Skip to content

Latest commit

 

History

History
427 lines (349 loc) · 19.7 KB

README.md

File metadata and controls

427 lines (349 loc) · 19.7 KB

Python Quick Start

Python 3.6+ setup (required if not already installed)

This package uses f-strings (more here), and so requires Python 3.6+.

If you have an older version of Python, you can download Python 3.9.1 (follow links below) and follow the instructions to set up Python for your machine. If you want to install a different version, visit the Python Downloads page and select the version you want.

Permissions for first run

This is required to make sure you can download and install the required Selenium binary dependencies.

On Windows: make sure you open Command Prompt or Powershell (both work) in "Run as Administrator" mode
  • shortcut: ⊞ Win + X + A
On Unix based machines (MacOS, Linux): make sure you have read and write access to /usr/local/bin/
  • if you're not sure, open terminal and run sudo chown $USER /usr/local/bin/

Using venv (optional)

While creating a virtual environment is not required to use this package, creating a virtual environment is useful for avoiding dependency conflicts with other projects. If you are sure you do not need to worry about dependency conflicts with other projects, skip this step.

Python has many ways to set up and use a virtual environment. The following instructions use the venv provided with the python standard library for simplicity. You do not need to use this particular implementation of a virtual environment, but virtual environments are outside of the scope of this project, so you will need to figure out how to set up and use a different implementation of python virtual environments on your own if you choose a different implementation of a virtual environment, since there are too many different variations to cover here.

### CREATING the virtual environment on MacOS/Linux ###

python3 -m venv ytvl-venv
source ytvl-venv/bin/activate
# python3 # enter the python shell inside this virtual environment
deactivate # exit this virtual environment

### USING the virtual environment on MacOS/Linux ###

# if ytvl-venv is in the directory you are currently in:
source ytvl-venv/bin/activate

# if ytvl-venv is NOT in the directory you are currently in:
source /absolute/path/to/ytvl-venv/bin/activate

deactivate # exit this virtual environment
### CREATING the virtual environment on Windows (NOT FOR git BASH) ###

python -m venv ytvl-venv
ytvl-venv\Scripts\activate
# python # enter the python shell inside this virtual environment
deactivate # exit this virtual environment

### USING the virtual environment on Windows (NOT FOR git BASH) ###

# if ytvl-venv is in the directory you are currently in:
ytvl-venv\Scripts\activate

# if ytvl-venv is NOT in the directory you are currently in:
## you may need to
##    include the .ps1 extenstion (activate.ps1) in Powershell
## or include the .bat extension (activate.bat) in Command Prompt
\absolute\path\to\ytvl-venv\Scripts\activate

deactivate # exit this virtual environment
### CREATING the virtual environment on Windows (FOR git BASH) ###

python -m venv ytvl-venv
source ytvl-venv/Scripts/activate
# python # enter the python shell inside this virtual environment
deactivate # exit this virtual environment

### USING the virtual environment on Windows (FOR git BASH) ###

# if ytvl-venv is in the directory you are currently in:
source ytvl-venv/Scripts/activate

# if ytvl-venv is NOT in the directory you are currently in:
source /absolute/path/to/ytvl-venv/Scripts/activate

deactivate # exit this virtual environment
Installing the package

After you install Python 3.6+ and ensure you have the required permissions as needed and have activated your virtual environment as required (if you decide to use a virtual environment - you do not need to use a virtual environment, but if you choose to use venv, follow the instructions above), enter the following in your command line:

# if something isn't working properly, try rerunning this
# the problem may have been fixed with a newer version

pip3 install -U yt-videos-list     # MacOS/Linux
pip  install -U yt-videos-list     # Windows


# if that doesn't work:

python3 -m pip install -U yt-videos-list     # MacOS/Linux
python  -m pip install -U yt-videos-list     # Windows
If you're on Windows: make sure you always open Command Prompt or Powershell (both work) in "Run as Administrator" mode!
  • shortcut: ⊞ Win + X + A
  • this allows yt_videos_list to update selenium webdriver binaries to be compatible with newer browser versions as browsers are updated (e.g. your Firefox browser updates from version 77 to version 82)
    • to see the commands being run, see the yt_videos_list/docs/dependencies.json file
Running the package from the python interpreter
python3     # MacOS/Linux
python      # Windows
from yt_videos_list import ListCreator


my_driver = 'firefox' # SUBSTITUTE DRIVER YOU WANT (options below)
lc = ListCreator(driver=my_driver, scroll_pause_time=0.8)


lc.create_list_for(url='https://www.youtube.com/user/schafer5')
lc.create_list_for(url='https://www.youtube.com/channel/UC8butISFwT-Wl7EV0hUK0BQ', log_silently=True)
# Set `log_silently` to `True` to mute program logging to the console.
# The program will log the prgram status and any program information
# to only the log file for the channel being scraped
# (this is useful when scraping multiple channels at once with multi-threading).
# By default, the program logs to both the log file for the channel being scraped AND the console.


