From 1f32a7cd861c14b737a36f6af6f0019fcb750912 Mon Sep 17 00:00:00 2001
From: Wei Ji <23487320+weiji14@users.noreply.github.com>
Date: Sat, 30 Dec 2023 19:01:01 +1300
Subject: [PATCH 01/25] Check passing pyarrow.array with string type to
 pygmt.text

Ensure that pyarrow.array objects with string type can be read by pygmt.text.
---
 pygmt/tests/test_text.py | 24 +++++++++++++++++++++---
 1 file changed, 21 insertions(+), 3 deletions(-)

diff --git a/pygmt/tests/test_text.py b/pygmt/tests/test_text.py
index 509dcd75cb1..bc28e180ef8 100644
--- a/pygmt/tests/test_text.py
+++ b/pygmt/tests/test_text.py
@@ -8,6 +8,12 @@
 from pygmt import Figure
 from pygmt.exceptions import GMTCLibError, GMTInvalidInput
 from pygmt.helpers import GMTTempFile
+from pygmt.helpers.testing import skip_if_no
+
+try:
+    import pyarrow as pa
+except ImportError:
+    pa = None
 
 TEST_DATA_DIR = os.path.join(os.path.dirname(__file__), "data")
 POINTS_DATA = os.path.join(TEST_DATA_DIR, "points.txt")
@@ -47,8 +53,20 @@ def test_text_single_line_of_text(region, projection):
 
 
 @pytest.mark.benchmark
-@pytest.mark.mpl_image_compare
-def test_text_multiple_lines_of_text(region, projection):
+@pytest.mark.mpl_image_compare(filename="test_text_multiple_lines_of_text.png")
+@pytest.mark.parametrize(
+    "array_func",
+    [
+        list,
+        pytest.param(np.array, id="numpy"),
+        pytest.param(
+            getattr(pa, "array", None),
+            marks=skip_if_no(package="pyarrow"),
+            id="pyarrow",
+        ),
+    ],
+)
+def test_text_multiple_lines_of_text(region, projection, array_func):
     """
     Place multiple lines of text at their respective x, y locations.
     """
@@ -58,7 +76,7 @@ def test_text_multiple_lines_of_text(region, projection):
         projection=projection,
         x=[1.2, 1.6],
         y=[0.6, 0.3],
-        text=["This is a line of text", "This is another line of text"],
+        text=array_func(["This is a line of text", "This is another line of text"]),
     )
     return fig
 

From 4c4e064c4b4d7fd3edbfda2bb75414824063238c Mon Sep 17 00:00:00 2001
From: Wei Ji <23487320+weiji14@users.noreply.github.com>
Date: Sat, 30 Dec 2023 19:49:53 +1300
Subject: [PATCH 02/25] Check passing pyarrow.array with string type to
 virtualfile_from_vectors

---
 pygmt/tests/test_clib_virtualfiles.py | 28 ++++++++++++++++++++++-----
 1 file changed, 23 insertions(+), 5 deletions(-)

diff --git a/pygmt/tests/test_clib_virtualfiles.py b/pygmt/tests/test_clib_virtualfiles.py
index 92a0f98b85b..a98df1fbb09 100644
--- a/pygmt/tests/test_clib_virtualfiles.py
+++ b/pygmt/tests/test_clib_virtualfiles.py
@@ -12,8 +12,14 @@
 from pygmt import clib
 from pygmt.exceptions import GMTCLibError, GMTInvalidInput
 from pygmt.helpers import GMTTempFile
+from pygmt.helpers.testing import skip_if_no
 from pygmt.tests.test_clib import mock
 
+try:
+    import pyarrow as pa
+except ImportError:
+    pa = None
+
 TEST_DATA_DIR = os.path.join(os.path.dirname(__file__), "data")
 POINTS_DATA = os.path.join(TEST_DATA_DIR, "points.txt")
 
@@ -210,16 +216,28 @@ def test_virtualfile_from_vectors(dtypes):
             assert output == expected
 
 
-@pytest.mark.parametrize("dtype", [str, object])
-def test_virtualfile_from_vectors_one_string_or_object_column(dtype):
+@pytest.mark.parametrize(
+    ("array_func", "dtype"),
+    [
+        pytest.param(np.array, {"dtype": str}, id="str"),
+        pytest.param(np.array, {"dtype": object}, id="object"),
+        pytest.param(
+            getattr(pa, "array", None),
+            {},  # pa.string()
+            marks=skip_if_no(package="pyarrow"),
+            id="pyarrow",
+        ),
+    ],
+)
+def test_virtualfile_from_vectors_one_string_or_object_column(array_func, dtype):
     """
-    Test passing in one column with string or object dtype into virtual file
-    dataset.
+    Test passing in one column with string (numpy/pyarrow) or object (numpy)
+    dtype into virtual file dataset.
     """
     size = 5
     x = np.arange(size, dtype=np.int32)
     y = np.arange(size, size * 2, 1, dtype=np.int32)
-    strings = np.array(["a", "bc", "defg", "hijklmn", "opqrst"], dtype=dtype)
+    strings = array_func(["a", "bc", "defg", "hijklmn", "opqrst"], **dtype)
     with clib.Session() as lib:
         with lib.virtualfile_from_vectors(x, y, strings) as vfile:
             with GMTTempFile() as outfile:

From d379e4667c5f8fcc907a3a56543ae734b8c605eb Mon Sep 17 00:00:00 2001
From: Wei Ji <23487320+weiji14@users.noreply.github.com>
Date: Fri, 11 Oct 2024 17:27:06 +1300
Subject: [PATCH 03/25] Enable passing pyarrow.StringArray to
 clib.Session.put_strings

Convert a pyarrow.StringArray via a Python list to a ctypes array in the strings_to_ctypes_array function. Updated docstrings and type hints in `clib.Session.put_strings` method and `clib.conversion.strings_to_ctypes_array` function. Added two parametrized unit tests to ensure that pyarrow.StringArray can be passed into the clib methods.
---
 pygmt/clib/conversion.py              | 18 ++++++++---
 pygmt/clib/session.py                 | 43 ++++++++++++++++-----------
 pygmt/tests/test_clib_put_strings.py  | 25 ++++++++++++++--
 pygmt/tests/test_clib_virtualfiles.py |  2 +-
 4 files changed, 63 insertions(+), 25 deletions(-)

diff --git a/pygmt/clib/conversion.py b/pygmt/clib/conversion.py
index 0739d767567..e924fad5332 100644
--- a/pygmt/clib/conversion.py
+++ b/pygmt/clib/conversion.py
@@ -9,6 +9,10 @@
 import numpy as np
 from pygmt.exceptions import GMTInvalidInput
 
+try:
+    import pyarrow as pa
+except ImportError:
+    pa = None
 
 def dataarray_to_matrix(grid):
     """
@@ -263,14 +267,15 @@ def sequence_to_ctypes_array(
     return (ctype * size)(*sequence)
 
 
-def strings_to_ctypes_array(strings: Sequence[str]) -> ctp.Array:
+def strings_to_ctypes_array(strings: Sequence[str] | pa.StringArray) -> ctp.Array:
     """
-    Convert a sequence (e.g., a list) of strings into a ctypes array.
+    Convert a sequence (e.g., a list) of strings or a pyarrow.StringArray into a ctypes
+    array.
 
     Parameters
     ----------
     strings
-        A sequence of strings.
+        A sequence of strings or a pyarrow.StringArray.
 
