Refactor of EgocentricMapper init

src/pupil_labs/egocentric_video_mapper/PL_EgocentricVideoMapper.ipynb

@@ -11,17 +11,9 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 1,
-   "metadata": {},
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "GPU 0: NVIDIA GeForce GTX 1080 Ti (UUID: GPU-72b7ddec-7074-1583-0076-9281c065ae9d)\n"
-     ]
-    }
-   ],
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
    "source": [
     "!nvidia-smi -L"
    ]
@@ -35,7 +27,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 2,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -51,7 +43,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 3,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -67,7 +59,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 4,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -95,7 +87,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 3,
+   "execution_count": 5,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -117,7 +109,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 4,
+   "execution_count": 6,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -136,7 +128,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 5,
+   "execution_count": 7,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -162,46 +154,9 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 6,
+   "execution_count": null,
    "metadata": {},
-   "outputs": [
-    {
-     "name": "stderr",
-     "output_type": "stream",
-     "text": [
-      "Calculating optic flow: 100%|██████████| 1262/1262 [00:45<00:00, 27.85it/s]\n",
-      "Calculating optic flow: 100%|██████████| 1291/1291 [01:04<00:00, 20.09it/s]\n"
-     ]
-    },
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "Estimated offset of alternative egocentric video with respect to Neon scene video: 2.016 seconds (Pearson correlation: 0.9192123717590794)\n"
-     ]
-    },
-    {
-     "name": "stderr",
-     "output_type": "stream",
-     "text": [
-      "Mapping gaze signal: 100%|██████████| 8155/8155 [01:44<00:00, 77.88it/s] "
-     ]
-    },
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "Gaze mapped to alternative camera video saved at /users/sof/test_main/test_note/alternative_camera_gaze.csv\n"
-     ]
-    },
-    {
-     "name": "stderr",
-     "output_type": "stream",
-     "text": [
-      "\n"
-     ]
-    }
-   ],
+   "outputs": [],
    "source": [
     "import pupil_labs.egocentric_video_mapper.__main__ as main\n",
     "from pathlib import Path\n",
@@ -221,9 +176,7 @@
     "\n",
     "args = Args()\n",
     "\n",
-    "main.main(args)\n",
-    "\n",
-    "# Hi, so the Egocentric Video Mapper Alpha Lab is ready for internal testing, I created a Google Colab for use the tool"
+    "main.main(args)"
    ]
   }
  ],

src/pupil_labs/egocentric_video_mapper/gaze_mapper.py

@@ -13,10 +13,8 @@
 class EgocentricMapper:
     def __init__(
         self,
-        neon_gaze_csv,
-        neon_video_path,
+        neon_timeseries_dir,
         alternative_video_path,
-        neon_timestamps,
         alternative_timestamps,
         image_matcher,
         image_matcher_parameters,
@@ -30,10 +28,8 @@ def __init__(
         """Class to map gaze from the Neon scene camera to an alternative camera. The gaze is mapped by calling the map_gaze method which uses correspondences between the two cameras to calculate the gaze transformation.
 
         Args:
-            neon_gaze_csv (str): Path to the gaze.csv file from the Neon recording. The file should come from the Timeseries Data + Scene Video download from Pupil Cloud.
-            neon_video_path (str): Path to the video recorded by the Neon scene camera. The video comes from the Timeseries Data + Scene Video download from Pupil Cloud.
+            neon_timeseries_dir (str): Path to the directory containing the Neon scene video, the world_timestamps.csv file and the gaze.csv file.
             alternative_video_path (str): Path to the video recorded by the alternative camera.
-            neon_timestamps (str): Path to the world_timestamps.csv from the Neon recording. The file should come from the Timeseries Data + Scene Video download from Pupil Cloud.
             alternative_timestamps (str): Path to the alternative_camera_timestamps.csv from the alternative camera. This file is create by utils.write_timestamp_csv
             image_matcher (str): Name of the image matcher to be used. For already implemented matcher check the feature_matcher module.
             image_matcher_parameters (dict): Set of specific parameters for the image matcher. The parameters are passed as a dictionary with the parameter name as the key and the parameter value as the value.
@@ -44,10 +40,14 @@ def __init__(
             alternative_fov (list, optional): Field of view in degrees of the Neon Scene camera (can be 2D or 1D). If 1D, it is assumed that the camera has the same field of view in both axes. If 2D, it is assumed that [fov_x, fov_y]. Defaults to [145, 76].
             logging_level (str, optional): Level of logging to be used. Defaults to "ERROR".
         """
+        # Video file name in the Time Series + Video Scene
+        neon_video_path = next(Path(neon_timeseries_dir).rglob("*.mp4"))
         self.neon_video = VideoHandler(neon_video_path)
         self.alt_video = VideoHandler(alternative_video_path)
 
-        self.neon_vid_ts_nanosec = pd.read_csv(neon_timestamps)
+        self.neon_vid_ts_nanosec = pd.read_csv(
+            Path(neon_timeseries_dir, "world_timestamps.csv")
+        )
         self.alt_vid_ts_nanosec = pd.read_csv(alternative_timestamps)
         self.alt2neon_offset_sec = (
             self.alt_vid_ts_nanosec["timestamp [ns]"].values[0]
@@ -59,7 +59,7 @@ def __init__(
         self.alt_opticflow = pd.read_csv(alternative_opticflow_csv, dtype=np.float32)
         self.alt_opticflow = self.alt_opticflow.iloc[1:]
 
-        self.neon_gaze = pd.read_csv(neon_gaze_csv)  # @ 200Hz
+        self.neon_gaze = pd.read_csv(Path(neon_timeseries_dir, "gaze.csv"))  # @ 200Hz
         self.alt_gaze = self._create_alternative_gaze_df()
 
         self.image_matcher = get_matcher(image_matcher, image_matcher_parameters)

src/pupil_labs/egocentric_video_mapper/utils.py

@@ -125,8 +125,6 @@ def generate_mapper_kwargs(
         "disk_lightglue": {"num_features": 2048, "gpu_num": 0},
         "dedode_lightglue": {"num_features": 5000, "gpu_num": 0},
     }
-    # Video file name in the Time Series + Video Scene
-    neon_vid_path = next(Path(neon_timeseries_dir).rglob("*.mp4"))
 
     alternative_timestamps_path = Path(output_dir, "alternative_camera_timestamps.csv")
     if not alternative_timestamps_path.exists():
@@ -140,15 +138,13 @@ def generate_mapper_kwargs(
         )
 
     mapper_kwargs = {
-        "neon_gaze_csv": Path(neon_timeseries_dir, "gaze.csv"),
-        "neon_video_path": neon_vid_path,
-        "neon_timestamps": Path(neon_timeseries_dir, "world_timestamps.csv"),
-        "neon_opticflow_csv": Path(output_dir, f"neon_optic_flow.csv"),
+        "neon_timeseries_dir": Path(neon_timeseries_dir),
         "alternative_video_path": alternative_vid_path,
         "alternative_timestamps": alternative_timestamps_path,
-        "alternative_opticflow_csv": Path(output_dir, f"alternative_optic_flow.csv"),
         "image_matcher": matcher_choice,
         "image_matcher_parameters": image_matcher_parameters[matcher_choice],
+        "neon_opticflow_csv": Path(output_dir, f"neon_optic_flow.csv"),
+        "alternative_opticflow_csv": Path(output_dir, f"alternative_optic_flow.csv"),
         "output_dir": Path(output_dir),
         "patch_size": 1000,
         "logging_level": logging_level,