Update utils.py

utils.py

@@ -181,14 +181,29 @@ def build_df5(oneVoice, multiVoice, sumNoVoice, sumOneVoice, sumMultiVoice, curr
     multiSpeakerList = [p[0] for p in mv_pairs]
     multiTimeList    = [p[1] for p in mv_pairs]
 
+    safeTotalTime  = currTotalTime if currTotalTime > 0 else 1
+    safeOneVoice   = sumOneVoice   if sumOneVoice   > 0 else 1
     summativeMulti = sum(multiTimeList) if multiTimeList else 1
-    safeOneVoice   = sumOneVoice if sumOneVoice > 0 else 1
 
-    safeTotalTime = currTotalTime if currTotalTime > 0 else 1
     base = [sumNoVoice / safeTotalTime, sumOneVoice / safeTotalTime, sumMultiVoice / safeTotalTime]
 
-    timeStrings      = su.timeToString(timeList)      if timeList      else []
-    multiTimeStrings = su.timeToString(multiTimeList) if multiTimeList else []
+    # Normalize child times so they sum exactly to their parent's total.
+    # This guarantees branchvalues="total" is satisfied regardless of whether
+    # the classifier's segment durations perfectly match sumOneVoice/sumMultiVoice.
+    sumTimeList = sum(timeList) if timeList else 0
+    if sumTimeList > 0:
+        normTimeList = [t / sumTimeList * sumOneVoice for t in timeList]
+    else:
+        normTimeList = timeList
+
+    sumMultiTimeList = sum(multiTimeList) if multiTimeList else 0
+    if sumMultiTimeList > 0:
+        normMultiTimeList = [t / sumMultiTimeList * sumMultiVoice for t in multiTimeList]
+    else:
+        normMultiTimeList = multiTimeList
+
+    timeStrings      = su.timeToString(normTimeList)      if normTimeList      else []
+    multiTimeStrings = su.timeToString(normMultiTimeList) if normMultiTimeList else []
     if isinstance(timeStrings, str):
         timeStrings = [timeStrings]
     if isinstance(multiTimeStrings, str):
@@ -202,18 +217,18 @@ def build_df5(oneVoice, multiVoice, sumNoVoice, sumOneVoice, sumMultiVoice, curr
         "labels": ["No Voice", "Single Voice", "Multi Voice"] + speakerList + multiSpeakerList,
         "parents":      ["", "", ""] + ["OV"] * n_ov + ["MV"] * n_mv,
         "parentNames":  ["Total", "Total", "Total"] + ["Single Voice"] * n_ov + ["Multi Voice"] * n_mv,
-        "values":       [sumNoVoice, sumOneVoice, sumMultiVoice] + timeList + multiTimeList,
+        "values":       [sumNoVoice, sumOneVoice, sumMultiVoice] + normTimeList + normMultiTimeList,
         "valueStrings": [
             su.timeToString(sumNoVoice),
             su.timeToString(sumOneVoice),
             su.timeToString(sumMultiVoice),
         ] + timeStrings + multiTimeStrings,
         "percentiles": [b * 100 for b in base]
-            + [(t * 100) / safeOneVoice * base[1] for t in timeList]
-            + [(t * 100) / summativeMulti * base[2] for t in multiTimeList],
+            + [t / safeTotalTime * 100 for t in normTimeList]
+            + [t / safeTotalTime * 100 for t in normMultiTimeList],
         "parentPercentiles": [b * 100 for b in base]
-            + [(t * 100) / safeOneVoice for t in timeList]
-            + [(t * 100) / summativeMulti for t in multiTimeList],
+            + [t / safeOneVoice * 100 for t in normTimeList]
+            + [t / summativeMulti * 100 for t in normMultiTimeList],
     })
 
 
@@ -275,7 +290,7 @@ def build_fig_sunburst(df5, catTypeColors, speakerColors, get_display_name_fn, c
     df5["parentNames"] = df5["parentNames"].apply(lambda s: get_display_name_fn(s, currFile))
     fig = px.sunburst(
         df5,
-        branchvalues="remainder",
+        branchvalues="total",
         names="labels", ids="ids", parents="parents",
         values="percentiles",
         custom_data=["labels", "valueStrings", "percentiles", "parentNames", "parentPercentiles"],
@@ -301,7 +316,7 @@ def build_fig_treemap(df5, catTypeColors, speakerColors, get_display_name_fn, cu
     df5["parentNames"] = df5["parentNames"].apply(lambda s: get_display_name_fn(s, currFile))
     fig = px.treemap(
         df5,
-        branchvalues="remainder",
+        branchvalues="total",
         names="labels", parents="parents", ids="ids",
         values="percentiles",
         custom_data=["labels", "valueStrings", "percentiles", "parentNames", "parentPercentiles"],