Added movieCLIP taxonomy details

data/shot_test_samples.txt

@@ -0,0 +1,9 @@
+/data/ambience/Condensed_Movies/Condensed_Movies_downloaded_data/2011/_2LMj-4PtdU.mkv
+/data/ambience/Condensed_Movies/Condensed_Movies_downloaded_data/2012/_6AGOfGHzeg.mkv
+/data/ambience/Condensed_Movies/Condensed_Movies_downloaded_data/2013/_8N1uJZUyaY.mkv
+/data/ambience/Condensed_Movies/Condensed_Movies_downloaded_data/2014/_epn5foR_Ts.mkv
+/data/ambience/Condensed_Movies/Condensed_Movies_downloaded_data/2015/_ItWcGtaJro.mkv
+/data/ambience/Condensed_Movies/Condensed_Movies_downloaded_data/2016/_emU23tTUAw.mkv
+/data/ambience/Condensed_Movies/Condensed_Movies_downloaded_data/2017/_6RI-8Ia4do.mkv
+/data/ambience/Condensed_Movies/Condensed_Movies_downloaded_data/2018/_HcDe70cSRU.mkv
+/data/ambience/Condensed_Movies/Condensed_Movies_downloaded_data/2019/_f7p28YFgvc.mkv

preprocess_scripts/check_shot_segment_files_overlap.py

@@ -0,0 +1,101 @@
+#helper scripts to check if the shot segment files overlap in two directories
+
+import os
+from tqdm import tqdm
+
+def generate_file_name(file_path):
+
+    #print(file_path)
+    scenes_index=file_path.index("Scenes")
+    file_name=file_path[0:scenes_index-1]
+    return(file_name)
+
+def generate_cmd_clip_list(folder_list,base_folder):
+
+    #print(folder_list)
+    cmd_clip_list=[]
+    for folder in tqdm(folder_list):
+        folder_path=os.path.join(base_folder,folder)
+        file_list=os.listdir(folder_path)
+        file_list=[os.path.join(folder_path,s) for s in file_list]
+        cmd_clip_list=cmd_clip_list+file_list
+
+    return(cmd_clip_list)
+
+
+shot_segment_file_v1="/data/digbose92/ambience_detection/codes/shot_segments/shot_segments_v1"
+shot_segment_file_v2="/data/digbose92/ambience_detection/codes/shot_segments/shot_segments_v2"
+folder="/data/ambience/Condensed_Movies/video_clips_shots_complete"
+shot_subfolder=os.listdir(folder)
+print(len(set(shot_subfolder)))
+
+#scene file list v1 and v2
+scene_file_list_v1=os.listdir(shot_segment_file_v1)
+scene_file_list_v1=[s for s in scene_file_list_v1 if s.endswith(".csv")]
+scene_file_list_v2=os.listdir(shot_segment_file_v2)
+
+#total scene file list
+total_scene_file_list=scene_file_list_v1+scene_file_list_v2
+total_scene_file_list.remove("Nan_label_top_250.csv")
+print(len(total_scene_file_list),len(set(total_scene_file_list)))
+
+#number of total scene files
+total_files=32484
+
+print('Total number of scene files in v1+v2: ',len(scene_file_list_v1)+len(scene_file_list_v2))
+print('Remaining files: ',total_files-len(scene_file_list_v1)-len(scene_file_list_v2))
+
+#check intersection between two lists
+intersection_list=list(set(scene_file_list_v1).intersection(scene_file_list_v2))
+print(len(intersection_list)) #currently zero
+
+#read the list of mkv files in the Condensed movies directory
+CMD_clip_file="/data/ambience/Condensed_Movies/Condensed_Movies_downloaded_data/clip_list.txt"
+with open(CMD_clip_file,'r') as f:
+    CMD_clip_list=f.readlines()
+
+CMD_clip_list_sample=[c.split("\n")[0].split("/")[-1] for c in CMD_clip_list]
+#print(CMD_clip_list[0:5])
+
+cnt_present_folder=0 #should be 28613
+cnt_mkv_files=0 #should be 28613
+
+subfold_present_list=[]
+for scene_file in tqdm(total_scene_file_list):
+
+    subfolder_name=generate_file_name(scene_file)
+    subfold_present_list.append(subfolder_name)
+    index_subfold=shot_subfolder.index(subfolder_name)
+    cnt_present_folder+=1
+    # except:
+    #     print('Here')
+    #     mkv_filename=subfolder_name+".mkv"
+    #     print(mkv_filename)
+    #     if mkv_filename in CMD_clip_list:
+    #         cnt_mkv_files+=1
+difference_folder=list(set(shot_subfolder)-set(subfold_present_list))
+filename_incomplete_list=[]
+for diff_fold in difference_folder:
+    mkv_filename=diff_fold+".mkv"
+    if mkv_filename in CMD_clip_list_sample:
+        cnt_mkv_files+=1
+        filename_incomplete_list.append(CMD_clip_list[CMD_clip_list_sample.index(mkv_filename)].split("\n")[0])
+    # else:
+    #     print(mkv_filename)
+
+# print(cnt_present_folder)
+print(cnt_mkv_files)
+print(len(difference_folder)-cnt_mkv_files)
+print(filename_incomplete_list)
+
+
+with open('../data/shots_rerun_incomplete_list.txt','w') as f:
+    for item in filename_incomplete_list:
+        f.write("%s \n" % item)
+
+
+# base_folder="/data/ambience/Condensed_Movies/Condensed_Movies_downloaded_data"
+# folder_list=['2011','2012','2013','2014','2015','2016','2017','2018','2019']
+# cmd_clip_list=generate_cmd_clip_list(folder_list,base_folder)#32333
+
+#print(len(cmd_clip_list))

