Support displaying 1D tensors

[Wumpf]

What

Technically two separate fixes:

• visualizability for 1 x n and n x 1 was blocked
• real single dimension tensors are now supported

Test code:

import rerun as rr
import numpy as np

rr.init("1dtensor", spawn=True)
x = np.linspace(0.0, 100.0, 100)
rr.log("n", rr.Tensor(x))
rr.log("n_x_1", rr.Tensor(np.reshape(x, (100, 1))))
rr.log("1_x_n", rr.Tensor(np.reshape(x, (1, 100))))

Before:

After first commit:

After second commit:

Screen.Recording.2024-04-08.at.11.07.18.mov

Checklist

• I have read and agree to Contributor Guide and the Code of Conduct
• I've included a screenshot or gif (if applicable)
• I have tested the web demo (if applicable):
  • Using newly built examples: rerun.io/viewer
  • Using examples from latest main build: rerun.io/viewer
  • Using full set of examples from nightly build: rerun.io/viewer
• The PR title and labels are set such as to maximize their usefulness for the next release's CHANGELOG
• If applicable, add a new check to the release checklist!

• PR Build Summary
• Recent benchmark results
• Wasm size tracking

[emilk]

1D tensors used to be visualized as bar charts - what happens to the heuristic around that now?

See e.g. docs/snippets/all/bar_chart.py

[Wumpf]

1D tensors used to be visualized as bar charts - what happens to the heuristic around that now?

not entirely true ever since we introduced indicators. You have to have a bar chart indicator in order to be picked up as a bar chart automatically
But let me run testing on bar charts and check if one can create a bar chat space view for those tensors in the test script!

[nikolausWest]

• Closes 1xN tensors are not visualizable #5839

[jleibs]

Talked through and tested this a bit and it makes sense. @Wumpf is still fixing 1 edge case with creating bar chart from an existing tensor.

[Wumpf]

Hit this very related bug, but not going to address it here or for the release in general:

• Tensors and bar charts can not be shown as images #5847