Fix anet video downloader (#601)

* add anet video downloader

* pass the test

tools/data/activitynet/README.md

@@ -19,16 +19,16 @@ For action detection, you can either use the ActivityNet rescaled feature provid
 We release both pipeline.
 Before we start, please make sure that current working directory is `$MMACTION2/tools/data/activitynet/`.
 
-## Step 1. Download Annotations
+## Option 1: Use the ActivityNet rescaled feature provided in this [repo](https://github.com/wzmsltw/BSN-boundary-sensitive-network#code-and-data-preparation)
+
+### Step 1. Download Annotations
 
 First of all, you can run the following script to download annotation files.
 
 ```shell
-bash download_annotations.sh
+bash download_feature_annotations.sh
 ```
 
-## Option 1: Use the ActivityNet rescaled feature provided in this [repo](https://github.com/wzmsltw/BSN-boundary-sensitive-network#code-and-data-preparation)
-
 ### Step 2. Prepare Videos Features
 
 Then, you can run the following script to download activitynet features.
@@ -40,24 +40,42 @@ bash download_features.sh
 ### Step 3. Process Annotation Files
 
 Next, you can run the following script to process the downloaded annotation files for training and testing.
-It first merges the two annotation files together and then seperates the annoations by `train`, `val` and `test`.
+It first merges the two annotation files together and then separates the annoations by `train`, `val` and `test`.
 
 ```shell
 python process_annotations.py
 ```
 
-## Option 2: Extract ActivityNet feature using MMAction2
+## Option 2: Extract ActivityNet feature using MMAction2 with all videos provided in official [website](http://activity-net.org/)
+
+### Step 1. Download Annotations
+
+First of all, you can run the following script to download annotation files.
+
+```shell
+bash download_annotations.sh
+```
 
 ### Step 2. Prepare Videos
 
 Then, you can run the following script to prepare videos.
 The codes are adapted from the [official crawler](https://github.com/activitynet/ActivityNet/tree/master/Crawler/Kinetics). Note that this might take a long time.
-Some videos in the ActivityNet dataset might be no longer available on YouTube, so that after video downloading, the downloading scripts update the annotation file to make sure every video in it exists.
 
 ```shell
 bash download_videos.sh
 ```
 
+Since some videos in the ActivityNet dataset might be no longer available on YouTube, official [website](http://activity-net.org/) has made the full dataset available on Google and Baidu drives.
+To accommodate missing data requests, you can fill in this [request form](https://docs.google.com/forms/d/e/1FAIpQLSeKaFq9ZfcmZ7W0B0PbEhfbTHY41GeEgwsa7WobJgGUhn4DTQ/viewform) provided in official [download page](http://activity-net.org/download.html) to have a 7-day-access to download the videos from the drive folders.
+
+We also provide download steps for annotations from [BSN repo](https://github.com/wzmsltw/BSN-boundary-sensitive-network#code-and-data-preparation)
+
+```shell
+bash download_bsn_videos.sh
+```
+
+For this case, the downloading scripts update the annotation file after downloading to make sure every video in it exists.
+
 ### Step 3. Extract RGB and Flow
 
 Before extracting, please refer to [install.md](/docs/install.md) for installing [denseflow](https://github.com/open-mmlab/denseflow).
@@ -119,18 +137,23 @@ mmaction2
 ├── configs
 ├── data
 │   ├── ActivityNet
+
+(if Option 1 used)
 │   │   ├── anet_anno_{train,val,test,full}.json
 │   │   ├── anet_anno_action.json
 │   │   ├── video_info_new.csv
-
-(if Option 1 used)
 │   │   ├── activitynet_feature_cuhk
 │   │   │   ├── csv_mean_100
 │   │   │   │   ├── v___c8enCfzqw.csv
 │   │   │   │   ├── v___dXUJsj3yo.csv
 │   │   │   |   ├── ..
 
