Missing graphics in meta-article

[asmacdo]

Following README instructions I re-generated the meta article on typhon.

I'm not sure if some of the graphics need to be pulled down via git-annex or if we just have a rendering error

Full article attached article-1.pdf

[TheChymera]

This means that the slices from the PDF diffs have not been generated. Are you sure you downloaded the article pdf data? Please post the output of:

 ls outputs/*/*pdf -lah

Following https://github.com/con/opfvta-replication-2023/commit/f1399d942501f2914f43d2218d636cca9b52a0f9 the PDF data should be auto-fetched. If re-running after having built the article with an older repo version, use make article -B or run make article-clean beforehand.

[asmacdo]

Looks like make container-article always does a clean, so that isn't the problem. (Btw, make article-clean only works with latex installed, so we might need to check that the container is doing this successfully?)

$ ls outputs/*/*pdf -lah

lrwxrwxrwx 1 asmacdo asmacdo 136 Sep 21 16:22 outputs/original/article.pdf -> ../../.git/annex/objects/f9/Pv/MD5E-s21852357--cda3904485734ab5c089894db1659fa0.pdf/MD5E-s21852357--cda3904485734ab5c089894db1659fa0.pdf
lrwxrwxrwx 1 asmacdo asmacdo 136 Sep 21 16:22 outputs/podman_20230906053037/article.pdf -> ../../.git/annex/objects/qW/zV/MD5E-s21529647--17365d13a65885d2fce3d65d1c96c9f7.pdf/MD5E-s21529647--17365d13a65885d2fce3d65d1c96c9f7.pdf
lrwxrwxrwx 1 asmacdo asmacdo 136 Sep 21 16:22 outputs/podman_20230908162849/article.pdf -> ../../.git/annex/objects/80/v6/MD5E-s21545132--cc38b0844b952a39cd8d4cbf96605105.pdf/MD5E-s21545132--cc38b0844b952a39cd8d4cbf96605105.pdf
lrwxrwxrwx 1 asmacdo asmacdo 136 Sep 21 16:22 outputs/podman_20230920004219/article.pdf -> ../../.git/annex/objects/PJ/0J/MD5E-s21532977--fd1bd9e0b630500be88a52f3880f7a96.pdf/MD5E-s21532977--fd1bd9e0b630500be88a52f3880f7a96.pdf
lrwxrwxrwx 1 asmacdo asmacdo 136 Sep 21 16:22 outputs/podman_20230920004416/article.pdf -> ../../.git/annex/objects/gx/Pj/MD5E-s21514941--990e16b8bb32fb21c65c71a73646301d.pdf/MD5E-s21514941--990e16b8bb32fb21c65c71a73646301d.pdf
-rw-r--r-- 1 asmacdo asmacdo 21M Sep 23 02:44 outputs/podman_20230922123546/article.pdf
lrwxrwxrwx 1 asmacdo asmacdo 136 Sep 21 16:22 outputs/singularity_20230904214036/article.pdf -> ../../.git/annex/objects/k1/6x/MD5E-s21487275--5ab7da033717219cf2b87dde21317777.pdf/MD5E-s21487275--5ab7da033717219cf2b87dde21317777.pdf
lrwxrwxrwx 1 asmacdo asmacdo 136 Sep 21 16:22 outputs/singularity_20230905040926/article.pdf -> ../../.git/annex/objects/16/fV/MD5E-s21487288--ad7b66af092f14b60d45b31ba5e09135.pdf/MD5E-s21487288--ad7b66af092f14b60d45b31ba5e09135.pdf
lrwxrwxrwx 1 asmacdo asmacdo 136 Sep 21 16:22 outputs/singularity_20230905204435/article.pdf -> ../../.git/annex/objects/wz/g9/MD5E-s21487302--9f76897f47716f6adda8c778917b920d.pdf/MD5E-s21487302--9f76897f47716f6adda8c778917b920d.pdf
lrwxrwxrwx 1 asmacdo asmacdo 136 Sep 21 16:22 outputs/singularity_20230906012936/article.pdf -> ../../.git/annex/objects/8v/q0/MD5E-s21551881--c92d2d5f2941e0a5a4eb25fca28a92ec.pdf/MD5E-s21551881--c92d2d5f2941e0a5a4eb25fca28a92ec.pdf
lrwxrwxrwx 1 asmacdo asmacdo 136 Sep 21 16:22 outputs/singularity_20230908122618/article.pdf -> ../../.git/annex/objects/K6/vv/MD5E-s21370913--dd4358584c78e366bc7940ed7c0965a8.pdf/MD5E-s21370913--dd4358584c78e366bc7940ed7c0965a8.pdf

I'm not in a clean git state though. After executing the analysis, commit steps aren't included in the readme. Is it safe to just commit these partial results so I can rebase onto your recent changes?

TODO: We should include commit points into the readme.

or

**TODO (asmacdo preference): We should change the make targets to datalad run

It does look like this is just an environment problem, since CI is building the pdf with images successfully.

[yarikoptic]

Use ls -lL to see if you actually fetched the data

[asmacdo]

$ ls -lLah outputs/*/*.pdf

-r--r--r-- 1 asmacdo asmacdo 21M Sep 25 10:36 outputs/original/article.pdf
-r--r--r-- 1 asmacdo asmacdo 21M Sep 25 10:33 outputs/podman_20230906053037/article.pdf
-r--r--r-- 1 asmacdo asmacdo 21M Sep 25 10:31 outputs/podman_20230908162849/article.pdf
-r--r--r-- 1 asmacdo asmacdo 21M Sep 25 10:35 outputs/podman_20230920004219/article.pdf
-r--r--r-- 1 asmacdo asmacdo 21M Sep 25 10:33 outputs/podman_20230920004416/article.pdf
-rw-r--r-- 1 asmacdo asmacdo 21M Sep 23 02:44 outputs/podman_20230922123546/article.pdf
-r--r--r-- 1 asmacdo asmacdo 21M Sep 25 10:32 outputs/singularity_20230904214036/article.pdf
-r--r--r-- 1 asmacdo asmacdo 21M Sep 25 10:37 outputs/singularity_20230905040926/article.pdf
-r--r--r-- 1 asmacdo asmacdo 21M Sep 25 10:32 outputs/singularity_20230905204435/article.pdf
-r--r--r-- 1 asmacdo asmacdo 21M Sep 25 10:34 outputs/singularity_20230906012936/article.pdf
-r--r--r-- 1 asmacdo asmacdo 21M Sep 25 10:35 outputs/singularity_20230908122618/article.pdf
asmacdo@typhon:~/tmp/opfvta-2023-9-21/opfvta-replication-2023$ 

[TheChymera]

After coworking session debugging it appears:

• The data is there
• Figures get generated with the correct aspect ratio (thank to it being set explicitly in ghostscipr)
• However, figures are blank
• Though, the actual PDF diffs from which the figures are sliced by ghostscipt, are not blank

I conclude that this is probably some sort of ghostscirpt issue. The artifact on github CI is also correct (i.e. it shows figure contents), so probably we're trying to find something which is present on both github CI and on my machine, but not on @asmacdo 's account on typhon. This is very strange but we need to figure it out. @yarikoptic can you try it locally as well?

[TheChymera]

It could be that ghostscript has trouble operating without a display, though even so it is strange that it would work locally on my machine both in and outside of the container, as well as on GitHub CI.

[TheChymera]

I think I have it, so we dynamically diff the latest 4 reexecutions, but for the figures an older one was hardcoded, which is why, in view of @asmacdo 's new reexecutions, the older pdf-diff which we were attempting to slice with ghostscript was no longer generated.