     Returns
     -------
@@ -286,7 +291,12 @@ def strings_to_ctypes_array(strings: Sequence[str]) -> ctp.Array:
     >>> [s.decode() for s in ctypes_array]
     ['first', 'second', 'third']
     """
-    return (ctp.c_char_p * len(strings))(*[s.encode() for s in strings])
+    try:
+        bytes_string_list = [s.encode() for s in strings]
+    except AttributeError:  # 'pyarrow.StringScalar' object has no attribute 'encode'
+        # Convert pyarrow.StringArray to Python list first
+        bytes_string_list = [s.encode() for s in strings.to_pylist()]
+    return (ctp.c_char_p * len(strings))(*bytes_string_list)
 
 
 def array_to_datetime(array):
diff --git a/pygmt/clib/session.py b/pygmt/clib/session.py
index a96ca84baae..959baac1540 100644
--- a/pygmt/clib/session.py
+++ b/pygmt/clib/session.py
@@ -34,6 +34,11 @@
     tempfile_from_image,
 )
 
+try:
+    import pyarrow as pa
+except ImportError:
+    pa = None
+
 FAMILIES = [
     "GMT_IS_DATASET",  # Entity is a data table
     "GMT_IS_GRID",  # Entity is a grid
@@ -936,39 +941,43 @@ def put_vector(self, dataset, column, vector):
                 f"in column {column} of dataset."
             )
 
-    def put_strings(self, dataset, family, strings):
+    def put_strings(
+        self,
+        dataset: ctp.c_void_p,
+        family: Literal["GMT_IS_VECTOR", "GMT_IS_MATRIX"],
+        strings: Sequence[str] | pa.StringArray,
+    ):
         """
-        Attach a numpy 1-D array of dtype str as a column on a GMT dataset.
+        Attach a 1-D numpy array of dtype str or pyarrow.StringArray as a column on a
+        GMT dataset.
 
-        Use this function to attach string type numpy array data to a GMT
-        dataset and pass it to GMT modules. Wraps ``GMT_Put_Strings``.
+        Use this function to attach string type array data to a GMT dataset and pass it
+        to GMT modules. Wraps ``GMT_Put_Strings``.
 
-        The dataset must be created by :meth:`pygmt.clib.Session.create_data`
-        first.
+        The dataset must be created by :meth:`pygmt.clib.Session.create_data` first.
 
         .. warning::
-            The numpy array must be C contiguous in memory. If it comes from a
-            column slice of a 2-D array, for example, you will have to make a
-            copy. Use :func:`numpy.ascontiguousarray` to make sure your vector
-            is contiguous (it won't copy if it already is).
+            The array must be C contiguous in memory. If it comes from a column slice of
+            a 2-D array, for example, you will have to make a copy. Use
+            :func:`numpy.ascontiguousarray` to make sure your vector is contiguous (it
+            won't copy if it already is).
 
         Parameters
         ----------
-        dataset : :class:`ctypes.c_void_p`
+        dataset
             The ctypes void pointer to a ``GMT_Dataset``. Create it with
             :meth:`pygmt.clib.Session.create_data`.
-        family : str
+        family
             The family type of the dataset. Can be either ``GMT_IS_VECTOR`` or
             ``GMT_IS_MATRIX``.
-        strings : numpy 1-D array
-            The array that will be attached to the dataset. Must be a 1-D C
-            contiguous array.
+        strings
+            The array that will be attached to the dataset. Must be a 1-D C contiguous
+            array.
 
         Raises
         ------
         GMTCLibError
-            If given invalid input or ``GMT_Put_Strings`` exits with
-            status != 0.
+            If given invalid input or ``GMT_Put_Strings`` exits with status != 0.
         """
         c_put_strings = self.get_libgmt_func(
             "GMT_Put_Strings",
diff --git a/pygmt/tests/test_clib_put_strings.py b/pygmt/tests/test_clib_put_strings.py
index 92287be3e29..0db47ca4a8a 100644
--- a/pygmt/tests/test_clib_put_strings.py
+++ b/pygmt/tests/test_clib_put_strings.py
@@ -8,12 +8,31 @@
 from pygmt import clib
 from pygmt.exceptions import GMTCLibError
 from pygmt.helpers import GMTTempFile
+from pygmt.helpers.testing import skip_if_no
+
+try:
+    import pyarrow as pa
+except ImportError:
+    pa = None
 
 
 @pytest.mark.benchmark
-def test_put_strings():
+@pytest.mark.parametrize(
+    ("array_func", "dtype"),
+    [
+        pytest.param(np.array, {"dtype": str}, id="str"),
+        pytest.param(
+            getattr(pa, "array", None),
+            {"type": pa.string()},
+            marks=skip_if_no(package="pyarrow"),
+            id="pyarrow",
+        ),
+    ],
+)
+def test_put_strings(array_func, dtype):
     """
-    Check that assigning a numpy array of dtype str to a dataset works.
+    Check that assigning a numpy array of dtype str, or a pyarrow.StringArray to a
+    dataset works.
     """
     with clib.Session() as lib:
         dataset = lib.create_data(
@@ -24,7 +43,7 @@ def test_put_strings():
         )
         x = np.array([1, 2, 3, 4, 5], dtype=np.int32)
         y = np.array([6, 7, 8, 9, 10], dtype=np.int32)
-        strings = np.array(["a", "bc", "defg", "hijklmn", "opqrst"], dtype=str)
+        strings = array_func(["a", "bc", "defg", "hijklmn", "opqrst"], **dtype)
         lib.put_vector(dataset, column=lib["GMT_X"], vector=x)
         lib.put_vector(dataset, column=lib["GMT_Y"], vector=y)
         lib.put_strings(
diff --git a/pygmt/tests/test_clib_virtualfiles.py b/pygmt/tests/test_clib_virtualfiles.py
index 0f70c06cb85..2d9b66b47d6 100644
--- a/pygmt/tests/test_clib_virtualfiles.py
+++ b/pygmt/tests/test_clib_virtualfiles.py
@@ -238,7 +238,7 @@ def test_virtualfile_from_vectors(dtypes):
         pytest.param(np.array, {"dtype": object}, id="object"),
         pytest.param(
             getattr(pa, "array", None),
-            {},  # pa.string()
+            {"type": pa.string()},
             marks=skip_if_no(package="pyarrow"),
             id="pyarrow",
         ),

From cfda386da577d0e2d2ab31a133f0e70770722393 Mon Sep 17 00:00:00 2001
From: Wei Ji <23487320+weiji14@users.noreply.github.com>
Date: Fri, 11 Oct 2024 18:16:59 +1300
Subject: [PATCH 04/25] Use "string" instead of pyarrow.string() in case
 pyarrow not installed

Fixes `AttributeError: 'NoneType' object has no attribute 'string'`
---
 pygmt/tests/test_clib_put_strings.py  | 2 +-
 pygmt/tests/test_clib_virtualfiles.py | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/pygmt/tests/test_clib_put_strings.py b/pygmt/tests/test_clib_put_strings.py
index 0db47ca4a8a..41f0dc921d4 100644
--- a/pygmt/tests/test_clib_put_strings.py
+++ b/pygmt/tests/test_clib_put_strings.py
@@ -23,7 +23,7 @@
         pytest.param(np.array, {"dtype": str}, id="str"),
         pytest.param(
             getattr(pa, "array", None),
-            {"type": pa.string()},
+            {"type": "string"},  # pa.string()
             marks=skip_if_no(package="pyarrow"),
             id="pyarrow",
         ),
diff --git a/pygmt/tests/test_clib_virtualfiles.py b/pygmt/tests/test_clib_virtualfiles.py
index 2d9b66b47d6..f0160d8bb82 100644
--- a/pygmt/tests/test_clib_virtualfiles.py
+++ b/pygmt/tests/test_clib_virtualfiles.py
@@ -238,7 +238,7 @@ def test_virtualfile_from_vectors(dtypes):
         pytest.param(np.array, {"dtype": object}, id="object"),
         pytest.param(
             getattr(pa, "array", None),
-            {"type": pa.string()},
+            {"type": "string"},  # pa.string()
             marks=skip_if_no(package="pyarrow"),
             id="pyarrow",
         ),

From 0a6cda5285f68b607e0fc49436dc76c84f00e598 Mon Sep 17 00:00:00 2001
From: Wei Ji <23487320+weiji14@users.noreply.github.com>
Date: Fri, 11 Oct 2024 18:17:42 +1300
Subject: [PATCH 05/25] Try to fix type hints