# to name the file using the channel ID instead of the channel name, set file_name='id'
# this is useful when scraping multiple channels with the same name:
lc.create_list_for(url='https://www.youtube.com/channel/UCb2EYjrzI6WpNAmPZeihhag', file_name='id')
lc.create_list_for(url='https://www.youtube.com/channel/UCDzYhlGOvGqsYw8IaTKDT8g', file_name='id')

# for more details about this method:
help(lc.create_list_for)


# see the new files that were just created:
import os
os.system('ls -lt | head')                      # MacOS/Linux
os.system('dir /O-D | find "_videos_list"')     # Windows

# for more information on using the module:
help(lc)
  • driver options include:
    • 'firefox'
    • 'opera'
    • 'safari' (MacOS only)
    • 'chrome'
    • 'brave'
    • 'edge' (Windows only!)
  • increase scroll_pause_time for laggy internet and decrease scroll_pause_time for fast internet

If you already scraped a channel and the channel uploaded a new video, simply rerun this program on that channel and this package updates your files to include the newer video(s)!

Scraping multiple channels from a file simultaneously with multi-threading

Add the url to every channel you want to extract information from in a txt file with every url placed on a new line.

  • example: channels.txt (NOTE this is a relative link, so this might not link properly on non-GitHub hosted sites!)

Enter the python interpreter:

python3     # MacOS/Linux
python      # Windows
from yt_videos_list import ListCreator

lc = ListCreator(driver='firefox', scroll_pause_time=1.2)
lc.create_list_from(path_to_channel_urls_file='channels.txt', number_of_threads=4)

# configuring settings:
lc.create_list_from(
  path_to_channel_urls_file='channels.txt',
  number_of_threads=4,
  min_sleep=1,
  max_sleep=5,
  after_n_channels_pause_for_s=(20, 10),
  log_subthread_status_silently=False,
  log_subthread_info_silently=False
)                                                                     # defaults (keyword argument form)
lc.create_list_from('channels.txt', 4, 1, 5, (20, 10), False, False)  # defaults (positional argument form)
lc.create_list_from('channels.txt', min_sleep=3, max_sleep=10)        # modifying only min_sleep and max_sleep

help(lc.create_list_from) # see API method details
Explicitly downloading all Selenium dependencies

Ideal if you use Selenium for other projects 😎

  • Make sure you already have the yt-videos-list package installed (follow directions above for getting set up), then run the following:
pip3 install -U yt-videos-list # MacOS/Linux: ensure latest package
python3                        # MacOS/Linux: enter python interpreter
pip install -U yt-videos-list  # Windows:     ensure latest package
python                         # Windows:     enter python interpreter
from yt_videos_list.download import selenium_webdriver_dependencies
selenium_webdriver_dependencies.download_all()

That's all! 🤓

More API information

NOTE that you can also access all the information below from the Python interpreter by entering

import yt_videos_list
help(yt_videos_list)

# default options for the ListCreator instance

ListCreator(
  txt=True,
  csv=True,
  md=True,
  file_suffix=True,
  all_video_data_in_memory=False,
  video_data_returned=False,
  video_id_only=False,
  reverse_chronological=True,
  headless=False,
  scroll_pause_time=0.8,
  driver='firefox',
  cookie_consent=False,
  verify_page_bottom_n_times=3,
  file_buffering=-1,
  )

There are a number of optional arguments you can specify during the instantiation of the ListCreator instance. The preceding arguments are run by default, but in case you want more flexibility, you can specify the:

