Supported media file types / extensions

[PatMyron]

These seem like the current supported file extensions:

https://github.com/dabutvin/ImgBot/blob/392f7b20e4117e363b78e2537e4d915cabefcc26/Common/KnownImgPatterns.cs#L5

Would love additional formats like .svg, .pdf, and audio+video formats like .mp4

[dabutvin]

hey @PatMyron Would love to see imgbot expanding out like this.

I had this thought the other day about svgs, but couldn't find a good dotnet optimizer for them and I naively started my own (https://github.com/dabutvin/SvgClean)

I bet we can find existing optimizers for videos + pdfs though! I hadn't considered them before. Do you have any suggestions for optimizers for any of these file types?

[PatMyron]

FFmpeg?

[geeknoid]

I recommend integrating with https://github.com/svg/svgo which does an excellent job in crunching down SVGs. It'd be great to have ImgBot take care of those.

[dabutvin]

@geeknoid that would be really cool. I wonder what the best way to fit that into the rest of the compression workflow would be, given that svgo is all JavaScript

[geeknoid]

svgo can run on a specific file, on a specific directory, or on a specific directory recursively.

I don't know how ImgBot is put together, but if I were to hack this together, I'd just kick off "svgo -r -f " as an external command. Any file it modifies, you include in the PR you post.

Now to run svgo, you would need to install it (npm install svgo -g) and then use the npx tool with it (npx svgo ). If this isn't working for some reason, an alternate solution is to create a docker image that contains svgo and use 'docker run' to execute it.

[dabutvin]

makes sense @geeknoid 👍

I just opened #341 which would make this a lot more feasible - along with some potential performance improvements

[dabutvin]

It would be great to get svgo runnable as an executable, alternatively we would have to install node on the container - might not be so bad

[dabutvin]

just opened #386 to start adding svg support with svgo. There's some more work to do, but it's a start

[stale[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[PatMyron]

@dabutvin re-open?

[dabutvin]

SVG support is in!!!!

[dabutvin]

see #488 for an example of an issue for adding a new compression library. We can use the same template for adding whole new file extensions.

[gleich]

Any update of pdf compression? It would be really nice for people that use LaTex.

[dabutvin]

Hi @Matt-Gleich !

I'm totally cool with adding pdf compression. You recommend any libraries or tools that we should look at?

[gleich]

A quick DuckDuckGo search brought up some useful tutorials on how to compress PDFs in C# https://duckduckgo.com/?q=how+to+compress+pdfs+with+c%23&atb=v181-1&ia=web

As far as anything that I have first-hand experience with, no I have done no work with C#.

[karlhorky]

@dabutvin is this project still being actively developed in 2023? I would be interested in the PDF compression/optimization support - is this already an accepted feature and just needs someone to take time to implement PDF compression in C#? Would guess that there would be libraries for this too... 🤔

[karlhorky]

Alternative 1: I've also created an issue in the GitHub issues for Calibre Image Actions:

• https://github.com/calibreapp/image-actions/issues/179

[karlhorky]

Alternative 2

A separate GitHub Actions workflow to do this using ghostscript (generate an access token for this and add it as a GitHub repo secret called PDF_COMPRESSION_GITHUB_TOKEN):

name: Compress PDFs
on:
  pull_request:
    # Run only when PDF files are added or changed
    paths:
      - '**.pdf'

jobs:
  build:
    # Only run on Pull Requests within the same repository, and not from forks.
    if: github.event.pull_request.head.repo.full_name == github.repository
    name: Compress PDFs
    runs-on: ubuntu-latest
    steps:
      - name: Run apt update
        run: sudo apt update
      - name: Install Ghostscript for compression
        run: sudo apt install -y ghostscript
      - name: Checkout Repo
        uses: actions/checkout@v3
        with:
          ref: ${{ github.head_ref }}

      - name: Compress PDFs
        run: |
          find . -type f -iname '*.pdf' -not -path "./node_modules/*" | while read file; do
            echo "Compressing $file"
            # Compress the file for screen using Ghostscript
            gs -sDEVICE=pdfwrite -dCompatibilityLevel=1.4 -dPDFSETTINGS=/screen -dNOPAUSE -dQUIET -dBATCH -sOutputFile="./temp.pdf" "$file"
            mv ./temp.pdf "$file"
          done
      - name: Commit files
        run: |
          git config user.email [email protected]
          git config user.name github-actions

          # Discard modified compressed files that are not more
          # than 10KB smaller than the previous version
          git ls-files -z --modified '*.pdf' | while IFS= read -r -d '' file; do
            old_size=$(git cat-file -s "HEAD:$file")
            new_size=$(wc -c < "$file" | tr -d ' ')
            if [[ $(($old_size - $new_size)) -lt 10000 ]]; then
              echo "Discarding changes to $file, compressed file size $new_size is not >10KB smaller than old size $old_size"
              git checkout -- "$file"
              continue
            fi
            echo "Compressed $file from $old_size to $new_size bytes"
          done

          git add **/*.pdf
          if [ -z "$(git status --porcelain)" ]; then
            exit 0
          fi
          git commit -m "Compress PDFs"
          git push origin HEAD:${{ github.head_ref }}
        env:
          GITHUB_TOKEN: ${{ secrets.PDF_COMPRESSION_GITHUB_TOKEN }}