 (if Option 2 used)
+│   │   ├── anet_train_video.txt
+│   │   ├── anet_val_video.txt
+│   │   ├── anet_train_clip.txt
+│   │   ├── anet_val_clip.txt
+│   │   ├── activity_net.v1-3.min.json
 │   │   ├── mmaction_feat
 │   │   │   ├── v___c8enCfzqw.csv
 │   │   │   ├── v___dXUJsj3yo.csv
@@ -142,10 +165,7 @@ mmaction2
 │   │   │   │   ├── flow_y_00000.jpg
 │   │   │   │   ├── ..
 │   │   │   ├── ..
-│   │   ├── anet_train_video.txt
-│   │   ├── anet_val_video.txt
-│   │   ├── anet_train_clip.txt
-│   │   ├── anet_val_clip.txt
+
 ```
 
 For training and evaluating on ActivityNet, please refer to [getting_started.md](/docs/getting_started.md).

tools/data/activitynet/download.py

@@ -1,6 +1,7 @@
 # This scripts is copied from
 # https://github.com/activitynet/ActivityNet/blob/master/Crawler/Kinetics/download.py  # noqa: E501
 # The code is licensed under the MIT licence.
+import argparse
 import os
 import ssl
 import subprocess
@@ -10,11 +11,19 @@
 
 ssl._create_default_https_context = ssl._create_unverified_context
 data_file = '../../../data/ActivityNet'
-video_list = f'{data_file}/video_info_new.csv'
-anno_file = f'{data_file}/anet_anno_action.json'
 output_dir = f'{data_file}/videos'
 
 
+def parse_args():
+    parser = argparse.ArgumentParser(description='ActivityNet downloader')
+    parser.add_argument(
+        '--bsn',
+        action='store_true',
+        help='download for BSN annotation or official one')
+    args = parser.parse_args()
+    return args
+
+
 def download_clip(video_identifier,
                   output_filename,
                   num_attempts=5,
@@ -72,7 +81,7 @@ def download_clip_wrapper(youtube_id, output_dir):
     return status
 
 
-def parse_activitynet_annotations(input_csv):
+def parse_activitynet_annotations(input_csv, is_bsn_case=False):
     """Returns a list of YoutubeID.
     arguments:
     ---------
@@ -85,16 +94,21 @@ def parse_activitynet_annotations(input_csv):
         List of all YoutubeIDs in ActivityNet.
 
     """
-    lines = open(input_csv).readlines()
-    lines = lines[1:]
-    # YoutubeIDs do not have prefix `v_`
-    youtube_ids = [x.split(',')[0][2:] for x in lines]
+    if is_bsn_case:
+        lines = open(input_csv).readlines()
+        lines = lines[1:]
+        # YoutubeIDs do not have prefix `v_`
+        youtube_ids = [x.split(',')[0][2:] for x in lines]
+    else:
+        data = mmcv.load(anno_file)['database']
+        youtube_ids = list(data.keys())
+
     return youtube_ids
 
 
-def main(input_csv, output_dir, anno_file, num_jobs=24):
+def main(input_csv, output_dir, anno_file, num_jobs=24, is_bsn_case=False):
     # Reading and parsing ActivityNet.
-    youtube_ids = parse_activitynet_annotations(input_csv)
+    youtube_ids = parse_activitynet_annotations(input_csv, is_bsn_case)
 
     # Creates folders where videos will be saved later.
     if not os.path.exists(output_dir):
@@ -114,10 +128,20 @@ def main(input_csv, output_dir, anno_file, num_jobs=24):
     annotation = mmcv.load(anno_file)
     downloaded = {status[0]: status[1] for status in status_list}
     annotation = {k: v for k, v in annotation.items() if downloaded[k]}
-    anno_file_bak = anno_file.replace('.json', '_bak.json')
-    os.system(f'mv {anno_file} {anno_file_bak}')
-    mmcv.dump(annotation, anno_file)
+
+    if is_bsn_case:
+        anno_file_bak = anno_file.replace('.json', '_bak.json')
+        os.system(f'mv {anno_file} {anno_file_bak}')
+        mmcv.dump(annotation, anno_file)
 
 
 if __name__ == '__main__':
-    main(video_list, output_dir, anno_file, 24)
