process-clips

#!/bin/bash

# Do a lot of steps to tag, sort, and catalog webp images.

# Given a file, list of files, or directory, including webp animated clips,
# Process them all by:
# * amending the files metadata to add a UUID and provenance.
# * if not already existing, derive a gif version of the webp
# * duplicate the metadata from the webp into the gif, so the gif has a reference to the original
# * archive the webp into a totally different place, giving it a filename derived from the cannonic UUID so that we can find it again later.

function processFile {

    local pid=$$
    local identifier="[PID $pid]"
    local exif_option="-quiet"

    local source_file="$1"

    if [[ ! -f "${source_file}" ]]; then
        echo "$identifier Error: $source_file is not a file"
        return 1
    fi

    local mime_type=$(exiftool -s -S -MimeType "${source_file}")
    echo "$identifier Processing $source_file ($mime_type)" >&2

    # Correct for misleading mime type. Do this before all else.
    new_file=$( "correct-webp-jpegs" "${source_file}" )

    if [ ! -z "$new_file" ] &&  [ ! "$new_file" == "$source_file" ]; then
        # a change happened.
        echo "new file after verification is '$new_file' " >&2
        source_file="$new_file"
        mime_type=$(exiftool -s -S -MimeType "${source_file}")
    fi

    if [[ "$mime_type" == "image/webp" ]] ; then
      echo "$identifier Processing webp '$source_file'" >&2

      new_file=$(processWebpFile "${source_file}")
      if [ ! -z "$new_file" ] &&  [ ! "$new_file" == "$source_file" ]; then
        # a change happened.
        echo "new file after verification is '$new_file' " >&2
        # either a jpeg or gif derivative is what we should work on from ow on.
        source_file="$new_file"
        mime_type=$(exiftool -s -S -MimeType "${source_file}")
        # continue below
      else
        return 0
      fi

    fi

    if [[ "$mime_type" == "image/gif" ]] ; then
      # Ensure its metadata is up to date
      duplicate-exif-from-webp "${source_file}"
      # Even if there wasn't a webp to scrape, can also append other metadata from filename or directory
      process-metadata "${source_file}"
      return 0
    fi

    return 0
}

function processWebpFile {
    local source_file="$1"

    # Ensure its metadata is populated
    process-metadata "${source_file}"
    # Ensure it's got a corresponding gif
    webp2gif "${source_file}"
    local gif_version="${source_file%.*}.gif"
    # Ensure the gif has a copy of the metadata
    duplicate-exif-from-webp "${gif_version}"
    # Shift the webp into the archive.
    move-file-to-uuid-archive "${source_file}"
    # Return the new derivative for any caller to work with.
    echo "${gif_version}"

    return 0
}

function parallalProcessFiles {
    # Find all files recursively in the input arguments.
    # This uses parallal processing - note that feedback from the sub-processes may get jumpbled.
    export -f processFile;
    export -f processWebpFile;
    # Additional options `-print0` & `-0` are here to manage spaces in filenames.
    # -P 4 is parallalization.
    find "$@" -type f -print0 | xargs -0 -P 4 -n 1 -I{} bash -c 'processFile "$1"' _ "{}"
}

# Call the processFiles function with the provided arguments
parallalProcessFiles "$@"