[viz] Use inlined byte array for heatmap png; remove VTK as dep

[EricCousineau-TRI]

As mentioned in #305

---

This change is 

[EricCousineau-TRI]

+@jwnimmer-tri Would you be up for review?
\cc @kunimatsu-tri @calderpg-tri @IanTheEngineer

Reviewable status: 0 of 7 files reviewed, all discussions resolved (waiting on @jwnimmer-tri)

[jwnimmer-tri]

Reviewable status: 0 of 7 files reviewed, 2 unresolved discussions (waiting on @EricCousineau-TRI)

a discussion (no related file):
The need for VTK in drake_ros is extremely minimal. The code only uses it to encode a constant RGB array into constant PNG bytes.

The simpler fix would be to put the PNG data into a constexpr char kHeapmapPng[] = ... and purge VTK entirely. Within the next release or two of Drake, we will have a PNG-to-memory writer public API, so you could revert to back using that pattern if you don't like the constexpr long-term.

---

---

.github/workflows/bazelized_drake_ros.yml line 57 at r1 (raw file):

        # but can't download the archive with bazel until we have the dependencies installed
        # so bazel_ros2_rules can inspect it.
        run: sudo apt-get install -y libconsole-bridge-dev libvtk9-dev

So this is adding the package to CI. How are users of drake-ros supposed to know that they need this package?

This PR seems only to unwind the deprecation console warning, but not actually solve the more acute problem of a build failure for users.

[EricCousineau-TRI]

Reviewable status: 0 of 14 files reviewed, 2 unresolved discussions (waiting on @jwnimmer-tri)

a discussion (no related file):

Previously, jwnimmer-tri (Jeremy Nimmer) wrote…

The need for VTK in drake_ros is extremely minimal. The code only uses it to encode a constant RGB array into constant PNG bytes.

The simpler fix would be to put the PNG data into a constexpr char kHeapmapPng[] = ... and purge VTK entirely. Within the next release or two of Drake, we will have a PNG-to-memory writer public API, so you could revert to back using that pattern if you don't like the constexpr long-term.

Done.

---

---

.github/workflows/bazelized_drake_ros.yml line 57 at r1 (raw file):

Previously, jwnimmer-tri (Jeremy Nimmer) wrote…

So this is adding the package to CI. How are users of drake-ros supposed to know that they need this package?

This PR seems only to unwind the deprecation console warning, but not actually solve the more acute problem of a build failure for users.

See TODO above - we can't (easily) use rosdep for this CI setup because of the current usage of tarball.
For developers, rosdep should handle it, as mentioned here:

drake-ros/drake_ros_examples/README.md

Lines 89 to 90 in 8e8f411

	# Install required dependencies.
	rosdep install --from-paths src -ryi

This is referenced by CONTRIBUTING.md.

(Moot now though)

[jwnimmer-tri]

it's so good.

Reviewed 14 of 14 files at r2, all commit messages.
Reviewable status: all files reviewed, 2 unresolved discussions (waiting on @EricCousineau-TRI)

---

drake_ros/viz/BUILD.bazel line 28 at r2 (raw file):

    hdrs = glob(
        ["*.h"],
        exclude = ["heamtap_png.raw.h"],

BTW If you make the *.inc change, this particular edit could disappear.

---

drake_ros/viz/contact_markers_system.cc line 91 at r2 (raw file):

std::vector<uint8_t> GenerateHeatmapPng() {
  return
#include "./heatmap_png.raw.h"

BTW In my experience, a more typical extension for this file would be *.inc as in #include "./heatmap_png.inc".

https://drake.mit.edu/styleguide/cppguide.html#Self_contained_Headers

[EricCousineau-TRI]

Reviewable status: 9 of 14 files reviewed, all discussions resolved (waiting on @jwnimmer-tri)

---

drake_ros/viz/BUILD.bazel line 28 at r2 (raw file):

Previously, jwnimmer-tri (Jeremy Nimmer) wrote…

BTW If you make the *.inc change, this particular edit could disappear.

Done, thanks!

---

drake_ros/viz/contact_markers_system.cc line 91 at r2 (raw file):

Previously, jwnimmer-tri (Jeremy Nimmer) wrote…

BTW In my experience, a more typical extension for this file would be *.inc as in #include "./heatmap_png.inc".

https://drake.mit.edu/styleguide/cppguide.html#Self_contained_Headers

Done

[jwnimmer-tri]

Reviewed 5 of 5 files at r3, all commit messages.
Reviewable status: complete! all files reviewed, all discussions resolved (waiting on @EricCousineau-TRI)

[EricCousineau-TRI]

CI failure not immediately clear, but assuming it's once again "out of disk space". Merging.