preprocess_scripts/duration_txt_files/test_durs.txt

@@ -0,0 +1,52 @@
+]633;C]633;E;for fl in `find /data/ambience/Condensed_Movies/test_shots_data/_8LrZ4NhPmk/ -iname *mp4`  Duration: 00:00:03.03
+  Duration: 00:00:01.36
+  Duration: 00:00:01.86
+  Duration: 00:00:00.61
+  Duration: 00:00:00.65
+  Duration: 00:00:05.82
+  Duration: 00:00:03.69
+  Duration: 00:00:01.15
+  Duration: 00:00:07.26
+  Duration: 00:00:02.02
+  Duration: 00:00:02.07
+  Duration: 00:00:01.40
+  Duration: 00:00:01.94
+  Duration: 00:00:01.86
+  Duration: 00:00:02.61
+  Duration: 00:00:04.15
+  Duration: 00:00:02.23
+  Duration: 00:00:05.19
+  Duration: 00:00:02.07
+  Duration: 00:00:02.19
+  Duration: 00:00:01.77
+  Duration: 00:00:01.07
+  Duration: 00:00:01.61
+  Duration: 00:00:02.23
+  Duration: 00:00:01.98
+  Duration: 00:00:01.69
+  Duration: 00:00:02.44
+  Duration: 00:00:03.07
+  Duration: 00:00:02.19
+  Duration: 00:00:03.32
+  Duration: 00:00:02.44
+  Duration: 00:00:03.19
+  Duration: 00:00:02.57
+  Duration: 00:00:01.82
+  Duration: 00:00:04.53
+  Duration: 00:00:01.40
+  Duration: 00:00:02.19
+  Duration: 00:00:04.57
+  Duration: 00:00:01.44
+  Duration: 00:00:01.86
+  Duration: 00:00:00.98
+  Duration: 00:00:04.40
+  Duration: 00:00:01.61
+  Duration: 00:00:01.77
+  Duration: 00:00:02.02
+  Duration: 00:00:01.61
+  Duration: 00:00:04.36
+  Duration: 00:00:01.52
+  Duration: 00:00:01.11
+  Duration: 00:00:01.48
+  Duration: 00:00:01.19
+  Duration: 00:00:02.61

