Skip to content

Latest commit

 

History

History
86 lines (69 loc) · 3.47 KB

TVNEWS_VIEWER_README.md

File metadata and controls

86 lines (69 loc) · 3.47 KB
Raw

Loading up the TV news viewer

Assuming everything went well and the metadata has been extracted, we will now load the data into a form that the TV news viewer can understand.

Note: the viewer assumes files are named with the following format CHANNEL_YYYYMMDD_hhmmss_SHOW.mp4.

Getting the viewer code

Clone the TV news viewer repository. https://github.com/scanner-research/tv-news-viewer.git

Perform the following steps in the README:

  1. Install Rust (see https://rustup.rs/)
  2. Clone submodules: git submodule init && git submodule update
  3. Run ./install_deps.sh to install the submodules
  • Specifically, if you look inside this script, we are installing the caption index(er) and a library for handling video interval intersections
  1. cd vgrid-widget and run ./install.sh. This will require npm and other js dependencies. Install them as needed. Once this succeeds, cd .. to return to the top level.
  2. Install python dependencies: pip3 install -r requirements.txt

Preparing the data

Run prepare_files_for_viewer.py <in_dir> <out_dir> with <in_dir> as the path to the directory containing output from the metadata extraction. This will put a number of files into the output directory <out_dir>. See the script for more detail as to what these files are.

Putting it all together

Navigate to the directory where you cloned the TV news viewer. Copy the data directory created by prepare_files_for_viewer.py as data here. Once that is done, run ./derive_data.py in the viewer repo.

You will want a place where the videos can be served. The easiest way to get started is to serve them from the same machine as static files for now. Symlink a directory containing the videos to static/videos in the viewer directory.

Now, take a look at develop.py in the viewer repository.

You will want to change video endpoint and the default index path lines:

DEFAULT_VIDEO_ENDPOINT = '/static/videos'
DEFAULT_INDEX_PATH = 'data/index'     # we saved the index under data

You also want to change some of the defaults in the call to build_app. See *** comments.

  app = build_app(
      data_dir, index_dir, video_endpoint, 
      video_auth_endpoint,                    # This is only needed if you have a
                                              # private and public video source
      fallback_to_archive=True,               # link videos directly from the archive
                                              # *** Make this False
      static_bbox_endpoint=None,
      static_caption_endpoint=None,
      host=None,
      min_date=datetime(2010, 1, 1),
      max_date=datetime(2029, 12, 31),
      tz=timezone('US/Eastern'),
      person_whitelist_file=None,
      min_person_screen_time=60 * 60,                     # *** Make this 0 (3600s by default)
      min_person_autocomplete_screen_time=10 * 60 * 60,   # *** Make this 0 (36000s by default)
      hide_person_tags=True,
      default_aggregate_by='month',           # *** Make this day unless you have years of data
      default_text_window=0,
      default_is_commercial=Ternary.false,
      default_color_gender_bboxes=True,
      allow_sharing=True,
      data_version='dev',
      show_uptime=True)

Once that is done, run ./develop.py and navigate to http://localhost:8080 in your browser. The app should now be serving.

Note: for a more sensible/permanent deployment take a look at wsgi.py and config.json. You can use this to serve the application with uwsgi, gunicorn, or any other wsgi server.