fix: Add validation for heatmap data loading in trace analysis scripts

* Raise ValueError if no data is available for reuse and size heatmaps.
* Ensure at least two objects are present for popularity data to prevent errors during analysis.
* Adjusted rotation of x-ticks in reuse plot for better readability.

Author: 1a1a11a

scripts/traceAnalysis/popularity.py

@@ -43,6 +43,11 @@ def load_popularity_data(datapath):
 
     ifile.close()
 
+    if len(sorted_freq) < 2:
+        raise ValueError(
+            f"insufficient popularity data in {datapath} "
+            "(need at least 2 objects, trace may have too few)"
+        )
     return sorted_freq, freq_cnt
 
 

scripts/traceAnalysis/reuse.py

@@ -127,7 +127,7 @@ def plot_reuse(datapath: str, figname_prefix: str = "") -> None:
     plt.xticks(
         np.array([60, 300, 3600, 86400, 86400 * 2, 86400 * 4]) / 3600,
         ["1 min", "5 min", "1 hour", "1 day", "", "4 day"],
-        rotation=28,
+        rotation=90,
     )
     plt.savefig(
         "{}/{}_reuse_rt_log.{}".format(FIG_DIR, figname_prefix, FIG_TYPE),

scripts/traceAnalysis/reuse_heatmap.py

@@ -90,7 +90,11 @@ def _load_reuse_heatmap_data(datapath: str) -> Tuple[np.ndarray, int, int, float
 
     ifile.close()
 
-    dim = max([len(l) for l in reuse_time_distribution_list])
+    if not reuse_time_distribution_list:
+        raise ValueError(
+            f"no reuse heatmap data in {datapath} (trace may have time span 0)"
+        )
+    dim = max(len(l) for l in reuse_time_distribution_list)
     plot_data = np.ones((len(reuse_time_distribution_list), dim))
     for idx, l in enumerate(reuse_time_distribution_list):
         plot_data[idx][: len(l)] = np.cumsum(np.array(l) / sum(l))

scripts/traceAnalysis/size_heatmap.py

@@ -68,7 +68,11 @@ def _load_size_heatmap_data(datapath) -> Tuple[np.ndarray, int, float, int]:
 
     ifile.close()
 
-    dim = max([len(l) for l in size_distribution_over_time])
+    if not size_distribution_over_time:
+        raise ValueError(
+            f"no size heatmap data in {datapath} (trace may have time span 0)"
+        )
+    dim = max(len(l) for l in size_distribution_over_time)
     plot_data = np.zeros((len(size_distribution_over_time), dim))
 
     for idx, l in enumerate(size_distribution_over_time):