Tesla-dashcam

Fixed:

- Merge of videos fails when a relative path instead of an absolute path is provided for --output Issue 62
- Issue during processing of metadata if files were missing
- Hidden files (files starting with period) on Mac/Linux were not ignored. This could cause issues as some programs might create these files when viewing the video.
- Checking if font path exist in Windows failed.

New:

- Support for rear camera (introduced in V10). This also results in layouts having been modified to allow inclusion of rear camera. Issue 71
- Support for hardware encoding for systems with supported Intel GPUs.
- Support for hardware encoding on Raspberry Pi (RPi) (H.264 only) Issue 66
- Layout CROSS with front camera top centered, side camera underneath it, and rear camera then underneath side cameras centered.
- Layout DIAMOND with front camera top centered, rear camera under front and side camera centered at the left and right of front&rear.
- Option --motion_only to fast-forward through the portions in the video that does not have anything motion (done through removal of duplicate frames).
- Option --skip_existing to skip creation of video files that already exist. Existence only is checked, not if layout etc are the same.
- Option --perspective for showing side cameras to be displayed in perspective mode irrespective of layout. Layout PERSPECTIVE is thus same as layout FULLSCREEN with --perspective option.
- Options --start_offset and --end_offset can be used to provide starting and ending offset in seconds for resulting video (at folder level).
- Options --start_timestamp and --end_timestamp can be used to restrict resulting video (and processing) to specific timestamps. This can be used in combination with --start_offset and/or --end_offset
- Options --no-front, --no-left, --no-right, and --no-rear to exclude camera(s) from the videos
- Option --gpu_type to provide GPU installed in the system for Windows/Linux. Current supported options are nvidia, intel, and RPi.
- Option --no-faststart for not setting the faststart flag in the video files as doing this can result in encoding failures on network shares Issue 62
- Option --temp_dir to provide a different path to store the temporary video files that are created Issue 67
- Description metadata to include video was created by tesla_dashcam with version number.

Changed:

- WIDESCREEN layout will now by default show the front camera on top with higher resolution compared to others due to incorporation of rear camera
- Include folder SentryClips in default source list if no source provided (SavedClips was already default).
- Check to ensure that Python version is at required level or higher (currently 3.7).
- Existence of font file (provided or default) will be checked and error returned if not existing.
- Existence of ffmpeg will be checked and error returned if not existing.
- If no filename provided for merged video then current date/time will be used for filename.

Fixed:

- Folders will now be deleted if there are 0-byte or corrupt video files within the folder Issue 40
- Providing a filename for --output would create a folder instead and not setting resulting file to filename provided Issue 52
- Thread exception in Windows that ToastNotifier does not have an attribute classAtom (potential fix). Issue 54
- Traceback when invalid output path (none-existing) is provided or when unable to create target folder in given path.
- Including sub dirs did not work correctly, it would only grab the 1st folder.
- When using monitor, if . was provided as source then nothing would be processed. Now it will process everything as intended.
- File created when providing a filename with --output and --monitor option did not put timestamp in filename to ensure unique filenames
- Argument to get release notes was provided incorrectly when checking for updates. Thank you to demonbane for fixing.

New:
- Added chapter markers in the concatenated movies. Folder ones will have a chapter marker for each intermediate clip, merged one has a chapter marker for each folder.
- Option --chapter_offset for use with --merge to offset the chapter marker in relation to the folder clip.
- Added flags -movstart and +faststart for video files better suited with browsers etc. (i.e. YouTube). Thanks to sf302 for suggestion.
- Option to add trigger (--monitor_trigger_file) to use existence of a file/folder/link for starting processing instead of USB/SD being inserted.

Changed:
- Method for concatenating the clips together has been changed resulting in massive performance improvement (less then 1 second to do concatenation). Big thanks to sf302!

Fixed:

- Traceback when getting ffmpeg path in Linux Issue 39
- Running tesla_dashcam when installed using pip. Issue 38
- Just providing a filename for output would result in traceback.
- When providing a folder as output it would be possible that the last folder name was stripped potentially resulting in error

Fixed:
- Issue where sometimes encoding with GPU would fail by also allowing software based encoding
- Traceback when unable to retrieve latest release from GitHub
- Running tesla_dashcam when installed using pip. Issue 23 and Issue 31
- Folder clip would be removed if only 1 set in folder with same name as folder name if keep_intermediate not specified
- Font issue in Windows (hopefully final fix) Issue 29
- Python version has to be 3.7 or higher due to use of capture_output Issue 19

New:
- Added scale option to set the scale of the clips and thus resulting video. (--scale)
- Added option to specify a parameter file using <filename> where parameters can be located in. (<filename>)
- One can now specify multiple sources instead of just 1.
- Individual file(s) can now be provided as a source as well (only 1 camera filename has to be provided to get all 3)
- Source is now optional, if not provided then it will be same as --monitor_once with as source SavedClips.