---
 pygmt/clib/conversion.py | 7 ++++++-
 pygmt/clib/session.py    | 7 ++++++-
 2 files changed, 12 insertions(+), 2 deletions(-)

diff --git a/pygmt/clib/conversion.py b/pygmt/clib/conversion.py
index e924fad5332..092ccab11bb 100644
--- a/pygmt/clib/conversion.py
+++ b/pygmt/clib/conversion.py
@@ -9,11 +9,16 @@
 import numpy as np
 from pygmt.exceptions import GMTInvalidInput
 
+StringArrayTypes = Sequence[str]
+
 try:
     import pyarrow as pa
+
+    StringArrayTypes |= pa.StringArray
 except ImportError:
     pa = None
 
+
 def dataarray_to_matrix(grid):
     """
     Transform an xarray.DataArray into a data 2-D array and metadata.
@@ -267,7 +272,7 @@ def sequence_to_ctypes_array(
     return (ctype * size)(*sequence)
 
 
-def strings_to_ctypes_array(strings: Sequence[str] | pa.StringArray) -> ctp.Array:
+def strings_to_ctypes_array(strings: StringArrayTypes) -> ctp.Array:
     """
     Convert a sequence (e.g., a list) of strings or a pyarrow.StringArray into a ctypes
     array.
diff --git a/pygmt/clib/session.py b/pygmt/clib/session.py
index 959baac1540..fe55d241536 100644
--- a/pygmt/clib/session.py
+++ b/pygmt/clib/session.py
@@ -34,11 +34,16 @@
     tempfile_from_image,
 )
 
+StringArrayTypes = Sequence[str]
+
 try:
     import pyarrow as pa
+
+    StringArrayTypes |= pa.StringArray
 except ImportError:
     pa = None
 
+
 FAMILIES = [
     "GMT_IS_DATASET",  # Entity is a data table
     "GMT_IS_GRID",  # Entity is a grid
@@ -945,7 +950,7 @@ def put_strings(
         self,
         dataset: ctp.c_void_p,
         family: Literal["GMT_IS_VECTOR", "GMT_IS_MATRIX"],
-        strings: Sequence[str] | pa.StringArray,
+        strings: StringArrayTypes,
     ):
         """
         Attach a 1-D numpy array of dtype str or pyarrow.StringArray as a column on a

From f59f93cf86d9f48ba4a8d7c310f73f5716391539 Mon Sep 17 00:00:00 2001
From: Wei Ji <23487320+weiji14@users.noreply.github.com>
Date: Fri, 11 Oct 2024 18:32:10 +1300
Subject: [PATCH 06/25] Add np.ndarray to StringArrayTypes and fix/ignore
 remaining type errors

---
 pygmt/clib/conversion.py | 4 ++--
 pygmt/clib/session.py    | 7 +++----
 2 files changed, 5 insertions(+), 6 deletions(-)

diff --git a/pygmt/clib/conversion.py b/pygmt/clib/conversion.py
index 092ccab11bb..c9eb070d1d7 100644
--- a/pygmt/clib/conversion.py
+++ b/pygmt/clib/conversion.py
@@ -9,7 +9,7 @@
 import numpy as np
 from pygmt.exceptions import GMTInvalidInput
 
-StringArrayTypes = Sequence[str]
+StringArrayTypes = Sequence[str] | np.ndarray
 
 try:
     import pyarrow as pa
@@ -300,7 +300,7 @@ def strings_to_ctypes_array(strings: StringArrayTypes) -> ctp.Array:
         bytes_string_list = [s.encode() for s in strings]
     except AttributeError:  # 'pyarrow.StringScalar' object has no attribute 'encode'
         # Convert pyarrow.StringArray to Python list first
-        bytes_string_list = [s.encode() for s in strings.to_pylist()]
+        bytes_string_list = [s.encode() for s in strings.to_pylist()]  # type: ignore[union-attr]
     return (ctp.c_char_p * len(strings))(*bytes_string_list)
 
 
diff --git a/pygmt/clib/session.py b/pygmt/clib/session.py
index fe55d241536..611b164715c 100644
--- a/pygmt/clib/session.py
+++ b/pygmt/clib/session.py
@@ -34,7 +34,7 @@
     tempfile_from_image,
 )
 
-StringArrayTypes = Sequence[str]
+StringArrayTypes = Sequence[str] | np.ndarray
 
 try:
     import pyarrow as pa
@@ -1005,9 +1005,8 @@ def put_strings(
             self.session_pointer, family_int, dataset, strings_pointer
         )
         if status != 0:
-            raise GMTCLibError(
-                f"Failed to put strings of type {strings.dtype} into dataset"
-            )
+            dtype = strings.dtype if hasattr(strings, "dtype") else type(strings)
+            raise GMTCLibError(f"Failed to put strings of type {dtype} into dataset")
 
     def put_matrix(self, dataset, matrix, pad=0):
         """

From 17c1e9ca28acbb9f35223b0bb770a8dc8bb75843 Mon Sep 17 00:00:00 2001
From: Wei Ji <23487320+weiji14@users.noreply.github.com>
Date: Fri, 11 Oct 2024 19:20:19 +1300
Subject: [PATCH 07/25] Move StringArrayTypes to pygmt/_typing.py

Also improve docstring of strings_to_ctypes_array function to mention np.ndarray as supported input.
---
 pygmt/_typing.py         | 10 ++++++++++
 pygmt/clib/conversion.py | 16 ++++------------
 pygmt/clib/session.py    | 11 +----------
 3 files changed, 15 insertions(+), 22 deletions(-)

diff --git a/pygmt/_typing.py b/pygmt/_typing.py
index bbc7d596c65..4a57c3c7678 100644
--- a/pygmt/_typing.py
+++ b/pygmt/_typing.py
@@ -2,7 +2,17 @@
 Type aliases for type hints.
 """
 
+import contextlib
+import importlib
+from collections.abc import Sequence
 from typing import Literal
 
+import numpy as np
+
 # Anchor codes
 AnchorCode = Literal["TL", "TC", "TR", "ML", "MC", "MR", "BL", "BC", "BR"]
+
+# String array types
+StringArrayTypes = Sequence[str] | np.ndarray
+with contextlib.suppress(ImportError):
+    StringArrayTypes |= importlib.import_module(name="pyarrow").StringArray
diff --git a/pygmt/clib/conversion.py b/pygmt/clib/conversion.py
index f4c8bb50fd2..5a3f708af1a 100644
--- a/pygmt/clib/conversion.py
+++ b/pygmt/clib/conversion.py
@@ -9,17 +9,9 @@
 import numpy as np
 import pandas as pd
 from packaging.version import Version
+from pygmt._typing import StringArrayTypes
 from pygmt.exceptions import GMTInvalidInput
 
-StringArrayTypes = Sequence[str] | np.ndarray
-
-try:
-    import pyarrow as pa
-
-    StringArrayTypes |= pa.StringArray
-except ImportError:
-    pa = None
-
 
 def dataarray_to_matrix(grid):
     """
@@ -292,13 +284,13 @@ def sequence_to_ctypes_array(
 
 def strings_to_ctypes_array(strings: StringArrayTypes) -> ctp.Array:
     """
-    Convert a sequence (e.g., a list) of strings or a pyarrow.StringArray into a ctypes
-    array.
+    Convert a sequence (e.g., a list) or numpy.ndarray of strings or a
+    pyarrow.StringArray into a ctypes array.
 
     Parameters
     ----------
     strings
-        A sequence of strings or a pyarrow.StringArray.
+        A sequence of strings, a numpy.ndarray of str dtype, or a pyarrow.StringArray.
 
