib','numpy','scipy'],
+    keywords = 'shear wave splitting seismic anisotropy seismology',
 )

splitwavepy/core/pair.py

@@ -211,7 +211,7 @@ def set_window(self,*args,**kwargs):
             time_centre = (start + end)/2
             time_width = end - start
             tcs = core.time2samps(time_centre,self.delta)
-            offset = tcs - self.centresamp()
+            offset = tcs - self._centresamp()
             # convert time to nsamples -- must be odd
             width = core.time2samps(time_width,self.delta,'odd')       
             self.window = Window(width,offset,**kwargs)