Stirling-Tools
switch images to alpine
License Agreement for Contributions
By submitting this pull request, I acknowledge and agree that my contributions will be included in Stirling-PDF and that they can be relicensed in the future under MPL 2.0 (Mozilla Public License Version 2.0) license.
(This does not change the general open-source nature of Stirling-PDF, simply moving from one license to another license)
Wow this is awesome work I could never get alpine images working due to inexperienced docker skills
Have you been able to test these and the functionality of python/libre office etc?
Just ran a quick run of all 3 docker files
Ultra lite seems to be missing java
Cool with that it seems they all load to the homepage
I will do some tests on OCR and conversions just to see if anything else wrong in parallel with your testing etc
From my tests (of the full image) everything works, except one thing; If I want to convert a pdf to word, I see in the docker logs: Error: source file could not be loaded and it downloads an empty zip file
doing docker compose i can expose the port
but when running in docker desktop i get
Without the Expose 8080
What is the normal expectation for docker files
yeah works via CMD for me as well, just find it odd that docker desktop UI doesnt work with it or let you define ports, so was wondering if this is something non standard
I've never used EXPOSE in a Dockerfile since it does nothing (usefull): https://docs.docker.com/engine/reference/builder/#expose
I've tried again, it downloads now the converted document, but it is unuseable (pdf => html):
Something is defo off with the outputs when compared to old version
(You can test here)
https://pdf.adminforge.de/pdf-to-word
pdf to work on your version gives me
vs
seems its just loaded the PDF file bytes into .doc format not actual converting
i see libreoffice-core is not avaible, i tried just 'libreoffice' and get a slightly better conversion now.. all 3 pages are overlayed onto single page and rotation is not kept.. so something weird.. but clearly something different about the libre package
So right now the conversions are not happening in a way which i am happy with, it seems to throw everything onto single page, not sure why... Until then we wont be able to merge this. However super appreciate your work, will try see whats up with libreoffice
it seems to be a bug in libreoffice, so it will also affect debian/ubuntu when they will update in the future
If it's the version and not general package can we pull the older version?
v7.3.7.2-r0 and earlier of libreoffice works, everything after v7.5.5.2-r0 not, so the bug seems to be added with v7.4 or v7.5
https://packages.ubuntu.com/jammy/libreoffice currently the dokcer image uses v7.3
using alpine:3.17 instead of :latest makes it work for lite and normal, ultra lite can stay at latest, but 3.17 is two versions old...
https://mirror1.hs-esslingen.de/pub/Mirrors/tdf/libreoffice/src/bugs-changelog-tag-libreoffice-7.6.4.1-release-7.6.4.1.log "PDF: Conversion of pdf to docx or doc collapses all content onto one page (tdf#157589) [Kevin Suo]"
using alpine:3.17 instead of :latest makes it work for lite and normal, ultra lite can stay at latest, but 3.17 is two versions old...
Based on this and you using 3.19 is this PR now blocked on resolution of alpine updating its package manager?
using alpine:3.17 instead of :latest makes it work for lite and normal, ultra lite can stay at latest, but 3.17 is two versions old...
Based on this and you using 3.19 is this PR now blocked on resolution of alpine updating its package manager?
yes blocked until they fix libreoffice, but from what I know, they are fast in fixing this
should now work, I will create a PR when a fixed libreoffice verion from outside edge can be used
btw wondering if i can get your advice, I am trying to get calibre installed on alpine
I get
apk add --no-cache calibre --repository=http://dl-cdn.alpinelinux.org/alpine/edge/testing
fetch http://dl-cdn.alpinelinux.org/alpine/edge/testing/x86_64/APKINDEX.tar.gz
fetch https://dl-cdn.alpinelinux.org/alpine/v3.18/main/x86_64/APKINDEX.tar.gz
fetch https://dl-cdn.alpinelinux.org/alpine/v3.18/community/x86_64/APKINDEX.tar.gz
ERROR: unable to select packages:
py3-pyqt6-webengine (no such package):
required by: calibre-7.3.0-r0[py3-pyqt6-webengine]
qt6-qtwebengine (no such package):
required by: calibre-7.3.0-r0[qt6-qtwebengine]
so:libicui18n.so.74 (no such package):
required by: calibre-7.3.0-r0[so:libicui18n.so.74]
so:libicuuc.so.74 (no such package):
required by: calibre-7.3.0-r0[so:libicuuc.so.74]
so:libpodofo.so.2 (no such package):
required by: calibre-7.3.0-r0[so:libpodofo.so.2]
As a solution i added
http://dl-cdn.alpinelinux.org/alpine/edge/main
http://dl-cdn.alpinelinux.org/alpine/edge/community
to
/etc/apk/repositories
is this okay?
I would do this: /etc/apk/repositories:
https://dl-cdn.alpinelinux.org/alpine/v3.19/main
https://dl-cdn.alpinelinux.org/alpine/v3.19/community
@testing https://dl-cdn.alpinelinux.org/alpine/edge/main
@testing https://dl-cdn.alpinelinux.org/alpine/edge/community
@testing https://dl-cdn.alpinelinux.org/alpine/edge/testing
apk add --no-cache calibre@testing
can I ask when this PR will be merged?-
Once i figure out calibre and wkhtmltopdf I added them as custom installs which users can do post docker image install https://github.com/Stirling-Tools/Stirling-PDF/pull/682/files#diff-8e6e46e251822ef45509ec061b133c5788a5719a0670ac3158b49dbb58e3dab6R69
If you can make it conditional in docker file via some arg in the init scripts or changed to match in the java file i linked then sure!