Fix the bug for passing text strings with numeric values (#3804)

* Add a regression text for passing strings with numeric values

* Do not try to convert string array to np.datetime64 dtype

* Fix the _to_numpy test for datetime strings

Author: seisman

pygmt/clib/conversion.py

@@ -194,15 +194,13 @@ def _to_numpy(data: Any) -> np.ndarray:
 
     array = np.ascontiguousarray(data, dtype=numpy_dtype)
 
-    # Check if a np.object_ or np.str_ array can be converted to np.datetime64.
-    if array.dtype.type in {np.object_, np.str_}:
-        with contextlib.suppress(TypeError, ValueError):
-            return np.ascontiguousarray(array, dtype=np.datetime64)
-
-    # Check if a np.object_ array can be converted to np.str_.
-    if array.dtype == np.object_:
-        with contextlib.suppress(TypeError, ValueError):
-            return np.ascontiguousarray(array, dtype=np.str_)
+    # Check if a np.object_ array can be converted to np.datetime64 or np.str_.
+    # Try np.datetime64 first then np.str_, because datetime-like objects usually have
+    # string representations.
+    if array.dtype.type == np.object_:
+        for dtype in [np.datetime64, np.str_]:
+            with contextlib.suppress(TypeError, ValueError):
+                return np.ascontiguousarray(array, dtype=dtype)
     return array
 
 

pygmt/tests/baseline/test_text_numeric_text.png.dvc

@@ -0,0 +1,5 @@
+outs:
+- md5: 97d7780216d535c0400f6117cf47ff65
+  size: 8657
+  hash: md5
+  path: test_text_numeric_text.png

pygmt/tests/test_clib_to_numpy.py

@@ -83,9 +83,6 @@ def test_to_numpy_python_types(data, expected_dtype):
 @pytest.mark.parametrize(
     "data",
     [
-        pytest.param(
-            ["2018", "2018-02", "2018-03-01", "2018-04-01T01:02:03"], id="iso8601"
-        ),
         pytest.param(
             [
                 datetime.date(2018, 1, 1),
@@ -144,6 +141,20 @@ def test_to_numpy_python_datetime(data):
     )
 
 
+def test_to_numpy_python_datetime_string():
+    """
+    Test the _to_numpy function with Python sequence of datetime strings.
+
+    Datetime strings are NOT converted to np.datetime64, but GMT can be process the
+    strings as datetime strings.
+    """
+    result = _to_numpy(["2018", "2018-02", "2018-03-01", "2018-04-01T01:02:03"])
+    _check_result(result, np.str_)
+    npt.assert_array_equal(
+        result, ["2018", "2018-02", "2018-03-01", "2018-04-01T01:02:03"]
+    )
+
+
 ########################################################################################
 # Test the _to_numpy function with NumPy arrays.
 #

pygmt/tests/test_text.py

@@ -426,6 +426,25 @@ def test_text_nonstr_text():
     return fig
 
 
+@pytest.mark.mpl_image_compare
+def test_text_numeric_text():
+    """
+    Test passing text strings that are numeric.
+
+    Regression test for https://github.com/GenericMappingTools/pygmt/issues/3803.
+    """
+    fig = Figure()
+    fig.text(
+        region=[0, 10, 0, 5],
+        projection="X10c/5c",
+        frame=True,
+        x=[1, 2, 3, 4],
+        y=[1, 2, 3, 4],
+        text=["2012", "2013", "2014", "2015"],
+    )
+    return fig
+
+
 @pytest.mark.mpl_image_compare(filename="test_text_nonascii.png")
 @pytest.mark.parametrize("encoding", ["ISOLatin1+", "Standard+"])
 def test_text_nonascii(encoding):