Update coersion to pandas for both biocframe and factor. Add method t…

…o initialize a factor form a list

src/biocframe/BiocFrame.py

@@ -12,6 +12,7 @@
 from biocutils import is_list_of_type, normalize_subscript
 
 from ._validators import validate_cols, validate_rows, validate_unique_list
+from .Factor import Factor
 from .types import SlicerArgTypes, SlicerTypes
 from .utils import _slice_or_index
 
@@ -784,8 +785,14 @@ def to_pandas(self):
         """
         from pandas import DataFrame
 
+        _data_copy = OrderedDict()
+        for col in self.column_names:
+            _data_copy[col] = self.column(col)
+            if isinstance(self.column(col), Factor):
+                _data_copy[col] = _data_copy[col].to_pandas()
+
         return DataFrame(
-            data=self._data, index=self._row_names, columns=self._column_names
+            data=_data_copy, index=self._row_names, columns=self._column_names
         )
 
     # TODO: very primitive implementation, needs very robust testing

src/biocframe/Factor.py

@@ -1,7 +1,8 @@
-from typing import List, Sequence, Union
 from copy import deepcopy
-from biocgenerics.combine import combine
+from typing import List, Sequence, Union
+
 import biocutils as ut
+from biocgenerics.combine import combine
 
 
 class Factor:
@@ -312,6 +313,36 @@ def __deepcopy__(self, memo) -> "Factor":
             validate=False,
         )
 
+    def to_pandas(self):
+        """Coerce to :py:class:`~pandas.Categorical` object.
+
+        Returns:
+            Categorical: A :py:class:`~pandas.Categorical` object.
+        """
+        from pandas import Categorical
+
+        return Categorical(
+            values=[self._levels[c] for c in self._codes],
+            ordered=self._ordered,
+        )
+
+    @staticmethod
+    def from_list(values: Sequence[str]) -> "Factor":
+        """Represent a categorical vector as a Factor.
+
+        Args:
+            values (Sequence[str]): List of strings
+
+        Raises:
+            ValueError: If values is not a list.
+
+        Returns:
+            Factor: A Factor object.
+        """
+        levels, indices = ut.factor(values)
+
+        return Factor(indices, levels=levels)
+
 
 @combine.register(Factor)
 def _combine_factors(*x: Factor):

tests/test_Factor.py

@@ -2,6 +2,7 @@
 from biocgenerics.combine import combine
 import pytest
 import copy
+import pandas as pd
 
 
 def test_Factor_basics():
@@ -176,3 +177,24 @@ def test_Factor_combine():
     f2 = Factor([1, 3, 2], levels=["D", "E", "F", "G"], ordered=True)
     out = combine(f1, f2)
     assert not out.get_ordered()
+
+
+def test_Factor_pandas():
+    f1 = Factor([0, 2, 4, 2, 0], levels=["A", "B", "C", "D", "E"])
+    pcat = f1.to_pandas()
+    assert pcat is not None
+    assert len(pcat) == len(f1)
+
+    f2 = Factor([1, 3, 2], levels=["D", "E", "F", "G"], ordered=True)
+    pcat = f2.to_pandas()
+    assert pcat is not None
+    assert len(pcat) == len(f2)
+    assert pcat.ordered == f2.get_ordered()
+
+
+def test_Factor_init_from_list():
+    f1 = Factor.from_list(["A", "B", "A", "B", "E"])
+
+    assert isinstance(f1, Factor)
+    assert len(f1) == 5
+    assert len(f1.get_levels()) == 3

tests/test_methods.py

@@ -1,7 +1,8 @@
 import numpy as np
 import pytest
-
+import pandas as pd
 from biocframe.BiocFrame import BiocFrame
+from biocframe.Factor import Factor
 
 __author__ = "jkanche"
 __copyright__ = "jkanche"
@@ -430,3 +431,33 @@ def test_nested_biocFrame_preserve_types():
     assert isinstance(sliced.column("nested"), BiocFrame)
     assert isinstance(sliced.row(0), dict)
     assert isinstance(sliced.column("column2"), np.ndarray)
+
+
+def test_export_pandas():
+    obj = BiocFrame(
+        {
+            "column1": [1, 2, 3],
+            "nested": BiocFrame(
+                {
+                    "ncol1": [4, 5, 6],
+                    "ncol2": ["a", "b", "c"],
+                    "deep": ["j", "k", "l"],
+                }
+            ),
+            "column2": np.array([1, 2, 3]),
+        }
+    )
+
+    pdf = obj.to_pandas()
+    assert pdf is not None
+    assert isinstance(pdf, pd.DataFrame)
+    assert len(pdf) == len(obj)
+    assert len(set(pdf.columns).difference(obj.colnames)) == 0
+
+    obj["factor"] = Factor([0, 2, 1], levels=["A", "B", "C"])
+    pdf = obj.to_pandas()
+    assert pdf is not None
+    assert isinstance(pdf, pd.DataFrame)
+    assert len(pdf) == len(obj)
+    assert len(set(pdf.columns).difference(obj.colnames)) == 0
+    assert pdf["factor"] is not None