+    args = parse_args()
+    is_bsn_case = args.bsn
+    if is_bsn_case:
+        video_list = f'{data_file}/video_info_new.csv'
+        anno_file = f'{data_file}/anet_anno_action.json'
+    else:
+        video_list = f'{data_file}/activity_net.v1-3.min.json'
+        anno_file = video_list
+    main(video_list, output_dir, anno_file, 24, is_bsn_case)

tools/data/activitynet/download_annotations.sh

@@ -7,10 +7,6 @@ fi
 
 cd ${DATA_DIR}
 
-wget https://raw.githubusercontent.com/wzmsltw/BSN-boundary-sensitive-network/master/data/activitynet_annotations/anet_anno_action.json
-
-wget https://raw.githubusercontent.com/wzmsltw/BSN-boundary-sensitive-network/master/data/activitynet_annotations/video_info_new.csv
-
-wget https://download.openmmlab.com/mmaction/localization/anet_activity_indexes_val.txt
+wget http://ec2-52-25-205-214.us-west-2.compute.amazonaws.com/files/activity_net.v1-3.min.json
 
 cd -

tools/data/activitynet/download_bsn_videos.sh

@@ -0,0 +1,13 @@
+#!/usr/bin/env bash
+
+# set up environment
+conda env create -f environment.yml
+source activate activitynet
+pip install --upgrade youtube-dl
+pip install mmcv
+
+DATA_DIR="../../../data/ActivityNet"
+python download.py --bsn
+
+source deactivate activitynet
+conda remove -n activitynet --all

tools/data/activitynet/download_feature_annotations.sh

@@ -0,0 +1,16 @@
+DATA_DIR="../../../data/ActivityNet/"
+
+if [[ ! -d "${DATA_DIR}" ]]; then
+  echo "${DATA_DIR} does not exist. Creating";
+  mkdir -p ${DATA_DIR}
+fi
+
+cd ${DATA_DIR}
+
+wget https://raw.githubusercontent.com/wzmsltw/BSN-boundary-sensitive-network/master/data/activitynet_annotations/anet_anno_action.json
+
+wget https://raw.githubusercontent.com/wzmsltw/BSN-boundary-sensitive-network/master/data/activitynet_annotations/video_info_new.csv
+
+wget https://download.openmmlab.com/mmaction/localization/anet_activity_indexes_val.txt
+
+cd -

tools/data/activitynet/generate_rawframes_filelist.py

@@ -1,81 +1,88 @@
+import json
 import os
 import os.path as osp
 
-import mmcv
-
 data_file = '../../../data/ActivityNet'
 video_list = f'{data_file}/video_info_new.csv'
 anno_file = f'{data_file}/anet_anno_action.json'
 rawframe_dir = f'{data_file}/rawframes'
 action_name_list = 'action_name.csv'
 
+train_rawframe_dir = rawframe_dir
+val_rawframe_dir = rawframe_dir
+
+json_file = f'{data_file}/activity_net.v1-3.min.json'
+
 
 def generate_rawframes_filelist():
-    anet_annotations = mmcv.load(anno_file)
-
-    videos = open(video_list).readlines()
-    videos = [x.strip().split(',') for x in videos]
-    attr_names = videos[0][1:]
-    # the first line is 'video,numFrame,seconds,fps,rfps,subset,featureFrame'
-    attr_names = [x.lower() for x in attr_names]
-    attr_types = [int, float, float, float, str, int]
-
-    video_annos = {}
-    for line in videos[1:]:
-        name = line[0]
-        data = {}
-        for attr_name, attr_type, attr_val in zip(attr_names, attr_types,
-                                                  line[1:]):
-            data[attr_name] = attr_type(attr_val)
-        video_annos[name] = data
-
-    # only keep downloaded videos
-    video_annos = {
-        k: v
-        for k, v in video_annos.items() if k in anet_annotations
-    }
-    # update numframe
-    for video in video_annos:
-        pth = osp.join(rawframe_dir, video)
-        num_imgs = len(os.listdir(pth))
-        # one more rgb img than flow
-        assert (num_imgs - 1) % 3 == 0
-        num_frames = (num_imgs - 1) // 3
-        video_annos[video]['numframe'] = num_frames
+    load_dict = json.load(open(json_file))
 
