V0.21.2 (#715)

* Doc improvements

* Adding piecwise regression fitter (no tests yet though)

* New default behaviour for `predict` in KaplanMeier and NelsonAalen: use the previous value, interpolate with interpolate flag.

* docs and changelog

* Adding some tests

* clean tests

* only tests /tests

* some cleanup

* change how I internally track how censoring is kept

* bump version

* fix tests

CHANGELOG.md

@@ -1,5 +1,18 @@
 ### Changelog
 
+#### 0.21.2 - 2019-04-26
+
+##### New features
+ - New regression model: `PiecewiseExponentialRegressionFitter` is available. See blog post here: https://dataorigami.net/blogs/napkin-folding/churn
+ - Regression models have a new method `log_likelihood_ratio_test` that computes, you guessed it, the log-likelihood ratio test. Previously this was an internal API that is being exposed.
+
+##### API changes
+ - The default behavior of the `predict` method on non-parametric estimators (`KaplanMeierFitter`, etc.) has changed from (previous) linear interpolation to (new) return last value. Linear interpolation is still possible with the `interpolate` flag.
+ - removing `_compute_likelihood_ratio_test` on regression models. Use `log_likelihood_ratio_test` now.
+
+##### Bug fixes
+
+
 #### 0.21.1 - 2019-04-26
 
 ##### New features

Makefile

@@ -9,15 +9,22 @@ else
 endif
 
 test:
-	py.test -rfs --cov=lifelines --block=False --cov-report term-missing
+	py.test tests/ -rfs --cov=lifelines --block=False --cov-report term-missing
 
 lint:
+ifeq ($(TRAVIS_PYTHON_VERSION), 2.7)
+		echo "Skip linting for Python2.7"
+else
+		make black
+		prospector --output-format grouped
+endif
+
+black:
 ifeq ($(TRAVIS_PYTHON_VERSION), 2.7)
 		echo "Skip linting for Python2.7"
 else
 		black lifelines/ -l 120 --fast
 		black tests/ -l 120 --fast
-		prospector --output-format grouped
 endif
 
 check_format: