[Predator] Move ``SingleFeatureScore`` to LLM.

appengine/findit/crash/loglinear/changelist_classifier.py

Index: appengine/findit/crash/loglinear/changelist_classifier.py
diff --git a/appengine/findit/crash/loglinear/changelist_classifier.py b/appengine/findit/crash/loglinear/changelist_classifier.py
index f25430adb419f556fe9994aa53b352e9a5f84df1..01ece27024cd002cc479f8e25216f29675cef7ed 100644
--- a/appengine/findit/crash/loglinear/changelist_classifier.py
+++ b/appengine/findit/crash/loglinear/changelist_classifier.py
@@ -49,10 +49,8 @@ class LogLinearChangelistClassifier(object):
         weights['MinDistance'],
     ]
 
-    self._model = UnnormalizedLogLinearModel(feature_function, weight_list)
-
-    # TODO(crbug.com/674262): remove the need for storing these weights.
-    self._weights = weights
+    self._model = UnnormalizedLogLinearModel(feature_function,
+                                             weight_list, weights)

[wrengr, 2017/01/09 19:31:51]

we shouldn't have to pass both the weight dict and list; the dict should be
sufficient.

 
   # TODO(crbug.com/673964): something better for detecting "close to log(0)".
   def _LogZeroish(self, x):
@@ -72,20 +70,6 @@ class LogLinearChangelistClassifier(object):
     """
     return x < 0 and math.isinf(x)
 
-  def _SingleFeatureScore(self, feature_value):
-    """Returns the score (aka weighted value) of a ``FeatureValue``.
-
-    This function assumes the report's stacktrace has already had any necessary
-    preprocessing (like filtering or truncating) applied.
-
-    Args:
-      feature_value (FeatureValue): the feature value to check.
-
-    Returns:
-      The score of the feature value.
-    """
-    return feature_value.value * self._weights.get(feature_value.name, 0.)
-
   def __call__(self, report):
     """Finds changelists suspected of being responsible for the crash report.
 
@@ -188,7 +172,7 @@ class LogLinearChangelistClassifier(object):
     """
     formatted_reasons = []
     for feature in features:
-      feature_score = self._SingleFeatureScore(feature)
+      feature_score = self._model.SingleFeatureScore(feature)
       if self._LogZeroish(feature_score): # pragma: no cover
         logging.debug('Discarding reasons from feature %s'
             ' because it has zero probability' % feature.name)
@@ -216,7 +200,8 @@ class LogLinearChangelistClassifier(object):
     """
     all_changed_files = {}
     for feature in features:
-      if self._LogZeroish(self._SingleFeatureScore(feature)): # pragma: no cover
+      if self._LogZeroish(
+          self._model.SingleFeatureScore(feature)): # pragma: no cover
         logging.debug('Discarding changed files from feature %s'
             ' because it has zero probability' % feature.name)
         continue