     anet_labels = open(action_name_list).readlines()
     anet_labels = [x.strip() for x in anet_labels[1:]]
 
-    train_videos, val_videos = {}, {}
-    for k, video in video_annos.items():
-        if video['subset'] == 'training':
-            train_videos[k] = video
-        elif video['subset'] == 'validation':
-            val_videos[k] = video
+    train_dir_list = [
+        osp.join(train_rawframe_dir, x) for x in os.listdir(train_rawframe_dir)
+    ]
+    val_dir_list = [
+        osp.join(val_rawframe_dir, x) for x in os.listdir(val_rawframe_dir)
+    ]
 
-    def simple_label(video_idx):
-        anno = anet_annotations[video_idx]
-        label = anno['annotations'][0]['label']
+    def simple_label(anno):
+        label = anno[0]['label']
         return anet_labels.index(label)
 
+    def count_frames(dir_list, video):
+        for dir_name in dir_list:
+            if video in dir_name:
+                return osp.basename(dir_name), len(os.listdir(dir_name))
+        return None, None
+
+    database = load_dict['database']
+    training = {}
+    validation = {}
+    key_dict = {}
+
+    for k in database:
+        data = database[k]
+        subset = data['subset']
+
+        if subset in ['training', 'validation']:
+            annotations = data['annotations']
+            label = simple_label(annotations)
+            if subset == 'training':
+                dir_list = train_dir_list
+                data_dict = training
+            else:
+                dir_list = val_dir_list
+                data_dict = validation
+
+        else:
+            continue
+
+        gt_dir_name, num_frames = count_frames(dir_list, k)
+        if gt_dir_name is None:
+            continue
+        data_dict[gt_dir_name] = [num_frames, label]
+        key_dict[k] = gt_dir_name
+
     train_lines = [
-        k + ' ' + str(train_videos[k]['numframe']) + ' ' +
-        str(simple_label(k)) for k in train_videos
+        k + ' ' + str(training[k][0]) + ' ' + str(training[k][1])
+        for k in training
     ]
     val_lines = [
-        k + ' ' + str(val_videos[k]['numframe']) + ' ' + str(simple_label(k))
-        for k in val_videos
+        k + ' ' + str(validation[k][0]) + ' ' + str(validation[k][1])
+        for k in validation
     ]
 
     with open(osp.join(data_file, 'anet_train_video.txt'), 'w') as fout:
         fout.write('\n'.join(train_lines))
     with open(osp.join(data_file, 'anet_val_video.txt'), 'w') as fout:
         fout.write('\n'.join(val_lines))
 
-    def clip_list(k, anno, vidanno):
-        num_seconds = anno['duration_second']
-        num_frames = vidanno['numframe']
-        fps = num_frames / num_seconds
+    def clip_list(k, anno, video_anno):
+        duration = anno['duration']
+        num_frames = video_anno[0]
+        fps = num_frames / duration
         segs = anno['annotations']
         lines = []
         for seg in segs:
@@ -90,10 +97,10 @@ def clip_list(k, anno, vidanno):
         return lines
 
     train_clips, val_clips = [], []
-    for k in train_videos:
-        train_clips.extend(clip_list(k, anet_annotations[k], train_videos[k]))
-    for k in val_videos:
-        val_clips.extend(clip_list(k, anet_annotations[k], val_videos[k]))
+    for k in training:
+        train_clips.extend(clip_list(k, database[key_dict[k]], training[k]))
+    for k in validation:
+        val_clips.extend(clip_list(k, database[key_dict[k]], validation[k]))
 
     with open(osp.join(data_file, 'anet_train_clip.txt'), 'w') as fout:
         fout.write('\n'.join(train_clips))