  • driver argument:
    • Firefox (default)
    • Opera
    • Safari (MacOS only)
    • Chrome
    • Brave
    • Edge (Windows only)
      • driver='firefox'
      • driver='opera'
      • driver='safari'
      • driver='chrome'
      • driver='brave'
      • driver='edge'
  • cookie_consent argument:
    • False (default) - block all cookie options if prompted by YouTube (at consent.youtube.com)
    • True - accept all cookie options if prompted by YouTube (also at consent.youtube.com)
      • cookie_consent=False (default) OR cookie_consent=True
  • txt, csv, md file type argument:
    • True (default) - create a file for the specified type
    • False - do not create a file for the specified type
      • txt=True (default) OR txt=False
      • csv=True (default) OR csv=False
      • md=True (default) OR md=False
  • file_suffix argument:
    • True (default) - add a file suffix to the output file name
      • ChannelName_reverse_chronological_videos_list.csv
      • ChannelName_chronological_videos_list.csv
    • False - do NOT add a file suffix to the output file name
      • this means if a reverse chronological file and a chronological file is made for the same channel, they will have the same name!
      • ChannelName.csv (reverse chronological output file)
      • ChannelName.csv (chronological output file) -> file_suffix=True (default) OR file_suffix=False
  • all_video_data_in_memory argument:
    • False (default) - do not scrape the entire page
    • True - scrape the entire page (must ALSO set the video_data_returned attribute to True to return this data!)
      • all_video_data_in_memory=False (default) OR all_video_data_in_memory=True
  • video_data_returned argument:
    • False (default) - do not return video data collected from the current scrape job (return dummy data instead: [[0, '', '', '']])
    • True - return video data collected from the current scrape job
      • if all_video_data_in_memory attribute set to False, the returned data MIGHT not be the full data, and video numbering MIGHT be incorrect
      • set all_video_data_in_memory attribute to True to return ALL video data for channel (video number will then also ALWAYS be correct)
        • video_data_returned=False (default) OR video_data_returned=True
  • video_id_only argument:
    • False (default) - include the full URL to video: https://www.youtube.com/watch?v=ElevenChars
    • True - include only the identifier parameter to video: ElevenChars
      • video_id_only=False (default) OR video_id_only=True
  • reverse_chronological argument:
    • True (default) - write the files in order from most recent video to the oldest video
    • False - write the files in order from oldest video to the most recent video
      • reverse_chronological=True (default) OR reverse_chronological=False
  • headless argument:
    • False (default) - run the driver with an open Selenium instance for viewing
    • True - run the driver in "invisible" mode
      • headless=False (default) OR headless=True
  • scroll_pause_time argument:
    • any float values greater than 0 (default 0.8)
      • The value you provide will be how long the program waits before trying to scroll the videos list page down for the channel you want to scrape. For fast internet connections, you may want to reduce the value, and for slow connections you may want to increase the value.
    • scroll_pause_time=0.8 (default)
    • CAUTION: reducing this value too much will result in the program not capturing all the videos, so be careful! Experiment :)
  • verify_page_bottom_n_times argument:
    • any int values greater than 0 (defaults to 3)
    • NOTE: this argument is only used when CREATING a new file for a new channel, and is unused when UPDATING an existing file for an already scraped channel.
    • The value you provide will be how many times the program needs to verify it acually reached the bottom of the page before accepting it is the bottom of the page, and starting to write the information to the output file(s).
    • For channels that have uploaded THOUSANDS of videos, increase this value to a large number that you think should be sufficient to verify the program reached the bottom of the page.
    • To determine HOW large of a value you should provide, determine the length of time you'd like to wait before being reasonably sure that you reached the bottom of the page and it's not just YouTube's server trying to fetch the response from an old database entry, and divide the time you decided to wait by the scroll_pause_time argument.
      • For example, if you want to wait 45 seconds and you set the scrioll_pause_time value to 1.0: -> your_time / scroll_pause_time -> 45 / 1.0 -> 45 -> therefore: verify_page_bottom_n_times=45
      • For channels with only a couple hundred videos (or less), the default value of verify_page_bottom_n_times=3 should be sufficient.
    • See commit a68f8f62e5c343cbb0641125e271bb96cc4f0750 for more details.
  • file_buffering argument:
scrapetube integration

scrapetube is a much more efficient backend developer tool that loads the videos uploaded by a channel. This package also supports loading information from playlists and searches, which yt-videos-list currently does not do. Integration with scrapetube will be available in a future yt-videos-list release!

To keep things backwards-compatible and maintainable, the scrapetube integration will be accessible through an almost identical, separate interface as the ListCreator interface, and the original ListCreator interface will continue to be available and continue to receive updates. 🤓

Cloning and running locally

To clone the repository and install the most updated version of the package that may not yet be available on the latest release through PyPI, clone this repository and run:

cd yt_videos_list/python # MacOS/Linux
python3 -m pip install . # MacOS/Linux

cd yt_videos_list\python # Windows
python -m pip install .  # Windows

To make your own changes to the yt_videos_list python package and run the changes locally:

# make changes to the codebase in the
# ===> /dev <=== directory
python3 minifier.py           # MacOS/Linux
pip3 install .                # MacOS/Linux

python minifier.py            # Windows
pip install .                 # Windows

NOTE: make the changes to the codebase in the yt_videos_list/python/dev directory!!

  • the code in the yt_videos_list/python/yt-videos-list directory is minified with
    • leading indents stipped to the minimum (1 space for each nested scope)
    • whitespace for padding (e.g. extra spaces to align variable assignments) stripped
    • comments stripped
  • as a result, the code in the yt_videos_list/python/yt-videos-list directory is NOT human readable, and the yt_videos_list/python/dev directory should be used for development instead!
    • the minifier.py module performs all the code preprocessing and packages the code from yt_videos_list/python/dev into the final version seen in the yt_videos_list/python/yt-videos-list directory
    • so running minifier.py before installing the local package with pip install . (Windows) or pip3 install . is essential!
Running tests

The tests use the custom ThreadWithResult subclass of threading.Thread provided by the save-thread-result package, so make sure you install that module using

pip3 install -U save-thread-result     # MacOS/Linux
pip  install -U save-thread-result     # Windows

# if that doesn't work:

python3 -m pip install -U save-thread-result     # MacOS/Linux
python  -m pip install -U save-thread-result     # Windows

Then, make sure you're in the yt_videos_list/python directory, then run:

tests\run_tests.bat     # Windows
####       Any shell on   MacOS/Linux
bash tests/run_tests.sh # this works
csh  tests/run_tests.sh # this works
dash tests/run_tests.sh # this works
ksh  tests/run_tests.sh # this also works
tcsh tests/run_tests.sh # this works too
zsh  tests/run_tests.sh # this works as well
# you can try other shells and
# they should work too, since
# there's no special syntax in
# the run_tests.sh file