YouTube Spammer Purge Versions Save

😤 [2.18] MAJOR New Feature: Much faster filter loading before scans

• Now instead of compiling all the regex filters before every scan, the program will save cached files of the longest lists
• This eliminates nearly all of the loading time before every scan
• The compiling will only need to be done once whenever new filter files are downloaded

🎉 [2.18] New Features:

• Encrypted Token File

  • Adds new encrypt_token_file option in config, which is True by default. If True, next time the program is run, or whenever the user logs in, the token file will be replaced by an encrypted version. The user will enter the password for this.
  • If False, it will not have the user enter a password to encrypt the file. But if an encrypted token file already exists, it will still prompt the user for the password.

• Dynamic Sample Text Length From Window Size

  • The printing of the comment samples list after a scan will now automatically adapt to the window size, so you can expand the window to see more of each comment sample

• Highlighted Spam List Matches in Samples

  • Text in the samples list that matches a known spam list will be highlighted so they are easier to identify

⚠️ [2.18] Other Notable Changes

• Disabled 'possible false positive' check because it was not reliable at all
• Changed moderator mode check to allow moderators fully automated holding of comments (#935)

---

Beta1 → Beta2:

• Fixed error when token expired
• Improve password flow, now shows asterisks as you type
• Should fix unescaped HTML characters showing in titles (#791)
• Changed moderator mode check to allow moderators fully automated holding of comments
• Added cached, pre-compiled regex, which will immensely speed up loading filters before scans. The longest loading now only has to happen once when new filter lists are downloaded
• Disabled 'possible false positive' check because it was not reliable at all
• The printing of the comment samples list after a scan will now automatically adapt to the window size, so you can expand the window to see more of each comment sample

🟢 Beta2 → Beta3:

• Add colorization of specific matches to known spam lists in the sample list

😤 [2.18] MAJOR New Feature: Much faster filter loading before scans

• Now instead of compiling all the regex filters before every scan, the program will save cached files of the longest lists
• This eliminates nearly all of the loading time before every scan
• The compiling will only need to be done once whenever new filter files are downloaded

🎉 [2.18] New Features:

• Encrypted Token File

  • Adds new encrypt_token_file option in config, which is True by default. If True, next time the program is run, or whenever the user logs in, the token file will be replaced by an encrypted version. The user will enter the password for this.
  • If False, it will not have the user enter a password to encrypt the file. But if an encrypted token file already exists, it will still prompt the user for the password.

• Dynamic Sample Text Length From Window Size

  • The printing of the comment samples list after a scan will now automatically adapt to the window size, so you can expand the window to see more of each comment sample

⚠️ [2.18] Other Notable Changes

• Disabled 'possible false positive' check because it was not reliable at all
• Changed moderator mode check to allow moderators fully automated holding of comments (#935)

---

🟢 Beta1 → Beta2:

• Fixed error when token expired
• Improve password flow, now shows asterisks as you type
• Should fix unescaped HTML characters showing in titles (#791)
• Changed moderator mode check to allow moderators fully automated holding of comments
• Added cached, pre-compiled regex, which will immensely speed up loading filters before scans. The longest loading now only has to happen once when new filter lists are downloaded
• Disabled 'possible false positive' check because it was not reliable at all
• The printing of the comment samples list after a scan will now automatically adapt to the window size, so you can expand the window to see more of each comment sample

😤 [2.17] MAJOR New Feature: Remote filter Update Without Needing Update

• The program will now automatically go and check the GitHub repo to see if there is a new version of filter_variables.py
• A full program update is no longer required to update the filters, which will make it much easier to make small and/or faster filter improvements

🎉 [2.17] New Features:

• Multi-Variable Spam Checking

  • The program now has new filtering functions that can check more complex spam patterns

• New Fuzzy Duplicate Detection

  • The program now uses the "RapidFuzz" module which can be many times faster than the previous method, meaning duplicate comments will be scanned MUCH faster

• Channel ID Lookup Tool

  • There is a new main menu option for "helpful tools". The only one right now allows you to retrieve the Channel ID for a channel by providing a YouTube video URL for that channel.

