Threshold can be a parameter of the PostProcessing.

Author: fnunnari

PostProcessing/PostProcessVideos.py

@@ -14,7 +14,8 @@
 from video import extract_video_info
 
 
-THRESHOLD_NS = 10 * 1000 * 1000  # millis * micros * nanos
+DEFAULT_THRESHOLD_MILLIS = 10
+DEFAULT_THRESHOLD_NANOS = DEFAULT_THRESHOLD_MILLIS * 1000 * 1000  # millis * micros * nanos
 
 
 def scan_session_dir(input_dir: Path) -> Tuple[List[str], List[pd.DataFrame], List[str]]:
@@ -66,14 +67,13 @@ def scan_session_dir(input_dir: Path) -> Tuple[List[str], List[pd.DataFrame], Li
 #
 #
 #
-def main(input_dir: Path, output_dir: Path):
+def main(input_dir: Path, output_dir: Path, threshold_ns: int):
 
     print(f"Scanning dir {str(input_dir)}...")
     clientIDs, df_list, mp4_list = scan_session_dir(input_dir)
 
     n_clients = len(clientIDs)
 
-
     #
     # Print collected info
     for i in range(n_clients):
@@ -97,18 +97,21 @@ def main(input_dir: Path, output_dir: Path):
     # Find time ranges
     min_common, max_common = compute_time_range(repaired_df_list)
     # Trim the data frames to the time range
-    trimmed_dataframes = trim_repaired_into_interval(repaired_df_list, min_common, max_common, THRESHOLD_NS)
+    trimmed_dataframes = trim_repaired_into_interval(repaired_df_list, min_common, max_common, threshold_ns)
 
     assert len(clientIDs) == len(trimmed_dataframes), f"Expected {len(clientIDs)} trimmed dataframes. Found f{len(trimmed_dataframes)}"
 
     # Check that all the resulting dataframes have the same number of rows
+    print("Checking if all clients we have the same number of frames in the repaired amd trimmed tables...")
     client0ID = clientIDs[0]
     client0size = len(trimmed_dataframes[0])
-    print(f"For client {client0ID}: {client0size} frames")
+    print(f"Client {client0ID} has {client0size} frames.")
     for cID, df in zip(clientIDs[1:], trimmed_dataframes[1:]):
         dfsize = len(df)
         if client0size != dfsize:
-            raise Exception(f"For client {cID}: expecting {client0size} frames, found {dfsize}")
+            raise Exception(f"For client {cID}: expecting {client0size} frames. Found {dfsize}."
+                            f" This might be due to an excessive phase offset during recording."
+                            f" Try to increase the threshold.")
 
     print("Good. All trimmed dataframes have the same number of entries.")
 
@@ -152,16 +155,25 @@ def main(input_dir: Path, output_dir: Path):
         "--outfolder", "-o", type=str, help="The folder where the repaired and aligned frames will be stored.",
         required=True
     )
+    parser.add_argument(
+        "--threshold", "-t", type=int, help="The allowed difference in ms between corresponding frames on different videos."
+                                            " Increase this is post processing fails with trimmed tables of different sizes."
+                                            f" Default is now {DEFAULT_THRESHOLD_MILLIS} ms.",
+        required=False,
+        default=DEFAULT_THRESHOLD_MILLIS
+    )
 
     args = parser.parse_args()
 
     infolder = Path(args.infolder)
     outfolder = Path(args.outfolder)
+    threshold_millis = args.threshold
+    threshold_nanos = threshold_millis * 1000 * 1000
 
     if not infolder.exists():
         raise Exception(f"Input folder '{infolder}' doesn't exist.")
 
     if not outfolder.exists():
         raise Exception(f"Output folder '{outfolder}' doesn't exist.")
 
-    main(infolder, outfolder)
+    main(infolder, outfolder, threshold_nanos)

PostProcessing/dataframes.py

@@ -120,4 +120,6 @@ def trim_repaired_into_interval(dfs, min_common, max_common, threshold) -> List[
         trimmed_df = df[selection_mask]
         trimmed_dataframes.append(trimmed_df)
 
+        assert len(trimmed_dataframes) <= len(df)
+
     return trimmed_dataframes

PostProcessing/test_functions.py

@@ -13,7 +13,7 @@
 
 from video import video_info
 
-from PostProcessVideos import THRESHOLD_NS
+from PostProcessVideos import DEFAULT_THRESHOLD_NANOS
 
 
 RECSYNCH_SESSION_DIR_VAR = "RECSYNCH_SESSION_DIR"
@@ -94,7 +94,7 @@ def test_df_reparation(client_data):
 
     tstamps = repaired_df["timestamp"]
     diffs = tstamps.diff().dropna()
-    assert (diffs <= (time_step + THRESHOLD_NS)).all(), "some timestamps difference is longer than expected"
+    assert (diffs <= (time_step + DEFAULT_THRESHOLD_NANOS)).all(), "some timestamps difference is longer than expected"
 
 
 def test_df_trimming(session_data):