Remove the deprecated load_usgs_quakes function (deprecated since v0.6.0)

[willschlitzer]

This replaces load_usgs_quakes with the internal function _load_usgs_quakes. It also adds additional checks to test_usgs_quakes.

Addresses: #2302

Reminders

• Run make format and make check to make sure the code follows the style guide.
• Add tests for new features or tests that would have caught the bug that you're fixing.
• Add new public functions/methods/classes to doc/api/index.rst.
• Write detailed docstrings for all functions/methods.
• If wrapping a new module, open a 'Wrap new GMT module' issue and submit reasonably-sized PRs.
• If adding new functionality, add an example to docstrings or tutorials.
• Use underscores (not hyphens) in names of Python files and directories.

Slash Commands

You can write slash commands (/command) in the first line of a comment to perform
specific operations. Supported slash commands are:

• /format: automatically format and lint the code
• /test-gmt-dev: run full tests on the latest GMT development version

[seisman]

I just realized that this sample dataset is not used in any examples. Should we remove it?

[yvonnefroehlich]

I just realized that this sample dataset is not used in any examples. Should we remove it?

Hm. Not sure, as there seem to be other sample datasets which are currently also not used in any example, e.g., maunaloa_co2, hotspots, or mars_shape.

[seisman]

I just realized that this sample dataset is not used in any examples. Should we remove it?

Hm. Not sure, as there seem to be other sample datasets which are currently also not used in any example, e.g., maunaloa_co2, hotspots, or mars_shape.

• mars_shape is used in pygmt/src/sphinterpolate.py,
• maunaloa_co2 is used in pygmt/tests/test_filter1d.py
• hotspots was added in Add function to import hotspot dataset #1386, and it seems there was a plan to use it in a sphdistance example.

[yvonnefroehlich]

I just realized that this sample dataset is not used in any examples. Should we remove it?

Hm. Not sure, as there seem to be other sample datasets which are currently also not used in any example, e.g., maunaloa_co2, hotspots, or mars_shape.

• mars_shape is used in pygmt/src/sphinterpolate.py,
• maunaloa_co2 is used in pygmt/tests/test_filter1d.py
• hotspots was added in Add function to import hotspot dataset #1386, and it seems there was a plan to use it in a sphdistance example.

Ah, I see. I just considered the "real" gallery examples and tutorials...

[willschlitzer]

I'm fine with removing load_usgs_quakes. I don't remember it, but it looks like I planned to use it with sphdistance and that took a long time to merge, so I forgot about it.

[weiji14]

A quick search on GitHub shows it being used before at https://github.com/GenericMappingTools/try-gmt-python/blob/master/try-gmt-python.ipynb. But it's 4 years old now.

How much of a burden is it to keep the function maintained? Maybe we could just keep it in case?

[willschlitzer]

It's not too much of an issue to keep it maintained. I just assumed that @seisman suggested removing it to avoid the precedent that all GMT remote files have an associated PyGMT function to load them.

[seisman]

It's not too much of an issue to keep it maintained. I just assumed that @seisman suggested removing it to avoid the precedent that all GMT remote files have an associated PyGMT function to load them.

I'm OK with maintaining this function, and I believe this dataset for global earthquakes can be used in examples in the future.

My major concern is that the returned data table contains too many columns and most of them are text strings and are unlikely used in any PyGMT examples. I suggest only keeping the "time", "latitude", "longitude", "depth" and "mag", similar to the data table returned by _load_japan_quakes.

[willschlitzer]

It's not too much of an issue to keep it maintained. I just assumed that @seisman suggested removing it to avoid the precedent that all GMT remote files have an associated PyGMT function to load them.

I'm OK with maintaining this function, and I believe this dataset for global earthquakes can be used in examples in the future.

My major concern is that the returned data table contains too many columns and most of them are text strings and are unlikely used in any PyGMT examples. I suggest only keeping the "time", "latitude", "longitude", "depth" and "mag", similar to the data table returned by _load_japan_quakes.

I've never studied seismology, so I'm not sure how important information that pertains to the errors is to someone hoping to use this dataset. But I don't see the need to load this dataset with a separate function if we're going to remove quite a few of the columns. If it's the same columns as _load_japan_quakes, we might as well just use the Japan quakes dataset for any example/test functions. I don't feel strongly about removing this function, as someone may already be using it in their examples, but I don't wish to pare down the information in the returned dataset.

Looking at the USGS Earthquake API, we could probably create an example that incorporates getting a JSON file from the API and using it to create a plot/perform operations. I think that would be a better example of a real application of earthquake data instead of usgs_quakes_22.txt.

[seisman]

It's not too much of an issue to keep it maintained. I just assumed that @seisman suggested removing it to avoid the precedent that all GMT remote files have an associated PyGMT function to load them.

I'm OK with maintaining this function, and I believe this dataset for global earthquakes can be used in examples in the future.
My major concern is that the returned data table contains too many columns and most of them are text strings and are unlikely used in any PyGMT examples. I suggest only keeping the "time", "latitude", "longitude", "depth" and "mag", similar to the data table returned by _load_japan_quakes.

I've never studied seismology, so I'm not sure how important information that pertains to the errors is to someone hoping to use this dataset. But I don't see the need to load this dataset with a separate function if we're going to remove quite a few of the columns. If it's the same columns as _load_japan_quakes, we might as well just use the Japan quakes dataset for any example/test functions. I don't feel strongly about removing this function, as someone may already be using it in their examples, but I don't wish to pare down the information in the returned dataset.

OK, I think it's fine to keep this function unchanged.

Looking at the USGS Earthquake API, we could probably create an example that incorporates getting a JSON file from the API and using it to create a plot/perform operations. I think that would be a better example of a real application of earthquake data instead of usgs_quakes_22.txt.

Seismologists usually use ObsPy's get_events() function to get earthquakes from USGS or other seismic data centers. The function returns a Catalog object which contains event information. This is related to the ObsPy integration in #967.

[seisman]

Need to remove "usgs_quakes": load_usgs_quakes at line 75.