make viz.fit_line a public function

Author: phobson

probscale/tests/test_viz.py

@@ -22,7 +22,7 @@
 BASELINE_DIR = 'baseline_images/test_viz'
 
 
-class Test__fit_line(object):
+class Test_fit_line(object):
     def setup(self):
         self.data = numpy.array([
             2.00,   4.0 ,   4.62,   5.00,   5.00,   5.50,   5.57,   5.66,
@@ -71,56 +71,56 @@ def setup(self):
     def test_xlinear_ylinear(self):
         scales = {'fitlogs': None, 'fitprobs': None}
         x, y = self.zscores, self.data
-        x_, y_, res = viz._fit_line(x, y, **scales)
+        x_, y_, res = viz.fit_line(x, y, **scales)
         nptest.assert_array_almost_equal(y_, self.known_y_linlin)
         assert isinstance(res, numpy.ndarray)
 
     def test_xlinear_ylog(self):
         scales = {'fitlogs': 'y', 'fitprobs': None}
         x, y = self.zscores, self.data
-        x_, y_, res = viz._fit_line(x, y, **scales)
+        x_, y_, res = viz.fit_line(x, y, **scales)
         nptest.assert_array_almost_equal(y_, self.known_y_linlog)
         assert isinstance(res, numpy.ndarray)
 
     def test_xlinear_yprob(self):
         scales = {'fitlogs': None, 'fitprobs': 'y'}
         x, y = self.data, self.probs
-        x_, y_, res = viz._fit_line(x, y, **scales)
+        x_, y_, res = viz.fit_line(x, y, **scales)
         nptest.assert_array_almost_equal(y_, self.known_y_linprob)
         assert isinstance(res, numpy.ndarray)
 
     def test_xlog_ylinear(self):
         scales = {'fitlogs': 'x', 'fitprobs': None}
         x, y = self.data, self.zscores
-        x_, y_, res = viz._fit_line(x, y, **scales)
+        x_, y_, res = viz.fit_line(x, y, **scales)
         nptest.assert_array_almost_equal(y_, self.known_y_loglin)
         assert isinstance(res, numpy.ndarray)
 
     def test_xlog_ylog(self):
         scales = {'fitlogs': 'both', 'fitprobs': None}
         x, y = self.data, self.y
-        x_, y_, res = viz._fit_line(x, y, **scales)
+        x_, y_, res = viz.fit_line(x, y, **scales)
         nptest.assert_array_almost_equal(y_, self.known_y_loglog)
         assert isinstance(res, numpy.ndarray)
 
     def test_xlog_yprob(self):
         scales = {'fitlogs': 'x', 'fitprobs': 'y'}
         x, y = self.data, self.probs
-        x_, y_, res = viz._fit_line(x, y, **scales)
+        x_, y_, res = viz.fit_line(x, y, **scales)
         nptest.assert_array_almost_equal(y_, self.known_y_logprob)
         assert isinstance(res, numpy.ndarray)
 
     def test_xprob_ylinear(self):
         scales = {'fitlogs': None, 'fitprobs': 'x'}
         x, y = self.probs, self.data
-        x_, y_, res = viz._fit_line(x, y, **scales)
+        x_, y_, res = viz.fit_line(x, y, **scales)
         nptest.assert_array_almost_equal(y_, self.known_y_problin)
         assert isinstance(res, numpy.ndarray)
 
     def test_xprob_ylog(self):
         scales = {'fitlogs': 'y', 'fitprobs': 'x'}
         x, y = self.probs, self.data
-        x_, y_, res = viz._fit_line(x, y, **scales)
+        x_, y_, res = viz.fit_line(x, y, **scales)
         nptest.assert_array_almost_equal(y_, self.known_y_problog)
         assert isinstance(res, numpy.ndarray)
 
@@ -139,23 +139,23 @@ def test_xprob_yprob(self):
 
         scales = {'fitlogs': None, 'fitprobs': 'both'}
         x, y = self.probs, p2,
-        x_, y_, res = viz._fit_line(x, y, **scales)
+        x_, y_, res = viz.fit_line(x, y, **scales)
         nptest.assert_array_almost_equal(y_, self.known_y_probprob)
         assert isinstance(res, numpy.ndarray)
 
     def test_bad_fitlogs(self):
         with pytest.raises(ValueError):
             x, y = self.zscores, self.data
-            x_, y_, res = viz._fit_line(x, y, fitlogs='junk')
+            x_, y_, res = viz.fit_line(x, y, fitlogs='junk')
 
     def test_bad_fitprobs(self):
         with pytest.raises(ValueError):
             x, y = self.zscores, self.data
-            x_, y_, res = viz._fit_line(x, y, fitprobs='junk')
+            x_, y_, res = viz.fit_line(x, y, fitprobs='junk')
 
     def test_custom_xhat(self):
         x, y = self.zscores, self.data
-        x_, y_, res = viz._fit_line(x, y, xhat=self.custom_xhat)
+        x_, y_, res = viz.fit_line(x, y, xhat=self.custom_xhat)
         nptest.assert_array_almost_equal(y_, self.known_custom_yhat)
 
 

probscale/viz.py

@@ -189,7 +189,7 @@ def probplot(data, ax=None, plottype='prob', dist=None, probax='x',
 
     # maybe do a best-fit and plot
     if bestfit:
-        xhat, yhat, modelres = _fit_line(x, y, fitprobs=fitprobs, fitlogs=fitlogs, dist=dist)
+        xhat, yhat, modelres = fit_line(x, y, fitprobs=fitprobs, fitlogs=fitlogs, dist=dist)
         ax.plot(xhat, yhat, **line_kws)
     else:
         xhat, yhat, modelres = (None, None, None)
@@ -337,7 +337,8 @@ def _set_prob_limits(ax, probax, N):
         ax.set_ylim(bottom=minval, top=100-minval)
 
 
-def _fit_line(x, y, xhat=None, fitprobs=None, fitlogs=None, dist=None):
+def fit_line(x, y, xhat=None, fitprobs=None, fitlogs=None, dist=None,
+             estimate_ci=False, niter=10000, alpha=0.05):
     """
     Fits a line to x-y data in various forms (linear, log, prob scales).