📈 [2.17] Other Improvements

• Improved update checking method should allow you to still be notified of the latest beta version, even if there was a stable release of the previous version after it

• Add support for YouTube @ handles

• Added automatic retry of downloads (program or list updates) on fail, using alternative download method

⚠️ [2.17] Other Notable Changes

• Updated requirements.txt because a package was renamed and was causing errors. If using python version, you may need to do pip uninstall python-levenshtein then pip install --upgrade levenshtein

🟢 2.17.0 → 2.17.1:

• Fix community post scanning

🎉 [2.18] New Features:

• Encrypted Token File

  • Adds new encrypt_token_file option in config, which is True by default. If True, next time the program is run, or whenever the user logs in, the token file will be replaced by an encrypted version. The user will enter the password for this.
  • If False, it will not have the user enter a password to encrypt the file. But if an encrypted token file already exists, it will still prompt the user for the password.

😞 [2.18] Known Issues

• When choosing the encryption password for the first time, it might ask you twice
• It will show an error when a token expires

😤 [2.17] MAJOR New Feature: Remote filter Update Without Needing Update

• The program will now automatically go and check the GitHub repo to see if there is a new version of filter_variables.py
• A full program update is no longer required to update the filters, which will make it much easier to make small and/or faster filter improvements

🎉 [2.17] New Features:

• Multi-Variable Spam Checking

  • The program now has new filtering functions that can check more complex spam patterns

• New Fuzzy Duplicate Detection

  • The program now uses the "RapidFuzz" module which can be many times faster than the previous method, meaning duplicate comments will be scanned MUCH faster

• Channel ID Lookup Tool

  • There is a new main menu option for "helpful tools". The only one right now allows you to retrieve the Channel ID for a channel by providing a YouTube video URL for that channel.

📈 [2.17] Other Improvements

• Improved update checking method should allow you to still be notified of the latest beta version, even if there was a stable release of the previous version after it

• Add support for YouTube @ handles

• Added automatic retry of downloads (program or list updates) on fail, using alternative download method

⚠️ [2.17] Other Notable Changes

• Updated requirements.txt because a package was renamed and was causing errors. If using python version, you may need to do pip uninstall python-levenshtein then pip install --upgrade levenshtein

😤 [2.17] MAJOR New Feature: Remote filter Update Without Needing Update

• The program will now automatically go and check the GitHub repo to see if there is a new version of filter_variables.py
• A full program update is no longer required to update the filters, which will make it much easier to make small and/or faster filter improvements

🎉 [2.17] New Features:

• Multi-Variable Spam Checking

  • The program now has new filtering functions that can check more complex spam patterns

• New Fuzzy Duplicate Detection

  • The program now uses the "RapidFuzz" module which can be many times faster than the previous method, meaning duplicate comments will be scanned MUCH faster

• Channel ID Lookup Tool

  • There is a new main menu option for "helpful tools". The only one right now allows you to retrieve the Channel ID for a channel by providing a YouTube video URL for that channel.

📈 [2.17] Other Improvements

• Improved update checking method should allow you to still be notified of the latest beta version, even if there was a stable release of the previous version after it

• Add support for YouTube @ handles

• Added automatic retry of downloads (program or list updates) on fail, using alternative download method

Beta1 → Beta2:

• Fix incorrect config version number that caused repeated incorrect warning about out-of-date config
• Updated filter variables and confusables

Beta2 → Beta3:

• Fixed replies not being checked against spam thread list, which caused many missed spam comments
• Fixed Comment Recovery feature not working
• Fixed validation error in recent video count
• Fix rare UnboundLocalError for VideoID
• Update custom confusables to check for more look-alike characters
• Updated requirements.txt
• Many grammatical corrections

Beta3 → Beta4:

• Fixed program not checking for filter_variables.py file updates when manually selecting to check for updates in main menu
• Fixed bug where the program would download the filter variables file even when the latest version already exists
• Fixed program not catching spam bot thread matches correctly
• Fixed infinite loop when invalid/unavailable video ID is used in videos_to_scan config setting

Beta4 → Beta6:

• Adds support for searching channels by YouTube handle
• Improves accuracy of searching for a channel using custom URL
• Added new main menu option for "helpful tools". The only one right now allows you to retrieve the Channel ID for a channel by providing a YouTube video URL for that channel.