     Returns
     -------
diff --git a/pygmt/clib/session.py b/pygmt/clib/session.py
index af5678aef09..f53a5af2cf5 100644
--- a/pygmt/clib/session.py
+++ b/pygmt/clib/session.py
@@ -17,6 +17,7 @@
 import numpy as np
 import pandas as pd
 import xarray as xr
+from pygmt._typing import StringArrayTypes
 from pygmt.clib.conversion import (
     array_to_datetime,
     dataarray_to_matrix,
@@ -34,16 +35,6 @@
     tempfile_from_image,
 )
 
-StringArrayTypes = Sequence[str] | np.ndarray
-
-try:
-    import pyarrow as pa
-
-    StringArrayTypes |= pa.StringArray
-except ImportError:
-    pa = None
-
-
 FAMILIES = [
     "GMT_IS_DATASET",  # Entity is a data table
     "GMT_IS_GRID",  # Entity is a grid

From 0105d64740a4f45d9a6337d28c4ebdda1dd7a795 Mon Sep 17 00:00:00 2001
From: Wei Ji <23487320+weiji14@users.noreply.github.com>
Date: Fri, 11 Oct 2024 19:29:13 +1300
Subject: [PATCH 08/25] Add pyarrow to docs CI

So that the pyarrow.StringArray type hint will show up on Readthedocs and the PyGMT docs page.
---
 .github/workflows/ci_docs.yml | 1 +
 ci/requirements/docs.yml      | 1 +
 2 files changed, 2 insertions(+)

diff --git a/.github/workflows/ci_docs.yml b/.github/workflows/ci_docs.yml
index 57cf9c8c75c..c0c8c8235e8 100644
--- a/.github/workflows/ci_docs.yml
+++ b/.github/workflows/ci_docs.yml
@@ -103,6 +103,7 @@ jobs:
             contextily
             geopandas<1.0
             ipython
+            pyarrow
             rioxarray
             make
             pip
diff --git a/ci/requirements/docs.yml b/ci/requirements/docs.yml
index 8214dbac8fe..6ae1595fe28 100644
--- a/ci/requirements/docs.yml
+++ b/ci/requirements/docs.yml
@@ -16,6 +16,7 @@ dependencies:
     - contextily
     - geopandas<1.0
     - ipython
+    - pyarrow
     - rioxarray
     # Development dependencies (general)
     - make

From 3ad0c86c20cd1079e2de99a417fcce59fb7a5370 Mon Sep 17 00:00:00 2001
From: Wei Ji <23487320+weiji14@users.noreply.github.com>
Date: Fri, 11 Oct 2024 20:40:04 +1300
Subject: [PATCH 09/25] Use np.asarray to convert pa.StringArray instead of
 .to_pylist()

Faster for longer string arrays.

Co-authored-by: Dongdong Tian <seisman.info@gmail.com>
---
 pygmt/clib/conversion.py | 7 +------
 1 file changed, 1 insertion(+), 6 deletions(-)

diff --git a/pygmt/clib/conversion.py b/pygmt/clib/conversion.py
index 5a3f708af1a..5cd50cc0734 100644
--- a/pygmt/clib/conversion.py
+++ b/pygmt/clib/conversion.py
@@ -306,12 +306,7 @@ def strings_to_ctypes_array(strings: StringArrayTypes) -> ctp.Array:
     >>> [s.decode() for s in ctypes_array]
     ['first', 'second', 'third']
     """
-    try:
-        bytes_string_list = [s.encode() for s in strings]
-    except AttributeError:  # 'pyarrow.StringScalar' object has no attribute 'encode'
-        # Convert pyarrow.StringArray to Python list first
-        bytes_string_list = [s.encode() for s in strings.to_pylist()]  # type: ignore[union-attr]
-    return (ctp.c_char_p * len(strings))(*bytes_string_list)
+    return (ctp.c_char_p * len(strings))(*[s.encode() for s in np.asarray(strings)])
 
 
 def array_to_datetime(array):

From 4bea288e34aee5cab520fa267e6ef9e9828cfcbe Mon Sep 17 00:00:00 2001
From: Wei Ji <23487320+weiji14@users.noreply.github.com>
Date: Fri, 11 Oct 2024 20:48:21 +1300
Subject: [PATCH 10/25] Update note to say that PyArrow string types are now
 supported

Still need to work on Duration and GeoArrow geometry types.
---
 doc/install.md | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

diff --git a/doc/install.md b/doc/install.md
index c64369c5b5a..c627cae853a 100644
--- a/doc/install.md
+++ b/doc/install.md
@@ -107,9 +107,10 @@ The following are optional dependencies:
 :::{note}
 If you have [PyArrow](https://arrow.apache.org/docs/python/index.html) installed, PyGMT
 does have some initial support for `pandas.Series` and `pandas.DataFrame` objects with
-Apache Arrow-backed arrays. Specifically, only uint/int/float and date32/date64 dtypes
-are supported for now. Support for string Arrow dtypes is still a work in progress.
-For more details, see [issue #2800](https://github.com/GenericMappingTools/pygmt/issues/2800).
+Apache Arrow-backed arrays. Specifically, only uint/int/float, date32/date64 and string
+types are supported for now. Support for Duration types and GeoArrow geometry types is
+still a work in progress. For more details, see
+[issue #2800](https://github.com/GenericMappingTools/pygmt/issues/2800).
 :::
 
 ## Installing GMT and other dependencies

From 371174a61e811a3c13200e953eb0d75ee5a35ef6 Mon Sep 17 00:00:00 2001
From: Wei Ji <23487320+weiji14@users.noreply.github.com>
Date: Fri, 11 Oct 2024 20:53:06 +1300
Subject: [PATCH 11/25] Add back pytest.mark.benchmark marker

Accidentally removed from `test_virtualfile_from_vectors_one_string_or_object_column` during merge conflict handling.
---
 pygmt/tests/test_clib_virtualfiles.py | 1 +
 1 file changed, 1 insertion(+)

diff --git a/pygmt/tests/test_clib_virtualfiles.py b/pygmt/tests/test_clib_virtualfiles.py
index f0160d8bb82..a7f449cbf32 100644
--- a/pygmt/tests/test_clib_virtualfiles.py
+++ b/pygmt/tests/test_clib_virtualfiles.py
@@ -231,6 +231,7 @@ def test_virtualfile_from_vectors(dtypes):
             assert output == expected
 
 
+@pytest.mark.benchmark
 @pytest.mark.parametrize(
     ("array_func", "dtype"),
     [

From b5887304bc955e4f60708e61456c29a4c0cad5d4 Mon Sep 17 00:00:00 2001
From: Wei Ji <23487320+weiji14@users.noreply.github.com>
Date: Fri, 11 Oct 2024 20:57:21 +1300
Subject: [PATCH 12/25] Add intersphinx link for pyarrow

---
 doc/conf.py | 1 +
 1 file changed, 1 insertion(+)

diff --git a/doc/conf.py b/doc/conf.py
index fc26641bd83..2024646c0d5 100644
--- a/doc/conf.py
+++ b/doc/conf.py
@@ -85,6 +85,7 @@
     "contextily": ("https://contextily.readthedocs.io/en/stable/", None),
     "geopandas": ("https://geopandas.org/en/stable/", None),
     "numpy": ("https://numpy.org/doc/stable/", None),
+    "pyarrow": ("https://arrow.apache.org/docs/", None),
     "python": ("https://docs.python.org/3/", None),
     "pandas": ("https://pandas.pydata.org/pandas-docs/stable/", None),
     "rasterio": ("https://rasterio.readthedocs.io/en/stable/", None),

From faf2065c47b15eafbb79218346c0a8176a79b80c Mon Sep 17 00:00:00 2001
From: Wei Ji <23487320+weiji14@users.noreply.github.com>
Date: Wed, 6 Nov 2024 14:38:07 +1300
Subject: [PATCH 13/25] Apply suggestions from code review

Co-authored-by: Dongdong Tian <seisman.info@gmail.com>
---
 pygmt/clib/session.py | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/pygmt/clib/session.py b/pygmt/clib/session.py
index f53a5af2cf5..5a0eb003a89 100644
--- a/pygmt/clib/session.py
+++ b/pygmt/clib/session.py
@@ -934,7 +934,7 @@ def put_vector(self, dataset, column, vector):
     def put_strings(
         self,
         dataset: ctp.c_void_p,
-        family: Literal["GMT_IS_VECTOR", "GMT_IS_MATRIX"],
+        family: str,
         strings: StringArrayTypes,
     ):
         """
@@ -955,8 +955,8 @@ def put_strings(
         Parameters
         ----------
         dataset
-            The ctypes void pointer to a ``GMT_Dataset``. Create it with
-            :meth:`pygmt.clib.Session.create_data`.
+            The ctypes void pointer to a ``GMT_VECTOR``/``GMT_MATRIX`` container. 
+            Create it with :meth:`pygmt.clib.Session.create_data`.
         family
             The family type of the dataset. Can be either ``GMT_IS_VECTOR`` or
             ``GMT_IS_MATRIX``.

