Plot square or cube by default for OGR/GMT files with Point/MultiPoint types

[yohaimagen]

To keep consistency with default style plotting of geoDataFrame with Point or MultiPoint geometry see #1405 adding the same behavior for .gmt files that are saved from geoDataFrame with .to_file('some/path', driver="OGR_GMT", mode="w")

Fixes #1373

Reminders

• Add the same if statement to plot3d.py

• Add tests for new features or tests that would have caught the bug that you're fixing.

[yohaimagen]

I am falling here on 5 tests:
when running make test I am getting
FAILED ../pygmt/tests/test_basemap.py::test_basemap_winkel_tripel
FAILED ../pygmt/tests/test_config.py::test_config_map_grid_cross_size
FAILED ../pygmt/tests/test_config.py::test_config_map_grid_pen
FAILED ../pygmt/tests/test_config.py::test_config_map_tick_length
FAILED ../pygmt/tests/test_config.py::test_config_map_tick_pen

all 5 tests do not involve plot.py which I modified and do involve basemap.py, also if I'm removing the added code to plot.py and run make test I'm getting the same test failure.
someone has a clue why do I get those results?

[seisman]

I am falling here on 5 tests:
when running make test I am getting
FAILED ../pygmt/tests/test_basemap.py::test_basemap_winkel_tripel
FAILED ../pygmt/tests/test_config.py::test_config_map_grid_cross_size
FAILED ../pygmt/tests/test_config.py::test_config_map_grid_pen
FAILED ../pygmt/tests/test_config.py::test_config_map_tick_length
FAILED ../pygmt/tests/test_config.py::test_config_map_tick_pen

all 5 tests do not involve plot.py which I modified and do involve basemap.py, also if I'm removing the added code to plot.py and run make test I'm getting the same test failure.
someone has a clue why do I get those results?

Maybe try to run dvc pull again to see if these images are up-to-date?

[yohaimagen]

Running dvc pull does not solve the issue. Also if I am making a fresh new development environment as described and run make test I am getting the same failed tests(on the main branch).
ping @weiji14 maybe you know how to solve this issue?

[weiji14]

Running dvc pull does not solve the issue. Also if I am making a fresh new development environment as described and run make test I am getting the same failed tests(on the main branch).
ping @weiji14 maybe you know how to solve this issue?

There is only 1 test failing on CI (test_legend_entries at https://github.com/GenericMappingTools/pygmt/pull/1438/checks?check_run_id=3302244951#step:11:735) instead of 5. The error message is as follows:

        if "S" in kwargs and kwargs["S"][0] in "vV" and direction is not None:
            extra_arrays.extend(direction)
        if "S" not in kwargs and kind == "file":
>           with open(data, "r") as file:
E           FileNotFoundError: [Errno 2] No such file or directory: '@Table_5_11.txt'

The problem is that your code tries to open the @Table_5_11.txt file, but because this is a remote GMT file from the server, Python can't find it. The easiest way to pass that test would be to add another if condition (if data.endswith(".gmt")) so that only OGR/GMT files with the .gmt extension are checked.

However, users may want to plot OGR/GMT files from https://github.com/GenericMappingTools/gmtserver-admin/tree/master/cache, so we'll need to handle files like @quakes.gmt too... Can you try to come up with another solution for that case?

[yohaimagen]

We can use os.path.exists to check if the file is on the system.

1. I don't know what is the import police can I import os at plot.py?
2. That means that we do not support remote GMT files from the server and doing something like

fig = pygmt.Figure()
fig.basemap(projection="M5c", frame="a", region=[0, 30, 0, 30])
fig.plot(data="@quakes.gmt")
fig.show()

will plot a line instead of squares.

as I understand from the code the import of files from the server is done on the GMT side so we can't support it anyway.

[yohaimagen]

Another option that we can use is to do something like that:

try:
    with open(data, 'r') as file:
        line = file.readline()
        if "@GMULTIPOINT" in line or "@GPOINT" in line:
            kwargs["S"] = "s0.2c"
except FileNotFoundError:
        pass

We do not import os in that solution, but we still do not support remote files, as I mentioned that's up to GMT.

[weiji14]

I like the try-except style, could you make the change please?

Another thing we could do (if we want to support remote OGR/GMT files) is to use pygmt.which to locate the local path of the @something.gmt file, and have the open function read that. Oh and by the way, it's ok to put import os in the plot.py file since it's a standard Python module.

[seisman]

Another thing we could do (if we want to support remote OGR/GMT files) is to use pygmt.which to locate the local path of the @something.gmt file, and have the open function read that.

I think this way is better.

It's not only about supporting remote files. GMT users can store their data (e.g., faults.gmt, plate_boundary.gmt) in any directory and set the environment variable GMT_DATADIR to that directory, GMT will also look for files in that directory. So users can use

fig.plot("faults.gmt")

even though the file faults.gmt is not in the current directory and is not a remote file.

[yohaimagen]

I changed the code to use pygmt.which and kept the try-except phrase so we will have consistency on the error thrown in case that really the file does not exists

[seisman]

@yohaimagen Need to resolve the conflicts first.

[yohaimagen]

O.K I added the encoding to the open commands in the tests and the # pylint: disable=too-many-locals to the functions.

And we passing the style check :)

Let me know if there is something else to do over here.

[weiji14]

Just some more typos, otherwise good for a final round of reviews! We'll merge in 24 hours if there are no other issues 😁

[yohaimagen]

We have one error: the docs were unable to build on windows, where it falls on something unrelated to this pull request.
It is falling while trying to read the iris data at scatterplot3d.py

df = pd.read_csv("https://github.com/mwaskom/seaborn-data/raw/master/iris.csv")

[seisman]

Looks good, except a few minor edits

[seisman]

We have one error: the docs were unable to build on windows, where it falls on something unrelated to this pull request.
It is falling while trying to read the iris data at scatterplot3d.py

df = pd.read_csv("https://github.com/mwaskom/seaborn-data/raw/master/iris.csv")

This error can be ignored, because we already know that sometimes GMT API on Windows may crash.

[michaelgrund]

Except the minor suggestions this PR looks good to me!

[weiji14]

Appreciate your help @yohaimagen, thanks again! 🥳

[yohaimagen]

Thanks to all three reviewers for reviewing my code.