Bump poetry from 1.1.11 to 1.2.0rc2 and refresh other conda/python dependencies

[weiji14]

Refresh to poetry 1.2.0rc2 and taking the opportunity to bump up lots of other dependencies like GMT and RAPIDS AI libraries. Supersedes #326.

TODO:

• Update environment.yml file:
  • Bump poetry to 1.2.0rc2, Pip to 22.3 (3bb50b5)
  • Bump gmt from 6.2.0 to 6.3.0 and RAPIDS AI libraries from 21.10 to 22.02 (21e0e99, 94dcff6)
• Update poetry.lock file:
  • Update dependencies in lock file by changing the ^ to == in pyproject.toml, running rm poetry.lock && poetry lock, change == back to ^, and then running poetry lock --no-update to update the hash
  • Fix broken tests due to dependency updates

[weiji14]

Segmentation fault on CI is caused by import cudf, bug reported upstream at rapidsai/cudf#11941.

[weiji14]

Next problem may or may not have something to do with dask. Crashing on https://github.com/weiji14/deepicedrain/actions/runs/3276952377/jobs/5393616432#step:9:48, deepicedrain/tests/test_subglacial_lake_crossover.py::test_subglacial_lake_anomalies somewhere along this chunk:

deepicedrain/deepicedrain/vizplots.py

Lines 413 to 441 in 0f405ba

	for track1_track2, indexes in tqdm.tqdm(crossovers.indices.items()):
	df_ = df.loc[indexes].sort_values(by=time_var)
	dhdt, _ = np.polyfit(
	x=df_[time_var].astype(np.int64), y=df_[elev_var], deg=1
	) * (365.25 * 24 * 60 * 60 * 1_000_000_000)
	if abs(dhdt) > elev_filter: # Plot only > 1 metre height change
	track1, track2 = track1_track2.split("x")
	fig.plot(
	x=df_[time_var],
	y=df_[elev_var],
	zvalue=dhdt,
	style="c0.1c",
	cmap=True,
	pen="thin+z",
	)
	# Plot line connecting points
	fig.plot(
	x=df_[time_var],
	y=df_[elev_var],
	zvalue=dhdt,
	pen=f"faint,+z,-",
	cmap=True,
	)
	# Get x, y and color (elev) for inset map
	if outline_points:
	_x, _y = df_.iloc[0][["x", "y"]]
	xover["x"].append(_x)
	xover["y"].append(_y)
	xover["dhdt"].append(dhdt)