From 9fd77dce62b10089670979b2a669faefd430b4cd Mon Sep 17 00:00:00 2001
From: Wei Ji <23487320+weiji14@users.noreply.github.com>
Date: Wed, 6 Nov 2024 15:01:26 +1300
Subject: [PATCH 14/25] format

---
 pygmt/clib/session.py | 5 +----
 1 file changed, 1 insertion(+), 4 deletions(-)

diff --git a/pygmt/clib/session.py b/pygmt/clib/session.py
index b795ec04efc..1248a09c6cb 100644
--- a/pygmt/clib/session.py
+++ b/pygmt/clib/session.py
@@ -1001,10 +1001,7 @@ def put_vector(self, dataset: ctp.c_void_p, column: int, vector: np.ndarray):
             raise GMTCLibError(msg)
 
     def put_strings(
-        self,
-        dataset: ctp.c_void_p,
-        family: str,
-        strings: StringArrayTypes,
+        self, dataset: ctp.c_void_p, family: str, strings: StringArrayTypes
     ):
         """
         Attach a 1-D numpy array of dtype str or pyarrow.StringArray as a column on a

From a927202acfcdb9a236a765534079bb821b798189 Mon Sep 17 00:00:00 2001
From: Wei Ji <23487320+weiji14@users.noreply.github.com>
Date: Fri, 15 Nov 2024 14:20:24 +1300
Subject: [PATCH 15/25] Revert "Enable passing pyarrow.StringArray to
 clib.Session.put_strings"

This reverts commit d379e4667c5f8fcc907a3a56543ae734b8c605eb.
---
 pygmt/clib/conversion.py              | 11 ++++----
 pygmt/clib/session.py                 | 40 +++++++++++++--------------
 pygmt/tests/test_clib_put_strings.py  | 25 ++---------------
 pygmt/tests/test_clib_virtualfiles.py |  2 +-
 4 files changed, 28 insertions(+), 50 deletions(-)

diff --git a/pygmt/clib/conversion.py b/pygmt/clib/conversion.py
index 87ebf17618b..68cddd63549 100644
--- a/pygmt/clib/conversion.py
+++ b/pygmt/clib/conversion.py
@@ -11,7 +11,6 @@
 import pandas as pd
 import xarray as xr
 from packaging.version import Version
-from pygmt._typing import StringArrayTypes
 from pygmt.exceptions import GMTInvalidInput
 
 
@@ -279,15 +278,15 @@ def sequence_to_ctypes_array(
     return (ctype * size)(*sequence)
 
 
-def strings_to_ctypes_array(strings: StringArrayTypes) -> ctp.Array:
+def strings_to_ctypes_array(strings: Sequence[str] | np.ndarray) -> ctp.Array:
     """
-    Convert a sequence (e.g., a list) or numpy.ndarray of strings or a
-    pyarrow.StringArray into a ctypes array.
+    Convert a sequence (e.g., a list) of strings or numpy.ndarray of strings into a
+    ctypes array.
 
     Parameters
     ----------
     strings
-        A sequence of strings, a numpy.ndarray of str dtype, or a pyarrow.StringArray.
+        A sequence of strings, or a numpy.ndarray of str dtype.
 
     Returns
     -------
@@ -303,7 +302,7 @@ def strings_to_ctypes_array(strings: StringArrayTypes) -> ctp.Array:
     >>> [s.decode() for s in ctypes_array]
     ['first', 'second', 'third']
     """
-    return (ctp.c_char_p * len(strings))(*[s.encode() for s in np.asarray(strings)])
+    return (ctp.c_char_p * len(strings))(*[s.encode() for s in strings])
 
 
 def array_to_datetime(array: Sequence[Any] | np.ndarray) -> np.ndarray:
diff --git a/pygmt/clib/session.py b/pygmt/clib/session.py
index 1248a09c6cb..7229ea170d8 100644
--- a/pygmt/clib/session.py
+++ b/pygmt/clib/session.py
@@ -16,7 +16,6 @@
 import numpy as np
 import pandas as pd
 import xarray as xr
-from pygmt._typing import StringArrayTypes
 from pygmt.clib.conversion import (
     array_to_datetime,
     dataarray_to_matrix,
@@ -1000,40 +999,39 @@ def put_vector(self, dataset: ctp.c_void_p, column: int, vector: np.ndarray):
             )
             raise GMTCLibError(msg)
 
-    def put_strings(
-        self, dataset: ctp.c_void_p, family: str, strings: StringArrayTypes
-    ):
+    def put_strings(self, dataset: ctp.c_void_p, family: str, strings: np.ndarray):
         """
-        Attach a 1-D numpy array of dtype str or pyarrow.StringArray as a column on a
-        GMT dataset.
+        Attach a numpy 1-D array of dtype str as a column on a GMT dataset.
 
-        Use this function to attach string type numpy array data to a GMT dataset and
-        pass it to GMT modules. Wraps ``GMT_Put_Strings``.
+        Use this function to attach string type numpy array data to a GMT
+        dataset and pass it to GMT modules. Wraps ``GMT_Put_Strings``.
 
-        The dataset must be created by :meth:`pygmt.clib.Session.create_data` first.
+        The dataset must be created by :meth:`pygmt.clib.Session.create_data`
+        first.
 
         .. warning::
-            The numpy array must be C contiguous in memory. If it comes from a column
-            slice of a 2-D array, for example, you will have to make a copy. Use
-            :func:`numpy.ascontiguousarray` to make sure your vector is contiguous (it
-            won't copy if it already is).
+            The numpy array must be C contiguous in memory. If it comes from a
+            column slice of a 2-D array, for example, you will have to make a
+            copy. Use :func:`numpy.ascontiguousarray` to make sure your vector
+            is contiguous (it won't copy if it already is).
 
         Parameters
         ----------
-        dataset
-            The ctypes void pointer to a ``GMT_VECTOR``/``GMT_MATRIX`` data container.
-            Create it with :meth:`pygmt.clib.Session.create_data`.
-        family
+        dataset : :class:`ctypes.c_void_p`
+            The ctypes void pointer to a ``GMT_Dataset``. Create it with
+            :meth:`pygmt.clib.Session.create_data`.
+        family : str
             The family type of the dataset. Can be either ``GMT_IS_VECTOR`` or
             ``GMT_IS_MATRIX``.
-        strings
-            The array that will be attached to the dataset. Must be a 1-D C contiguous
-            array.
+        strings : numpy 1-D array
+            The array that will be attached to the dataset. Must be a 1-D C
+            contiguous array.
 
