Remove unzip and zip and use ZipArchives.jl instead

[nhz2]

Fixes #14 by using ZipArchives.jl

ZipArchives.jl is a new package I made to read Zarr data, and it works for PPTX as well.

Lastly, in this PR I removed the ability to use template directories, only files are allowed.

[matthijscox-asml]

You put some serious effort into this!

I don't know if @jvieyras and/or @xvries have some time to review as well?

[codecov]

Codecov Report

Attention: Patch coverage is 97.40260% with 2 lines in your changes are missing coverage. Please review.

Project coverage is 95.84%. Comparing base (1f1adda) to head (4f175f1).
Report is 1 commits behind head on main.

❗ Current head 4f175f1 differs from pull request most recent head 2a195f3. Consider uploading reports for the commit 2a195f3 to get more accurate results

Files	Patch %	Lines
src/write.jl	96.87%	2 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main      #31      +/-   ##
==========================================
+ Coverage   94.62%   95.84%   +1.21%     
==========================================
  Files          10       10              
  Lines         521      505      -16     
==========================================
- Hits          493      484       -9     
+ Misses         28       21       -7     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[nhz2]

I updated this PR to the latest version of main.

@matthijscox-asml @jaakkor2 This PR also resolves the problems in #47. With this PR there is no need to mess with file permissions or directories because everything stays as zip files.

[matthijscox-asml]

@matthijscox-asml @jaakkor2 This PR also resolves the problems in #47. With this PR there is no need to mess with file permissions or directories because everything stays as zip files.

That would be another good reason to go with your changes :)

But you do not expect issues with editing the zip files, if the zip file is read-only from the Pkg.add ?

[nhz2]

In this PR the only thing I do with the template file is read, I never use the cp function.

To avoid changing an existing zip file, I keep track of two separate zip file objects when writing the presentation.

One is template_reader = ZipBufferReader(read(template_path))
This is a read-only ZipReader of the template.

The other is ZipWriter(temp_out; own_io=true) do w
This is an append-only ZipWriter for the new file.

w starts out completely empty, and files that need to be different from the template are added to it in:

update_presentation_state!(p, template_reader)
write_relationships!(w, p)
write_presentation!(w, p)
write_slides!(w, p, template_reader)
write_shapes!(w, p)
update_table_style!(w, template_reader)
add_contenttypes!(w, template_reader)

Then only after all of the files that needed to be edited were written, do I copy over all other files from the template that are not yet in w.

# copy over any files from the template
# but don't overwrite any files in w
for i in zip_nentries(template_reader):-1:1
    local name = zip_name(template_reader, i)
    if !endswith(name,"/")
        if !zip_name_collision(w, name)
            local compress = zip_iscompressed(template_reader, i)
            zip_data = zip_readentry(template_reader, i)
            zip_newfile(w, name; compress)
            write(w, zip_data)
            zip_commitfile(w)
        end
    end
end

In this way no zip file entry needs to be edited or deleted, only read and append operations are needed.

[matthijscox-asml]

okay, i agree this looks good! hmm, I've decided to go with it.

do you want to do a release 1.0 of ZipArchives.jl first before merging this?

[nhz2]

Sure, I'll make sure to do that. Thanks for reviewing this.

[matthijscox-asml]

I was ill for a while so this took a bit longer than expected.

I see that there's a silly error in the docs test that needs a fix (although I don't know why it suddenly needs DataFrames.DataFrame instead of DataFrame, but it's easy enough to change that)

If you want you can also update the Project.toml and bump it to v0.7.0, then I'll register a new version after merging this PR.