more faffing about prepping for release

MANIFEST.in

@@ -1,2 +1,5 @@
 include .coveragerc tox.ini
-include TODO COPYING
+include TODO COPYING README
+recursive-include tools *.py
+recursive-include doc *
+prune doc/_build

README

@@ -1,6 +1,9 @@
 This is patsy, a Python library for describing statistical models
 and building design matrices.
 
+Documentation:
+  http://patsy.readthedocs.org/
+
 Dependencies:
   numpy
   Python 2.4 or later (including Python 3)
@@ -12,7 +15,7 @@ Install:
   pip install patsy
 
 Code and bug tracker:
-  https://github.com/patsy/patsy
+  https://github.com/pydata/patsy
 
 Mailing list:
   pydata@googlegroups.com

doc/formulas.rst

@@ -726,7 +726,7 @@ for this reference manual, but here's a sketch of the analysis.
 Recall that our goal is to maintain two invariants: the design matrix
 column space should include the space associated with each term, and
 should avoid "structural redundancy", i.e. it should be full rank on
-at least some data sets. It's easy to see the this algorithm will
+at least some data sets. It's easy to see the above algorithm will
 never "lose" columns, since the only time it eliminates a subspace is
 when it has previously processed that exact subspace within the same
 design. (So long as the subspace merging is correctly specified etc.;

doc/library-developers.rst

@@ -103,6 +103,9 @@ Example
 Here's a simplified class for doing ordinary least-squares regression,
 demonstrating the above techniques:
 
+.. warning:: This code has not been validated for numerical
+   correctness.
+
 .. include:: _examples/example_lm.py
    :code:
 

doc/overview.rst

@@ -37,7 +37,9 @@ it:
 * Has an incremental mode to handle data sets which are too large to
   fit into memory at one time,
 * Provides a language for symbolic, human-readable specification of
-  linear constraint matrices, and
+  linear constraint matrices,
+* Has a thorough test suite and solid underlying theory, allowing it
+  to correctly handle corner cases that even R gets wrong, and
 * Features a simple API for integration into statistical packages.
 
 What Patsy *won't* do is, well, statistics --- it just lets you

release-checklist.txt

@@ -2,13 +2,16 @@
 * make sure docs are up to date
 * check MANIFEST.in
 * update version in doc/changes.rst, setup.py, patsy/__init__.py
-* make a clean checkout
-* run tox tests
+* make sure there are no uncommitted changes
+* clone a clean source directory (so as to get a clean checkout
+  without killing the expensive-to-recreate .tox dir)
+* python sdist
+* unzip sdist to a clean directory, and run tox in it
 * tag
 * upload source dist to pypi
-* build and upload docs
 * announce release on:
-  py
+  pydata@googlegroups.com
+  statsmodels
   scipy-dev?
   pypi
 * update version in setup.py, patsy/__init__.py again (add "+dev")