Merge pull request #136 from Janelia-Trackathon-2023/framebuffer

Change `frame_buffer` kwarg to `max_frame_buffer`

src/traccuracy/cli.py

@@ -187,7 +187,7 @@ def run_divisions_on_iou(
         gt_data,
         pred_data,
         IOUMatcher(iou_threshold=match_threshold),
-        [DivisionMetrics(frame_buffer=frame_buffer_tuple)],
+        [DivisionMetrics(max_frame_buffer=frame_buffer_tuple)],
     )
     with open(out_path, "w") as fp:
         json.dump(result, fp)
@@ -242,7 +242,7 @@ def run_divisions_on_ctc(
         gt_data,
         pred_data,
         CTCMatcher(),
-        [DivisionMetrics(frame_buffer=frame_buffer_tuple)],
+        [DivisionMetrics(max_frame_buffer=frame_buffer_tuple)],
     )
     with open(out_path, "w") as fp:
         json.dump(result, fp)

src/traccuracy/metrics/_divisions.py

@@ -57,14 +57,15 @@ class DivisionMetrics(Metric):
         3, 734559 (2021).
 
     Args:
-        frame_buffer (tuple(int), optional): Tuple of integers. Value used as n_frames
-            to tolerate in correct_shifted_divisions. Defaults to (0).
+        max_frame_buffer (int, optional): Maximum value of frame buffer to use in correcting
+            shifted divisions. Divisions will be evaluated for all integer values of frame
+            buffer between 0 and max_frame_buffer
     """
 
     needs_one_to_one = True
 
-    def __init__(self, frame_buffer=(0,)):
-        self.frame_buffer = frame_buffer
+    def __init__(self, max_frame_buffer=0):
+        self.frame_buffer = max_frame_buffer
 
     def _compute(self, data: Matched):
         """Runs `_evaluate_division_events` and calculates summary metrics for each frame buffer
@@ -78,7 +79,7 @@ def _compute(self, data: Matched):
         """
         div_annotations = _evaluate_division_events(
             data,
-            frame_buffer=self.frame_buffer,
+            max_frame_buffer=self.frame_buffer,
         )
 
         return {

src/traccuracy/track_errors/divisions.py

@@ -239,7 +239,7 @@ def _correct_shifted_divisions(matched_data: Matched, n_frames=1):
     return new_matched
 
 
-def _evaluate_division_events(matched_data: Matched, frame_buffer=(0)):
+def _evaluate_division_events(matched_data: Matched, max_frame_buffer=0):
     """Classify division errors and correct shifted divisions according to frame_buffer
 
     Note: A copy of matched_data will be created for each frame_buffer other than 0.
@@ -248,8 +248,9 @@ def _evaluate_division_events(matched_data: Matched, frame_buffer=(0)):
     Args:
         matched_data (Matched): Matched data object containing gt and pred graphs
             with their associated mapping
-        frame_buffer (tuple, optional): Tuple of integers. Value used as n_frames
-            to tolerate in correct_shifted_divisions. Defaults to (0).
+        max_frame_buffer (int, optional): Maximum value of frame buffer to use in correcting
+            shifted divisions. Divisions will be evaluated for all integer values of frame
+            buffer between 0 and max_frame_buffer
 
     Returns:
         dict {frame_buffer: matched_data}: A dictionary where each key corresponds to a frame
@@ -263,11 +264,7 @@ def _evaluate_division_events(matched_data: Matched, frame_buffer=(0)):
     div_annotations[0] = matched_data
 
     # Correct shifted divisions for each nonzero value in frame_buffer
-    for delta in frame_buffer:
-        # Skip 0 because we used that in baseline classification
-        if delta == 0:
-            continue
-
+    for delta in range(1, max_frame_buffer + 1):
         corrected_matched = _correct_shifted_divisions(matched_data, n_frames=delta)
         div_annotations[delta] = corrected_matched
 

tests/metrics/test_divisions.py

@@ -12,13 +12,13 @@ def test_DivisionMetrics():
         TrackingGraph(g_pred),
         mapper,
     )
-    frame_buffer = (0, 1, 2)
+    frame_buffer = 2
 
-    results = DivisionMetrics(frame_buffer=frame_buffer)._compute(matched)
+    results = DivisionMetrics(max_frame_buffer=frame_buffer)._compute(matched)
 
     for name, r in results.items():
         buffer = int(name[-1:])
-        assert buffer in frame_buffer
+        assert buffer in list(range(frame_buffer + 1))
         if buffer in (0, 1):
             # No corrections
             assert r["True Positive Divisions"] == 0

tests/track_errors/test_divisions.py

@@ -213,10 +213,10 @@ def test_fp_early(self):
 
 def test_evaluate_division_events():
     g_gt, g_pred, mapper = get_division_graphs()
-    frame_buffer = (0, 1, 2)
+    frame_buffer = 2
 
     matched_data = Matched(TrackingGraph(g_gt), TrackingGraph(g_pred), mapper)
 
-    results = _evaluate_division_events(matched_data, frame_buffer=frame_buffer)
+    results = _evaluate_division_events(matched_data, max_frame_buffer=frame_buffer)
 
     assert np.all([isinstance(k, int) for k in results.keys()])