Bug for missing context score fixed.

felixleopoldo · Feb 7, 2024 · 02f14a1 · 02f14a1
1 parent e4f67f0
commit 02f14a1
Show file tree

Hide file tree

Showing 2 changed files with 8 additions and 4 deletions.
diff --git a/src/cstrees/learning.py b/src/cstrees/learning.py
@@ -165,7 +165,8 @@ def _optimal_staging_at_level(
             # here we (=I) anyway extract just the context, so the stage format
             # is a bit redundant.
             stage_context = sc._stage_to_context_key(stage, order)
-            score = context_scores["scores"][var][stage_context]
+            if stage_context in context_scores["scores"][var]:
+                score = context_scores["scores"][var][stage_context]
             staging_score += score
 
         # Update the max score and the max staging

diff --git a/src/cstrees/scoring.py b/src/cstrees/scoring.py
@@ -10,6 +10,8 @@
 import cstrees.stage as st
 import cstrees.dependence as csi_rel
 
+import pp
+
 
 def _counts_at_level(cstree: ct.CStree, level: int, data):
     """Collect all the observed counts at a specific level by stages.
@@ -264,9 +266,8 @@ def _context_score_tables(
                 # get the counts
                 testdf = test.to_frame().rename(
                     columns={var: str(var) + " counts"})
-                for index, r in testdf.iterrows():
+                for index, r in testdf.iterrows(): #index represents the context variables and the value of the variable
                     value = None
-
                     # Sort variables
                     context = ""
                     if len(active_labels) > 0:
@@ -518,7 +519,9 @@ def order_score_tables(
                 for stage in staging:
                     # OK! even when restricting to some possible cvars
                     stage_context = _stage_to_context_key(stage, subset)
-                    staging_unnorm_post += context_scores["scores"][var][stage_context]
+                    if stage_context in context_scores["scores"][var]:
+                        # If the context is not in the data it can be ignored as it would vanish in the score.
+                        staging_unnorm_post += context_scores["scores"][var][stage_context]
 
                 if i == 0:
                     order_scores["scores"][var][subset_str] = staging_unnorm_post