🟢 Beta6 → Beta7:

• User Tools menu now has better warning when an invalid choice is made.
• The video ID validation function now displays more specific error messages depending on whether the entry appears to be a URL or not
• Added handling for OSError when checking for updates
• Added automatic retry of download on fail, using alternative download method (stream true/false, depending on previous attempt)
• Updated requirements.txt because a package was renamed and was causing errors. May need to do pip uninstall python-levenshtein then pip install --upgrade levenshtein

😤 [2.17] MAJOR New Feature: Remote filter Update Without Needing Update

• The program will now automatically go and check the GitHub repo to see if there is a new version of filter_variables.py
• A full program update is no longer required to update the filters, which will make it much easier to make small and/or faster filter improvements

🎉 [2.17] New Features:

• Multi-Variable Spam Checking

  • The program now has new filtering functions that can check more complex spam patterns

• New Fuzzy Duplicate Detection

  • The program now uses the "RapidFuzz" module which can be many times faster than the previous method, meaning duplicate comments will be scanned MUCH faster

• Channel ID Lookup Tool

  • There is a new main menu option for "helpful tools". The only one right now allows you to retrieve the Channel ID for a channel by providing a YouTube video URL for that channel.

📈 [2.17] Other Improvements

• Improved update checking method should allow you to still be notified of the latest beta version, even if there was a stable release of the previous version after it

• Add support for YouTube @ handles

Beta1 → Beta2:

• Fix incorrect config version number that caused repeated incorrect warning about out-of-date config
• Updated filter variables and confusables

Beta2 → Beta3:

• Fixed replies not being checked against spam thread list, which caused many missed spam comments
• Fixed Comment Recovery feature not working
• Fixed validation error in recent video count
• Fix rare UnboundLocalError for VideoID
• Update custom confusables to check for more look-alike characters
• Updated requirements.txt
• Many grammatical corrections

Beta3 → Beta4:

• Fixed program not checking for filter_variables.py file updates when manually selecting to check for updates in main menu
• Fixed bug where the program would download the filter variables file even when the latest version already exists
• Fixed program not catching spam bot thread matches correctly
• Fixed infinite loop when invalid/unavailable video ID is used in videos_to_scan config setting

🟢 Beta4 → Beta6:

• Adds support for searching channels by YouTube handle
• Improves accuracy of searching for a channel using custom URL
• Added new main menu option for "helpful tools". The only one right now allows you to retrieve the Channel ID for a channel by providing a YouTube video URL for that channel.

😤 [2.17] MAJOR New Feature: Remote filter Update Without Needing Update

• The program will now automatically go and check the GitHub repo to see if there is a new version of filter_variables.py
• A full program update is no longer required to update the filters, which will make it much easier to make small and/or faster filter improvements

🎉 [2.17] New Features:

• Multi-Variable Spam Checking

  • The program now has new filtering functions that can check more complex spam patterns

• New Fuzzy Duplicate Detection

  • The program now uses the "RapidFuzz" module which can be many times faster than the previous method, meaning duplicate comments will be scanned MUCH faster

📈 [2.17] Other Improvements

• Improved update checking method should allow you to still be notified of the latest beta version, even if there was a stable release of the previous version after it

Beta1 → Beta2:

• Fix incorrect config version number that caused repeated incorrect warning about out-of-date config
• Updated filter variables and confusables

Beta2 → Beta3:

• Fixed replies not being checked against spam thread list, which caused many missed spam comments
• Fixed Comment Recovery feature not working
• Fixed validation error in recent video count
• Fix rare UnboundLocalError for VideoID
• Update custom confusables to check for more look-alike characters
• Updated requirements.txt
• Many grammatical corrections

🟢 Beta3 → Beta4:

• Fixed program not checking for filter_variables.py file updates when manually selecting to check for updates in main menu
• Fixed bug where the program would download the filter variables file even when the latest version already exists
• Fixed program not catching spam bot thread matches correctly
• Fixed infinite loop when invalid/unavailable video ID is used in videos_to_scan config setting

😤 [2.17] MAJOR New Feature: Remote filter Update Without Needing Update

• The program will now automatically go and check the GitHub repo to see if there is a new version of filter_variables.py
• A full program update is no longer required to update the filters, which will make it much easier to make small and/or faster filter improvements

🎉 [2.17] New Features:

• Multi-Variable Spam Checking

  • The program now has new filtering functions that can check more complex spam patterns

• New Fuzzy Duplicate Detection

  • The program now uses the "RapidFuzz" module which can be many times faster than the previous method, meaning duplicate comments will be scanned MUCH faster