         Raises
         ------
         GMTCLibError
-            If given invalid input or ``GMT_Put_Strings`` exits with a non-zero status.
+            If given invalid input or ``GMT_Put_Strings`` exits with
+            status != 0.
         """
         c_put_strings = self.get_libgmt_func(
             "GMT_Put_Strings",
diff --git a/pygmt/tests/test_clib_put_strings.py b/pygmt/tests/test_clib_put_strings.py
index ccc0189c780..bd2687bb7c0 100644
--- a/pygmt/tests/test_clib_put_strings.py
+++ b/pygmt/tests/test_clib_put_strings.py
@@ -8,31 +8,12 @@
 from pygmt import clib
 from pygmt.exceptions import GMTCLibError
 from pygmt.helpers import GMTTempFile
-from pygmt.helpers.testing import skip_if_no
-
-try:
-    import pyarrow as pa
-except ImportError:
-    pa = None
 
 
 @pytest.mark.benchmark
-@pytest.mark.parametrize(
-    ("array_func", "dtype"),
-    [
-        pytest.param(np.array, {"dtype": np.str_}, id="str"),
-        pytest.param(
-            getattr(pa, "array", None),
-            {"type": "string"},  # pa.string()
-            marks=skip_if_no(package="pyarrow"),
-            id="pyarrow",
-        ),
-    ],
-)
-def test_put_strings(array_func, dtype):
+def test_put_strings():
     """
-    Check that assigning a numpy array of dtype str, or a pyarrow.StringArray to a
-    dataset works.
+    Check that assigning a numpy array of dtype str to a dataset works.
     """
     with clib.Session() as lib:
         dataset = lib.create_data(
@@ -43,7 +24,7 @@ def test_put_strings(array_func, dtype):
         )
         x = np.array([1, 2, 3, 4, 5], dtype=np.int32)
         y = np.array([6, 7, 8, 9, 10], dtype=np.int32)
-        strings = array_func(["a", "bc", "defg", "hijklmn", "opqrst"], **dtype)
+        strings = np.array(["a", "bc", "defg", "hijklmn", "opqrst"], dtype=str)
         lib.put_vector(dataset, column=lib["GMT_X"], vector=x)
         lib.put_vector(dataset, column=lib["GMT_Y"], vector=y)
         lib.put_strings(
diff --git a/pygmt/tests/test_clib_virtualfiles.py b/pygmt/tests/test_clib_virtualfiles.py
index 42284b08c78..6277f80e153 100644
--- a/pygmt/tests/test_clib_virtualfiles.py
+++ b/pygmt/tests/test_clib_virtualfiles.py
@@ -153,7 +153,7 @@ def test_open_virtual_file():
         pytest.param(np.array, {"dtype": np.object_}, id="object"),
         pytest.param(
             getattr(pa, "array", None),
-            {"type": "string"},  # pa.string()
+            {},  # pa.string()
             marks=skip_if_no(package="pyarrow"),
             id="pyarrow",
         ),

From 7b002488b7c46b9c793c6e522ae87312a020b2c0 Mon Sep 17 00:00:00 2001
From: Wei Ji <23487320+weiji14@users.noreply.github.com>
Date: Fri, 15 Nov 2024 14:27:58 +1300
Subject: [PATCH 16/25] Reduce diff from messy revert handling

---
 pygmt/clib/session.py                | 31 ++++++++++++++--------------
 pygmt/tests/test_clib_put_strings.py |  2 +-
 2 files changed, 16 insertions(+), 17 deletions(-)

diff --git a/pygmt/clib/session.py b/pygmt/clib/session.py
index 7229ea170d8..b8d6ff6feb1 100644
--- a/pygmt/clib/session.py
+++ b/pygmt/clib/session.py
@@ -1001,31 +1001,30 @@ def put_vector(self, dataset: ctp.c_void_p, column: int, vector: np.ndarray):
 
     def put_strings(self, dataset: ctp.c_void_p, family: str, strings: np.ndarray):
         """
-        Attach a numpy 1-D array of dtype str as a column on a GMT dataset.
+        Attach a 1-D numpy array of dtype str as a column on a GMT dataset.
 
-        Use this function to attach string type numpy array data to a GMT
-        dataset and pass it to GMT modules. Wraps ``GMT_Put_Strings``.
+        Use this function to attach string type numpy array data to a GMT dataset and
+        pass it to GMT modules. Wraps ``GMT_Put_Strings``.
 
-        The dataset must be created by :meth:`pygmt.clib.Session.create_data`
-        first.
+        The dataset must be created by :meth:`pygmt.clib.Session.create_data` first.
 
         .. warning::
-            The numpy array must be C contiguous in memory. If it comes from a
-            column slice of a 2-D array, for example, you will have to make a
-            copy. Use :func:`numpy.ascontiguousarray` to make sure your vector
-            is contiguous (it won't copy if it already is).
+            The numpy array must be C contiguous in memory. If it comes from a column
+            slice of a 2-D array, for example, you will have to make a copy. Use
+            :func:`numpy.ascontiguousarray` to make sure your vector is contiguous (it
+            won't copy if it already is).
 
         Parameters
         ----------
-        dataset : :class:`ctypes.c_void_p`
-            The ctypes void pointer to a ``GMT_Dataset``. Create it with
-            :meth:`pygmt.clib.Session.create_data`.
-        family : str
+        dataset
+            The ctypes void pointer to a ``GMT_VECTOR``/``GMT_MATRIX`` data container.
+            Create it with :meth:`pygmt.clib.Session.create_data`.
+        family
             The family type of the dataset. Can be either ``GMT_IS_VECTOR`` or
             ``GMT_IS_MATRIX``.
-        strings : numpy 1-D array
-            The array that will be attached to the dataset. Must be a 1-D C
-            contiguous array.
+        strings
+            The array that will be attached to the dataset. Must be a 1-D C contiguous
+            array.
 
         Raises
         ------
diff --git a/pygmt/tests/test_clib_put_strings.py b/pygmt/tests/test_clib_put_strings.py
index bd2687bb7c0..81b27b06117 100644
--- a/pygmt/tests/test_clib_put_strings.py
+++ b/pygmt/tests/test_clib_put_strings.py
@@ -24,7 +24,7 @@ def test_put_strings():
         )
         x = np.array([1, 2, 3, 4, 5], dtype=np.int32)
         y = np.array([6, 7, 8, 9, 10], dtype=np.int32)
-        strings = np.array(["a", "bc", "defg", "hijklmn", "opqrst"], dtype=str)
+        strings = np.array(["a", "bc", "defg", "hijklmn", "opqrst"], dtype=np.str_)
         lib.put_vector(dataset, column=lib["GMT_X"], vector=x)
         lib.put_vector(dataset, column=lib["GMT_Y"], vector=y)
         lib.put_strings(

From 7dc353b61ee1c9393d87c79e91c23ab03fe30a93 Mon Sep 17 00:00:00 2001
From: Wei Ji <23487320+weiji14@users.noreply.github.com>
Date: Fri, 15 Nov 2024 14:31:55 +1300
Subject: [PATCH 17/25] Revert support of pyarrow.array inputs to put_strings

---
 pygmt/clib/session.py | 6 ++----
 1 file changed, 2 insertions(+), 4 deletions(-)

diff --git a/pygmt/clib/session.py b/pygmt/clib/session.py
index b8d6ff6feb1..10c8770adaa 100644
--- a/pygmt/clib/session.py
+++ b/pygmt/clib/session.py
@@ -1029,8 +1029,7 @@ def put_strings(self, dataset: ctp.c_void_p, family: str, strings: np.ndarray):
         Raises
         ------
         GMTCLibError
-            If given invalid input or ``GMT_Put_Strings`` exits with
-            status != 0.
+            If given invalid input or ``GMT_Put_Strings`` exits with a non-zero status.
         """
         c_put_strings = self.get_libgmt_func(
             "GMT_Put_Strings",
@@ -1051,8 +1050,7 @@ def put_strings(self, dataset: ctp.c_void_p, family: str, strings: np.ndarray):
             self.session_pointer, family_int, dataset, strings_pointer
         )
         if status != 0:
-            dtype = strings.dtype if hasattr(strings, "dtype") else type(strings)
-            msg = f"Failed to put strings of type {dtype} into dataset."
+            msg = f"Failed to put strings of type {strings.dtype} into dataset."
             raise GMTCLibError(msg)
 
     def put_matrix(self, dataset: ctp.c_void_p, matrix: np.ndarray, pad: int = 0):

From ce761520995677a5193007a4502fdb438ed2680a Mon Sep 17 00:00:00 2001
From: Wei Ji <23487320+weiji14@users.noreply.github.com>
Date: Fri, 15 Nov 2024 14:44:00 +1300
Subject: [PATCH 18/25] Remove StringArrayTypes type hint

