Extra layer when running two read/writes within one pipeline

[lewi0622]

File for reference and reproduction

Single pipeline

vpype \
read -a id 'C://Users//lewi0//Downloads//Demo1.svg'  occult  -i  write 'C://Users//lewi0//Downloads//Demo1_PROCESSED.svg' \
read -a stroke 'C://Users//lewi0//Downloads//Demo1_PROCESSED.svg'  \
eval "k=0" \
forlayer write "C://Users//lewi0//Downloads//Demo1_PROCESSED"%_i%".svg" end

Produces 4 files when it should produce 3:

1. Demo1_PROCESSED.svg
   1. Occulted version of the svg, with two colors and three layers
2. Demo1_PROCESSED0.svg
   1. Just the green layers
3. Demo1_PROCESSED1.svg
   1. Just the red layer
4. Demo1_PROCESSED2.svg
   1. The topmost green layer again for some reason

Two separate pipelines

vpype \
read -a id 'C://Users//lewi0//Downloads//Demo1.svg'  occult  -i  write 'C://Users//lewi0//Downloads//Demo1_PROCESSED.svg'

vpype \
read -a stroke 'C://Users//lewi0//Downloads//Demo1_PROCESSED.svg'  \
eval "k=0" \
forlayer write "C://Users//lewi0//Downloads//Demo1_PROCESSED"%_i%".svg" end

Produces 3 files correctly

[abey79]

This is working as intended.

The write command does not flush the pipeline, so the result of the initial read + occult (ie 3 layers) is still in the pipeline by the time you reach forlayer. The second read just adds (redundant) lines in layer 1 and 2. You can see that for yourself by inserting a show command:

vpype \
read -a id 'C://Users//lewi0//Downloads//Demo1.svg'  occult  -i  write 'C://Users//lewi0//Downloads//Demo1_PROCESSED.svg' \
show \
read -a stroke 'C://Users//lewi0//Downloads//Demo1_PROCESSED.svg'  \
eval "k=0" \
forlayer write "C://Users//lewi0//Downloads//Demo1_PROCESSED"%_i%".svg" end

It's possible to run the pipeline in a single line, but it requires the knowledge of which layers to merge:

vpype \
read -a id 'C://Users//lewi0//Downloads//Demo1.svg'  occult  -i  \
lmove 3 1  \
eval "k=0" \
forlayer write "C://Users//lewi0//Downloads//Demo1_PROCESSED"%_i%".svg" end

It's possible to automatically group layers by color, but the related pipeline is rather complicated IIRC (it's somewhere in the cookbook).

If you want to avoid the intermediate file, you could pipe the two command together (if that works on on windows):

vpype \
read -a id 'C://Users//lewi0//Downloads//Demo1.svg'  occult  -i  write - | vpype read -a stroke -  \
eval "k=0" \
forlayer write "C://Users//lewi0//Downloads//Demo1_PROCESSED"%_i%".svg" end

Note the use of - instead of a file name. In that case, standard in and out are used.