Changed:
- Timestamp within video will now be used for concatenation of the clips at folder level and all (--merge option) instead of filename. This will ensure that even when crossing timezones the order of the video is still accurate.
- --delete_source will delete source files when specified even when --monitor or --monitor_once is not specified `Issue 28 <https://github.com/ehendrix23/tesla_dashcam/issues/28>`_
- Output will default to Videos\Tesla_Dashcam (Windows) Movies/Tesla_Dashcam (MacOS), or Videos\Tesla_Dashcam (Linux) if not output folder specified.
- Filename for the folder video files will not have start and end timestamp in local timezone instead of just folder name. `Issue 30 <https://github.com/ehendrix23/tesla_dashcam/issues/30>`_ and `Issue #33 <https://github.com/ehendrix23/tesla_dashcam/issues/33>`_
- Updated release notes for each release better identifying what is new, changed, and fixed.

Fixed:
- Issue where sometimes encoding with GPU would fail by also allowing software based encoding
- Traceback when unable to retrieve latest release from GitHub
- Running tesla_dashcam when installed using pip. Issue 23 and Issue 31
- Folder clip would be removed if only 1 set in folder with same name as folder name if keep_intermediate not specified
- Font issue in Windows (hopefully final fix) Issue 29
- Python version has to be 3.7 or higher due to use of capture_output Issue 19

New:
- Added scale option to set the scale of the clips and thus resulting video. (--scale)
- Added option to specify a parameter file using <filename> where parameters can be located in. (<filename>)
- One can now specify multiple sources instead of just 1.
- Individual file(s) can now be provided as a source as well (only 1 camera filename has to be provided to get all 3)
- Source is now optional, if not provided then it will be same as --monitor_once with as source SavedClips.

Changed:
- Timestamp within video will now be used for concatenation of the clips at folder level and all (--merge option) instead of filename. This will ensure that even when crossing timezones the order of the video is still accurate.
- --delete_source will delete source files when specified even when --monitor or --monitor_once is not specified `Issue 28 <https://github.com/ehendrix23/tesla_dashcam/issues/28>`_
- Output will default to Videos\Tesla_Dashcam (Windows) Movies/Tesla_Dashcam (MacOS), or Videos\Tesla_Dashcam (Linux) if not output folder specified.
- Filename for the folder video files will not have start and end timestamp in local timezone instead of just folder name. `Issue 30 <https://github.com/ehendrix23/tesla_dashcam/issues/30>`_ and `Issue #33 <https://github.com/ehendrix23/tesla_dashcam/issues/33>`_
- Updated release notes for each release better identifying what is new, changed, and fixed.

Biggest change is the ability for it to scan all subfolders and create the merged video files, and by having it monitor for the USB (or SD) card to be inserted and then automatically start processing.
This means that it could be setup to have it just run continuously, and anytime the Tesla DashCam USB (or SD) is inserted it will start creating the merged video files, and upon completion provide a notification.

This version will now also check for new releases, and inform if a new release is available.

List of changes:
- Added scanning of sub-folders clip files. Each folder will be processed and resulting movie file created. This can be disabled through parameter --exclude_subdirs
- Added option to merge the video files from multiple sub-folders into 1 movie file. Use parameter --merge to enable.
- Added option to monitor if the USB drive (or SD card) is inserted in the PC and then automatically start processing the files. Use parameter --monitor to enable.
Parameter --monitor_once will stop monitoring and exit after 1st time drive was inserted.
Parameter --delete_source will delete the source files and folder once the movie file for that folder has been created.
- Added update checker to determine if there is a newer version, additional arguments to just perform check (--check_for_update), include test releases (--include_test), or disable always checking for updates (--no-check_for_update)
- Formatted output to easily show progress
- Time it took to create the video files will now be provided upon completion of processing.
- Will now handle it much better if a video file from a camera is corrupt (i.e. zero-byte file).
- ffmpeg is part of the tesla_dashcam executable
- Desktop notification when processing starts (when using monitor) and when it completes.
- DockerFile added making it easy to run tesla_dashcam within Docker (jeanfabrice)
- Fixed: combining clips to movie would not use GPU or provided encoding.
- Added additional check that video file exist before merging into movie.