Reverts 0a6cda5285f68b607e0fc49436dc76c84f00e598, f59f93cf86d9f48ba4a8d7c310f73f5716391539, and 17c1e9ca28acbb9f35223b0bb770a8dc8bb75843
---
 pygmt/_typing.py | 10 ----------
 1 file changed, 10 deletions(-)

diff --git a/pygmt/_typing.py b/pygmt/_typing.py
index 4a57c3c7678..bbc7d596c65 100644
--- a/pygmt/_typing.py
+++ b/pygmt/_typing.py
@@ -2,17 +2,7 @@
 Type aliases for type hints.
 """
 
-import contextlib
-import importlib
-from collections.abc import Sequence
 from typing import Literal
 
-import numpy as np
-
 # Anchor codes
 AnchorCode = Literal["TL", "TC", "TR", "ML", "MC", "MR", "BL", "BC", "BR"]
-
-# String array types
-StringArrayTypes = Sequence[str] | np.ndarray
-with contextlib.suppress(ImportError):
-    StringArrayTypes |= importlib.import_module(name="pyarrow").StringArray

From ef431afd065f3c8af35d94c6c8c62f301bf698c7 Mon Sep 17 00:00:00 2001
From: Wei Ji <23487320+weiji14@users.noreply.github.com>
Date: Fri, 15 Nov 2024 14:52:13 +1300
Subject: [PATCH 19/25] Revert "Remove StringArrayTypes type hint"

This reverts commit ce761520995677a5193007a4502fdb438ed2680a.
---
 pygmt/_typing.py | 10 ++++++++++
 1 file changed, 10 insertions(+)

diff --git a/pygmt/_typing.py b/pygmt/_typing.py
index bbc7d596c65..4a57c3c7678 100644
--- a/pygmt/_typing.py
+++ b/pygmt/_typing.py
@@ -2,7 +2,17 @@
 Type aliases for type hints.
 """
 
+import contextlib
+import importlib
+from collections.abc import Sequence
 from typing import Literal
 
+import numpy as np
+
 # Anchor codes
 AnchorCode = Literal["TL", "TC", "TR", "ML", "MC", "MR", "BL", "BC", "BR"]
+
+# String array types
+StringArrayTypes = Sequence[str] | np.ndarray
+with contextlib.suppress(ImportError):
+    StringArrayTypes |= importlib.import_module(name="pyarrow").StringArray

From acaf3509b169a18b62d919fd01adc83dba7c7937 Mon Sep 17 00:00:00 2001
From: Wei Ji <23487320+weiji14@users.noreply.github.com>
Date: Fri, 15 Nov 2024 14:54:53 +1300
Subject: [PATCH 20/25] Improve type-hint of text parameter in
 pygmt.Figure.text

Include pyarrow.StringArray besides Sequence[str] or np.ndarray.
---
 pygmt/src/text.py | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/pygmt/src/text.py b/pygmt/src/text.py
index 2ed475c9ac2..c15c6b2e499 100644
--- a/pygmt/src/text.py
+++ b/pygmt/src/text.py
@@ -5,7 +5,7 @@
 from collections.abc import Sequence
 
 import numpy as np
-from pygmt._typing import AnchorCode
+from pygmt._typing import AnchorCode, StringArrayTypes
 from pygmt.clib import Session
 from pygmt.exceptions import GMTInvalidInput
 from pygmt.helpers import (
@@ -48,7 +48,7 @@ def text_(  # noqa: PLR0912
     x=None,
     y=None,
     position: AnchorCode | None = None,
-    text=None,
+    text: str | StringArrayTypes = None,
     angle=None,
     font=None,
     justify: bool | None | AnchorCode | Sequence[AnchorCode] = None,
@@ -104,7 +104,7 @@ def text_(  # noqa: PLR0912
 
         For example, ``position="TL"`` plots the text at the Top Left corner
         of the map.
-    text : str or 1-D array
+    text
         The text string, or an array of strings to plot on the figure.
     angle: float, str, bool or list
         Set the angle measured in degrees counter-clockwise from

From 6ad6eb96d7d0a0bcbe8e01386d80059a6d70a104 Mon Sep 17 00:00:00 2001
From: Wei Ji <23487320+weiji14@users.noreply.github.com>
Date: Fri, 15 Nov 2024 14:58:44 +1300
Subject: [PATCH 21/25] Move pa.array parametrizations to another file

The test_virtualfile_from_vectors_one_string_or_object_column unit test was moved to test_clib_virtualfile_from_vectors.py. Xref https://github.com/GenericMappingTools/pygmt/pull/3512
---
 .../test_clib_virtualfile_from_vectors.py     | 25 ++++++++++----
 pygmt/tests/test_clib_virtualfiles.py         | 34 -------------------
 2 files changed, 19 insertions(+), 40 deletions(-)

diff --git a/pygmt/tests/test_clib_virtualfile_from_vectors.py b/pygmt/tests/test_clib_virtualfile_from_vectors.py
index 041bc7a803c..453f222b6bc 100644
--- a/pygmt/tests/test_clib_virtualfile_from_vectors.py
+++ b/pygmt/tests/test_clib_virtualfile_from_vectors.py
@@ -53,17 +53,30 @@ def test_virtualfile_from_vectors(dtypes):
 
 
 @pytest.mark.benchmark
-@pytest.mark.parametrize("dtype", [str, object])
-def test_virtualfile_from_vectors_one_string_or_object_column(dtype):
-    """
-    Test passing in one column with string or object dtype into virtual file dataset.
+@pytest.mark.parametrize(
+    ("array_func", "dtype"),
+    [
+        pytest.param(np.array, {"dtype": np.str_}, id="str"),
+        pytest.param(np.array, {"dtype": np.object_}, id="object"),
+        pytest.param(
+            getattr(pa, "array", None),
+            {},  # pa.string()
+            marks=skip_if_no(package="pyarrow"),
+            id="pyarrow",
+        ),
+    ],
+)
+def test_virtualfile_from_vectors_one_string_or_object_column(array_func, dtype):
+    """
+    Test passing in one column with string (numpy/pyarrow) or object (numpy)
+    dtype into virtual file dataset.
     """
     size = 5
     x = np.arange(size, dtype=np.int32)
     y = np.arange(size, size * 2, 1, dtype=np.int32)
-    strings = np.array(["a", "bc", "defg", "hijklmn", "opqrst"], dtype=dtype)
+    strings = array_func(["a", "bc", "defg", "hijklmn", "opqrst"], **dtype)
     with clib.Session() as lib:
-        with lib.virtualfile_from_vectors((x, y, strings)) as vfile:
+        with lib.virtualfile_from_vectors(x, y, strings) as vfile:
             with GMTTempFile() as outfile:
                 lib.call_module("convert", [vfile, f"->{outfile.name}"])
                 output = outfile.read(keep_tabs=True)
diff --git a/pygmt/tests/test_clib_virtualfiles.py b/pygmt/tests/test_clib_virtualfiles.py
index 6277f80e153..c747f763024 100644
--- a/pygmt/tests/test_clib_virtualfiles.py
+++ b/pygmt/tests/test_clib_virtualfiles.py
@@ -143,37 +143,3 @@ def test_open_virtual_file():
         bounds = "\t".join([f"<{col.min():.0f}/{col.max():.0f}>" for col in data.T])
         expected = f"<matrix memory>: N = {shape[0]}\t{bounds}\n"
         assert output == expected
-
-
-@pytest.mark.benchmark
-@pytest.mark.parametrize(
-    ("array_func", "dtype"),
-    [
-        pytest.param(np.array, {"dtype": np.str_}, id="str"),
-        pytest.param(np.array, {"dtype": np.object_}, id="object"),
-        pytest.param(
-            getattr(pa, "array", None),
-            {},  # pa.string()
-            marks=skip_if_no(package="pyarrow"),
-            id="pyarrow",
-        ),
-    ],
-)
-def test_virtualfile_from_vectors_one_string_or_object_column(array_func, dtype):
-    """
-    Test passing in one column with string (numpy/pyarrow) or object (numpy)
-    dtype into virtual file dataset.
-    """
-    size = 5
-    x = np.arange(size, dtype=np.int32)
-    y = np.arange(size, size * 2, 1, dtype=np.int32)
-    strings = array_func(["a", "bc", "defg", "hijklmn", "opqrst"], **dtype)
-    with clib.Session() as lib:
-        with lib.virtualfile_from_vectors(x, y, strings) as vfile:
-            with GMTTempFile() as outfile:
-                lib.call_module("convert", [vfile, f"->{outfile.name}"])
-                output = outfile.read(keep_tabs=True)
-        expected = "".join(
-            f"{i}\t{j}\t{k}\n" for i, j, k in zip(x, y, strings, strict=True)
-        )
-        assert output == expected

