Add function to load the USGS quakes data (#207)

This is a dataset from the GMT tutorial of global earthquakes.

Author: leouieda

gmt/datasets/__init__.py

@@ -1,5 +1,5 @@
 """
 Load sample data included with GMT (downloaded from the GMT cache server).
 """
-from .tutorial import load_japan_quakes
+from .tutorial import load_japan_quakes, load_usgs_quakes
 from .earth_relief import load_earth_relief

gmt/datasets/tutorial.py

@@ -36,3 +36,24 @@ def load_japan_quakes():
         "magnitude",
     ]
     return data
+
+
+def load_usgs_quakes():
+    """
+    Load a table of global earthquakes form the USGS as a pandas.Dataframe.
+
+    This is the ``@usgs_quakes_22.txt`` dataset used in the GMT tutorials.
+
+    The data are downloaded to a cache directory (usually ``~/.gmt/cache``) the
+    first time you invoke this function. Afterwards, it will load the data from
+    the cache. So you'll need an internet connection the first time around.
+
+    Returns
+    -------
+    data :  pandas.Dataframe
+        The data table. Use ``print(data.describe())`` to see the available columns.
+
+    """
+    fname = which("@usgs_quakes_22.txt", download="c")
+    data = pd.read_csv(fname)
+    return data

gmt/tests/test_datasets.py

@@ -5,7 +5,7 @@
 import numpy as np
 import numpy.testing as npt
 
-from ..datasets import load_japan_quakes, load_earth_relief
+from ..datasets import load_japan_quakes, load_earth_relief, load_usgs_quakes
 from ..exceptions import GMTInvalidInput
 
 
@@ -22,6 +22,12 @@ def test_japan_quakes():
     assert summary.loc["max", "day"] == 31
 
 
+def test_usgs_quakes():
+    "Check that the dataset loads without errors"
+    data = load_usgs_quakes()
+    assert data.shape == (1197, 22)
+
+
 def test_earth_relief_fails():
     "Make sure earth relief fails for invalid resolutions"
     resolutions = "1m 1d bla 60d 01s 03s 001m 03".split()