📈 [2.17] Other Improvements

• Improved update checking method should allow you to still be notified of the latest beta version, even if there was a stable release of the previous version after it

Beta1 → Beta2:

• Fix incorrect config version number that caused repeated incorrect warning about out-of-date config
• Updated filter variables and confusables

🟢 Beta2 → Beta3:

• Fixed replies not being checked against spam thread list, which caused many missed spam comments
• Fixed Comment Recovery feature not working
• Fixed validation error in recent video count
• Fix rare UnboundLocalError for VideoID
• Update custom confusables to check for more look-alike characters
• Updated requirements.txt
• Many grammatical corrections

🎉 New Features:

• Stolen Comment Detection

  • New ability to detect users who copied and reposted an existing comment
  • By default, is only enabled in Sensitive-Smart mode, but can be enabled in any mode through config option stolen_comments_check_modes
  • By default, it detects exact matches for speed. But with config setting fuzzy_stolen_comment_detection, you can check comments that are similar. Uses the same levenshtein_distance value as when detecting duplicates
    • Note: Fuzzy stolen comment detection is much slower, and required time is exponentially proportional to the number of comments

• Ability to Set Custom Directory for Config Files

  • You can now use a new config setting called configs_path to choose where to put secondary config files
  • If you keep the new default location as 'configs', you can also put the main config file in there too
    • Otherwise, at least the main config file must be kept in the same place as the program, so it knows where to look

📈 Other Improvements

• 10x Improvement in API quota efficiency when fetching recent videos. Now fetches 50 videos at once, instead of 5. This means one could theoretically fetch up to 5000 recent videos per day, instead of just 500.

• Match sample lists are now sorted by number of comments per author (highest to lowest)

• Match sample list now shows warning if some comments are possible false positives, and highlights them

• Filter mode selection menu now better illustrates which modes can be used in "Not Your Channel Mode" and Moderator Mode

• Log files will now include a list of the videos or community posts that were scanned

• Log files now display timestamps of comments (except not community posts for now)

• New colors_enabled config setting, which allows disabling colored text if necessary

• New minimum_duplicate_length config setting, to set comment length threshold for duplicates

• Many filtering and detection improvements

Other Changes & Fixes

• Increased default minimum duplicates amount threshold from 4 to 8
• Added a few additional startup messages to help narrow down slow startups for some people
• When displaying spam threads sample list, if the video is title has investment related keywords, added notice to be extra vigilant in checking false positives of spam thread results
• Completely re-worked filter variable intake. Variables no longer use encoded obfuscation, instead precompiled regex is directly loaded, improving loading speed. Filter variables now exist in filter_variables.py
• Fixed a blocking prompt on chosenVideos mode even if skip_confirm_video is True
• Fixed issue where excluding multiple times back to back would cause duplicate whitelist entries

Patch 2.16.0 → 2.16.1:

• Fix 'NoneType AttributeError' crash on startup when certain config settings are used, such as channel_to_scan
• 10x Improvement in API quota efficiency when fetching recent videos. Now fetches 50 videos at once, instead of 5. This means one could theoretically fetch up to 5000 recent videos per day.
• Fix rare crash caused by comments from deleted channels when writing JSON log

Patch 2.16.1 → 2.16.2:

• Fixed crash that would occur if no 'configs' folder exists, and user tries to create second config file

2.16.2 → 2.16.3:

• Several updates to filtering to significantly improve detection of certain types of spam, and reduction of false positives

2.16.3 → 2.164:

• Fixed crash when manually entering list of comment IDs to delete
• Reduce false positives in auto smart mode

2.16.4 → 2.16.6:

• Fixed crash when checking success of deleted or recovered comments
• If an error occurs while checking if comment was deleted, now displays the reason if available
• (v2.16.5 skipped because of missing packaged dll)

2.16.6 → 2.16.7:

• Improved filter variables
• Improved recognition of look-alike characters

2.16.7 → 2.16.8:

• Fixed custom regex filter mode not working, and also causing TypeError when logging
• Fixed UnboundLocalError when logging in some circumstances

2.16.8 → 2.16.9:

• Fix very rare crash from ZeroDivisionError

🟢 2.16.9 → 2.16.10:

• Fixed major issue that was preventing the comment recovery functionality from working
• Fixed rare error that resulted from 'videoID' variable having no value for unknown reasons
• Fix minor issue with input validation for recent video count
• Updated to latest filter variables