preprocess_scripts/duration_txt_files/test_durs_orig.txt

@@ -0,0 +1,52 @@
+]633;C]633;E;for fl in `find /data/ambience/Condensed_Movies/video_clips_shots_complete/_8LrZ4NhPmk/ -iname *mp4`  Duration: 00:00:03.03
+  Duration: 00:00:01.36
+  Duration: 00:00:01.86
+  Duration: 00:00:00.61
+  Duration: 00:00:00.65
+  Duration: 00:00:05.82
+  Duration: 00:00:03.69
+  Duration: 00:00:01.15
+  Duration: 00:00:07.26
+  Duration: 00:00:02.02
+  Duration: 00:00:02.07
+  Duration: 00:00:01.40
+  Duration: 00:00:01.94
+  Duration: 00:00:01.86
+  Duration: 00:00:02.61
+  Duration: 00:00:04.15
+  Duration: 00:00:02.23
+  Duration: 00:00:05.19
+  Duration: 00:00:02.07
+  Duration: 00:00:02.19
+  Duration: 00:00:01.77
+  Duration: 00:00:01.07
+  Duration: 00:00:01.61
+  Duration: 00:00:02.23
+  Duration: 00:00:01.98
+  Duration: 00:00:01.69
+  Duration: 00:00:02.44
+  Duration: 00:00:03.07
+  Duration: 00:00:02.19
+  Duration: 00:00:03.32
+  Duration: 00:00:02.44
+  Duration: 00:00:03.19
+  Duration: 00:00:02.57
+  Duration: 00:00:01.82
+  Duration: 00:00:04.53
+  Duration: 00:00:01.40
+  Duration: 00:00:02.19
+  Duration: 00:00:04.57
+  Duration: 00:00:01.44
+  Duration: 00:00:01.86
+  Duration: 00:00:00.98
+  Duration: 00:00:04.40
+  Duration: 00:00:01.61
+  Duration: 00:00:01.77
+  Duration: 00:00:02.02
+  Duration: 00:00:01.61
+  Duration: 00:00:04.36
+  Duration: 00:00:01.52
+  Duration: 00:00:01.11
+  Duration: 00:00:01.48
+  Duration: 00:00:01.19
+  Duration: 00:00:02.61

preprocess_scripts/duration_txt_files/test_durs_original_ItWcGtaJro.txt

@@ -0,0 +1,75 @@
+]633;C]633;E;for fl in `find /data/ambience/Condensed_Movies/video_clips_shots_complete/_ItWcGtaJro -iname *mp4`  Duration: 00:00:01.61
+  Duration: 00:00:01.15
+  Duration: 00:00:02.44
+  Duration: 00:00:02.23
+  Duration: 00:00:02.73
+  Duration: 00:00:01.36
+  Duration: 00:00:04.36
+  Duration: 00:00:03.19
+  Duration: 00:00:01.69
+  Duration: 00:00:01.27
+  Duration: 00:00:01.73
+  Duration: 00:00:01.94
+  Duration: 00:00:03.86
+  Duration: 00:00:03.65
+  Duration: 00:00:01.73
+  Duration: 00:00:02.19
+  Duration: 00:00:02.02
+  Duration: 00:00:01.69
+  Duration: 00:00:02.11
+  Duration: 00:00:02.15
+  Duration: 00:00:00.67
+  Duration: 00:00:03.78
+  Duration: 00:00:03.19
+  Duration: 00:00:02.65
+  Duration: 00:00:02.02
+  Duration: 00:00:01.73
+  Duration: 00:00:01.02
+  Duration: 00:00:02.19
+  Duration: 00:00:01.73
+  Duration: 00:00:00.98
+  Duration: 00:00:03.44
+  Duration: 00:00:04.49
+  Duration: 00:00:02.48
+  Duration: 00:00:01.48
+  Duration: 00:00:01.23
+  Duration: 00:00:02.69
+  Duration: 00:00:03.94
+  Duration: 00:00:01.44
+  Duration: 00:00:01.86
+  Duration: 00:00:01.27
+  Duration: 00:00:01.48
+  Duration: 00:00:01.57
+  Duration: 00:00:01.73
+  Duration: 00:00:01.65
+  Duration: 00:00:01.94
+  Duration: 00:00:01.77
+  Duration: 00:00:03.07
+  Duration: 00:00:01.23
+  Duration: 00:00:01.61
+  Duration: 00:00:01.36
+  Duration: 00:00:02.32
+  Duration: 00:00:01.57
+  Duration: 00:00:02.36
+  Duration: 00:00:00.94
+  Duration: 00:00:01.19
+  Duration: 00:00:01.36
+  Duration: 00:00:01.36
+  Duration: 00:00:01.44
+  Duration: 00:00:02.02
+  Duration: 00:00:02.78
+  Duration: 00:00:02.11
+  Duration: 00:00:01.02
+  Duration: 00:00:01.44
+  Duration: 00:00:01.32
+  Duration: 00:00:03.03
+  Duration: 00:00:01.44
+  Duration: 00:00:01.23
+  Duration: 00:00:01.23
+  Duration: 00:00:03.23
+  Duration: 00:00:01.11
+  Duration: 00:00:01.11
+  Duration: 00:00:01.61
+  Duration: 00:00:01.57
+  Duration: 00:00:03.61
+  Duration: 00:00:01.52

