Set start_time to oldest event time in dataframe file if file exists

projects/online/online/monitor/main.py

@@ -5,9 +5,6 @@
 from .utils.parse_logs import estimate_tb, pipeline_online
 from .pages import EventPage, SummaryPage
 
-from gwpy.time import tconvert
-from datetime import datetime, timezone
-
 
 def main(
     run_dir: Path,
@@ -25,7 +22,9 @@ def main(
         online_args: Configuration parameters for the online search.
         start_time:
             The earliest GPS time to consider for processing events.
-            If None, the current GPS time is used.
+            If None and an event dataframe file exists, the oldest
+            event time will be used. If the file does not exist,
+            the current GPS time is used.
         update_cadence:
             The interval in seconds to check for new events.
         logger: Logger object for standardizing logging output
@@ -36,26 +35,23 @@ def main(
 
     detected_event_dir = run_dir / "output" / "events"
 
-    if start_time is None:
-        start_time = float(tconvert(datetime.now(timezone.utc)))
-
+    summary_page = SummaryPage(start_time, run_dir, out_dir, logger)
     # Estimate analysis live time since the given start time
-    tb = estimate_tb(run_dir, start_time)
+    tb = estimate_tb(run_dir, summary_page.start_time)
     logger.info(f"Estimated analysis live time: {tb:.2f} seconds")
 
     previous_update_time = time.time()
 
-    summary_page = SummaryPage(start_time, run_dir, out_dir, logger)
     while True:
         detected_events = [
             event
             for event in detected_event_dir.iterdir()
-            if float(event.name.split("_")[1]) >= start_time
+            if float(event.name.split("_")[1])
         ]
 
         # The event page will be created/updated only if the event directory
         # is missing expected plots. Otherwise it will be skipped.
-        for event in detected_events:
+        for event in sorted(detected_events):
             event_page = EventPage(
                 event, online_args, run_dir, out_dir, logger
             )

projects/online/online/monitor/pages/summary.py

@@ -21,6 +21,20 @@ def __init__(
     ) -> None:
         super().__init__(*args, **kwargs)
         self.start_time = start_time
+        if self.start_time is None:
+            if self.dataframe_file.exists():
+                df = pd.read_hdf(self.dataframe_file)
+                self.start_time = min(df["gpstime"])
+                self.logger.info(
+                    "start_time was None, setting start_time to the oldest "
+                    f"event time in {self.dataframe_file}, {self.start_time}"
+                )
+            else:
+                self.start_time = float(tconvert(datetime.now(timezone.utc)))
+                self.logger.info(
+                    "start_time was None, setting start_time to current time"
+                )
+
         self.plots_dir = self.out_dir / "plots"
         if not self.plots_dir.exists():
             self.plots_dir.mkdir(exist_ok=True, parents=True)
@@ -83,6 +97,12 @@ def update_summary_plots(self, tb: float):
             tb: Total time in seconds that the pipeline has been running.
         """
         df = pd.read_hdf(self.dataframe_file)
+        df = df[df["gpstime"] >= self.start_time]
+        if df.empty:
+            self.logger.warning(
+                "No events found in the DataFrame after the start time."
+            )
+            return
         latency_plot(self.plots_dir, df)
         ifar_plot(self.plots_dir, df, tb)
         event_rate_plots(self.plots_dir, df)

scripts/aframe_init.py

@@ -110,7 +110,8 @@ def create_online_runfile(path: Path):
     monitor_cmd = "apptainer run "
     monitor_cmd += f" --bind {path} "
     monitor_cmd += "$AFRAME_CONTAINER_ROOT/online.sif /opt/env/bin/monitor "
-    monitor_cmd += f"--run_dir {path} --out_dir $MONITOR_OUTDIR &"
+    monitor_cmd += f"--run_dir {path} --out_dir $MONITOR_OUTDIR "
+    monitor_cmd += ">> summary_pages.log 2>&1 &"
 
     content = f"""
     #!/bin/bash