From d88accd2aa8ea47a61730fb2edc44d022deaf7d4 Mon Sep 17 00:00:00 2001
From: Wei Ji <23487320+weiji14@users.noreply.github.com>
Date: Fri, 15 Nov 2024 15:00:38 +1300
Subject: [PATCH 22/25] Pass a tuple of vectors to virtualfile_from_vectors

---
 pygmt/tests/test_clib_virtualfile_from_vectors.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/pygmt/tests/test_clib_virtualfile_from_vectors.py b/pygmt/tests/test_clib_virtualfile_from_vectors.py
index 453f222b6bc..238a0f5cb9c 100644
--- a/pygmt/tests/test_clib_virtualfile_from_vectors.py
+++ b/pygmt/tests/test_clib_virtualfile_from_vectors.py
@@ -76,7 +76,7 @@ def test_virtualfile_from_vectors_one_string_or_object_column(array_func, dtype)
     y = np.arange(size, size * 2, 1, dtype=np.int32)
     strings = array_func(["a", "bc", "defg", "hijklmn", "opqrst"], **dtype)
     with clib.Session() as lib:
-        with lib.virtualfile_from_vectors(x, y, strings) as vfile:
+        with lib.virtualfile_from_vectors(vectors=(x, y, strings)) as vfile:
             with GMTTempFile() as outfile:
                 lib.call_module("convert", [vfile, f"->{outfile.name}"])
                 output = outfile.read(keep_tabs=True)

From 265132ece545279bdf2e3734f9a4e736b357150f Mon Sep 17 00:00:00 2001
From: Wei Ji <23487320+weiji14@users.noreply.github.com>
Date: Fri, 15 Nov 2024 15:14:23 +1300
Subject: [PATCH 23/25] Move skip_if_no and pyarrow import to
 test_clib_virtualfile_from_vectors

---
 pygmt/tests/test_clib_virtualfile_from_vectors.py | 6 ++++++
 pygmt/tests/test_clib_virtualfiles.py             | 6 ------
 2 files changed, 6 insertions(+), 6 deletions(-)

diff --git a/pygmt/tests/test_clib_virtualfile_from_vectors.py b/pygmt/tests/test_clib_virtualfile_from_vectors.py
index 238a0f5cb9c..55702f114b6 100644
--- a/pygmt/tests/test_clib_virtualfile_from_vectors.py
+++ b/pygmt/tests/test_clib_virtualfile_from_vectors.py
@@ -11,6 +11,12 @@
 from pygmt.clib.session import DTYPES_NUMERIC
 from pygmt.exceptions import GMTInvalidInput
 from pygmt.helpers import GMTTempFile
+from pygmt.helpers.testing import skip_if_no
+
+try:
+    import pyarrow as pa
+except ImportError:
+    pa = None
 
 
 @pytest.fixture(scope="module", name="dtypes")
diff --git a/pygmt/tests/test_clib_virtualfiles.py b/pygmt/tests/test_clib_virtualfiles.py
index c747f763024..a45a662de71 100644
--- a/pygmt/tests/test_clib_virtualfiles.py
+++ b/pygmt/tests/test_clib_virtualfiles.py
@@ -10,14 +10,8 @@
 from pygmt.clib.session import DTYPES_NUMERIC
 from pygmt.exceptions import GMTCLibError, GMTInvalidInput
 from pygmt.helpers import GMTTempFile
-from pygmt.helpers.testing import skip_if_no
 from pygmt.tests.test_clib import mock
 
-try:
-    import pyarrow as pa
-except ImportError:
-    pa = None
-
 POINTS_DATA = Path(__file__).parent / "data" / "points.txt"
 
 

From edb3438e81f0ac261973cc006dd0a40df366ffb0 Mon Sep 17 00:00:00 2001
From: Wei Ji <23487320+weiji14@users.noreply.github.com>
Date: Fri, 15 Nov 2024 15:15:31 +1300
Subject: [PATCH 24/25] The text argument can be None

---
 pygmt/src/text.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/pygmt/src/text.py b/pygmt/src/text.py
index c15c6b2e499..ad98711824b 100644
--- a/pygmt/src/text.py
+++ b/pygmt/src/text.py
@@ -48,7 +48,7 @@ def text_(  # noqa: PLR0912
     x=None,
     y=None,
     position: AnchorCode | None = None,
-    text: str | StringArrayTypes = None,
+    text: str | StringArrayTypes | None = None,
     angle=None,
     font=None,
     justify: bool | None | AnchorCode | Sequence[AnchorCode] = None,

From 8172102c4ea460e0dcc3de8a79cdc0cdf77de5fa Mon Sep 17 00:00:00 2001
From: Wei Ji <23487320+weiji14@users.noreply.github.com>
Date: Fri, 15 Nov 2024 15:32:05 +1300
Subject: [PATCH 25/25] Simplify to remove getattr(pa, "array", None) call

Co-Authored-By: Dongdong Tian <seisman.info@gmail.com>
---
 pygmt/tests/test_clib_virtualfile_from_vectors.py |  8 +++++---
 pygmt/tests/test_text.py                          | 10 ++++------
 2 files changed, 9 insertions(+), 9 deletions(-)

diff --git a/pygmt/tests/test_clib_virtualfile_from_vectors.py b/pygmt/tests/test_clib_virtualfile_from_vectors.py
index 55702f114b6..b76a9bfe168 100644
--- a/pygmt/tests/test_clib_virtualfile_from_vectors.py
+++ b/pygmt/tests/test_clib_virtualfile_from_vectors.py
@@ -15,8 +15,10 @@
 
 try:
     import pyarrow as pa
+
+    pa_array = pa.array
 except ImportError:
-    pa = None
+    pa_array = None
 
 
 @pytest.fixture(scope="module", name="dtypes")
@@ -65,8 +67,8 @@ def test_virtualfile_from_vectors(dtypes):
         pytest.param(np.array, {"dtype": np.str_}, id="str"),
         pytest.param(np.array, {"dtype": np.object_}, id="object"),
         pytest.param(
-            getattr(pa, "array", None),
-            {},  # pa.string()
+            pa_array,
+            {},  # {"type": pa.string()}
             marks=skip_if_no(package="pyarrow"),
             id="pyarrow",
         ),
diff --git a/pygmt/tests/test_text.py b/pygmt/tests/test_text.py
index ce59d7d4bd6..593c07a7b4d 100644
--- a/pygmt/tests/test_text.py
+++ b/pygmt/tests/test_text.py
@@ -13,8 +13,10 @@
 
 try:
     import pyarrow as pa
+
+    pa_array = pa.array
 except ImportError:
-    pa = None
+    pa_array = None
 
 TEST_DATA_DIR = Path(__file__).parent / "data"
 POINTS_DATA = TEST_DATA_DIR / "points.txt"
@@ -60,11 +62,7 @@ def test_text_single_line_of_text(region, projection):
     [
         list,
         pytest.param(np.array, id="numpy"),
-        pytest.param(
-            getattr(pa, "array", None),
-            marks=skip_if_no(package="pyarrow"),
-            id="pyarrow",
-        ),
+        pytest.param(pa_array, marks=skip_if_no(package="pyarrow"), id="pyarrow"),
     ],
 )
 def test_text_multiple_lines_of_text(region, projection, array_func):