preprocess_scripts/extract_scenes_condensed_movies_clips.py

@@ -0,0 +1,65 @@
+import os 
+import pandas as pd 
+import numpy as np 
+from tqdm import tqdm 
+import pickle
+import multiprocessing as mp
+import argparse 
+
+destination_scenes_folder='/data/ambience/Condensed_Movies/video_clips_shots_complete'
+csv_scenes_folder="/data/ambience/Condensed_Movies/video_clips_shots_stats_complete"
+file_list_pickle_file="/data/digbose92/ambience_detection/pkl-files/Condensed_Movies_updated_list_large_set.pkl"
+
+with open(file_list_pickle_file,"rb") as f:
+    file_list=pickle.load(f)
+
+# def extract_scene_clips(idx):
+#     vid_file=file_list[idx]
+#     file_key=vid_file.split("/")[-1][:-4]
+#     subfolder=os.path.join(destination_scenes_folder,file_key)
+#     csv_scenes_file=os.path.join(csv_scenes_folder,file_key+".csv")
+
+#     if(os.path.exists(csv_scenes_file) is False):
+#         os.mkdir(subfolder)
+#         scene_detect_command="scenedetect --input "+vid_file+ " -s "+csv_scenes_file+" detect-content list-scenes split-video -o "+subfolder
+#         os.system(scene_detect_command)
+def extract_scene_clips(idx):
+    vid_file=file_list[idx]
+    file_key=vid_file.split("/")[-1][:-4]
+    subfolder=os.path.join(destination_scenes_folder,file_key)
+    csv_scenes_file=os.path.join(csv_scenes_folder,file_key+".csv")
+
+    if(os.path.exists(subfolder) is False):
+        os.mkdir(subfolder)
+        scene_detect_command="scenedetect --input "+vid_file+ " -s "+csv_scenes_file+" detect-content list-scenes split-video -o "+subfolder
+        os.system(scene_detect_command)
+
+#condensed_movies_folder='/data/ambience/Condensed_Movies/video_clips_downsampled'
+
+
+# #print(len(condensed_movies_folder))
+# for vid_file in tqdm(file_list):
+#     file_key=vid_file.split("/")[-1][:-4]
+#     subfolder=os.path.join(destination_scenes_folder,file_key)
+#     csv_scenes_file=os.path.join(csv_scenes_folder,file_key+".csv")
+
+#     if(os.path.exists(csv_scenes_file) is False):
+#         os.mkdir(subfolder)
+#         scene_detect_command="scenedetect --input "+vid_file+ " -s "+csv_scenes_file+" detect-content list-scenes split-video -o "+subfolder
+#         os.system(scene_detect_command)
+        #print(scene_detect_command)
+    #scene_det
+    #print(subfolder)
+    #print(file_key)
+def main(args):
+#    data = [x.rstrip().split(',') for x in open('/data/movies/movie_sounds_50_mturk_test.csv').readlines()[1:]]
+#   data = [x.rstrip().split(',') for x in open('/data/rajatheb/sound_events/isound_event_labels.csv').readlines()[1:]]
+    pool = mp.Pool(args.nj)
+    pool.map(extract_scene_clips, list(range(len(file_list))))
+    pool.close()
+
+if __name__ == '__main__':
+    parser = argparse.ArgumentParser(description='')
+    parser.add_argument('--nj', default=16, type=int, help='number of parallel processes')
+    args = parser.parse_args()
+    main(args)

preprocess_scripts/test_shot_detection.py

@@ -0,0 +1,13 @@
+import os 
+import pandas as pd 
+import numpy as np 
+from tqdm import tqdm 
+import pickle
+
+destination_folder="/data/ambience/Condensed_Movies/test_shots_data"
+video_file="/data/ambience/Condensed_Movies/Condensed_Movies_downloaded_data/2012/_8LrZ4NhPmk.mkv"
+subfolder=os.path.join(destination_folder,os.path.splitext(video_file.split("/")[-1])[0])
+csv_file="test.csv"
+scene_detect_command="scenedetect --input "+video_file+ " -s "+csv_file+" detect-content list-scenes split-video -o "+subfolder
+
+os.system(scene_detect_command)