[RTADPA BW Tests] New Sprucing bandwidth test using latest-available HLT2 input.

cc: @rjhunter.

related to: lhcb/Moore!2763 (merged) + lhcb-core/LHCbPR2HD!270 (merged) + lhcb-core/LHCbNightlyConf!1121 (merged)

Changes:

Adds a new sprucing test, Moore_spruce_latest_bandwidth, which uses the Hlt2 nominal full-stream output which was uploaded to eos on the previous day.

Added a new script, generate_hlt2_fullstream_metadata.py, which runs only for Hlt2 nominal tests and writes the metadata.yaml based on information from the full-stream output of the test currently running. This metadata.yaml is used as an input config for the new sprucing test.

• a little safety around division in zero from the calculation of overlap_matrix, never comes up in the actual tests because every stream has a non-zero retention in 100k events, but when running test jobs it can come up and it was annoying me.

• changes to facilitate the new test + metadata.

ToDo:

• Maybe revisit the N_EVTS and calculate via a gaudipython job?
  • Wait to see if BWTesters know a way to tell the N_EVTS available in a file.
  • Else, adapt list_event_numbers to extract the N_EVTS by iterating a gaudipython job.
• Rebase after !362 (merged) was merged

Closes #12 (closed)

python/MooreTests/calculate_stream_overlap.py

@@ -67,8 +67,8 @@ def calculate_overlap_matrix(df):
     for target_stream in df.columns:
         for comparison_stream in df.columns:
             cond_prob_per_stream[target_stream].append(
-                sum(df[comparison_stream] * df[target_stream]) / sum(
-                    df[target_stream]))
+                sum(df[comparison_stream] * df[target_stream]) /
+                sum(df[target_stream]) if sum(df[target_stream]) > 0 else 0)
     overlap_matrix_df = pd.DataFrame(
         cond_prob_per_stream, columns=df.columns, index=df.columns)
     return overlap_matrix_df