ENH: multirow naive implementation Styler.to_latex part1 (#43369)

* multirow naive implementation

* multirow naive implementation

* fix tests

* fixture

* clean up tests with fixture

* (ivanomg req) ref tests

* ivoanovmg req

Co-authored-by: JHM Darbyshire (iMac) <attack68@users.noreply.github.com>

doc/source/whatsnew/v1.4.0.rst

@@ -78,6 +78,7 @@ Styler
   - :meth:`.Styler.to_html` introduces keyword arguments ``sparse_index``, ``sparse_columns``, ``bold_headers``, ``caption`` (:issue:`41946`, :issue:`43149`).
   - Keyword arguments ``level`` and ``names`` added to :meth:`.Styler.hide_index` and :meth:`.Styler.hide_columns` for additional control of visibility of MultiIndexes and index names (:issue:`25475`, :issue:`43404`, :issue:`43346`)
   - Global options have been extended to configure default ``Styler`` properties including formatting and encoding and mathjax options and LaTeX (:issue:`41395`)
+  - Naive sparsification is now possible for LaTeX without the multirow package (:issue:`43369`)
 
 Formerly Styler relied on ``display.html.use_mathjax``, which has now been replaced by ``styler.html.mathjax``.
 

pandas/core/config_init.py

@@ -889,7 +889,7 @@ def register_converter_cb(key):
         "latex.multirow_align",
         "c",
         styler_multirow_align,
-        validator=is_one_of_factory(["c", "t", "b"]),
+        validator=is_one_of_factory(["c", "t", "b", "naive"]),
     )
 
     cf.register_option(

pandas/io/formats/style.py

@@ -520,10 +520,13 @@ def to_latex(
             Whether to sparsify the display of a hierarchical index. Setting to False
             will display each explicit level element in a hierarchical key for each
             column. Defaults to ``pandas.options.styler.sparse.columns`` value.
-        multirow_align : {"c", "t", "b"}, optional
+        multirow_align : {"c", "t", "b", "naive"}, optional
             If sparsifying hierarchical MultiIndexes whether to align text centrally,
-            at the top or bottom. If not given defaults to
-            ``pandas.options.styler.latex.multirow_align``
+            at the top or bottom using the multirow package. If not given defaults to
+            ``pandas.options.styler.latex.multirow_align``. If "naive" is given renders
+            without multirow.
+
+            .. versionchanged:: 1.4.0
         multicol_align : {"r", "c", "l"}, optional
             If sparsifying hierarchical MultiIndex columns whether to align text at
             the left, centrally, or at the right. If not given defaults to

pandas/io/formats/style_render.py

@@ -1412,6 +1412,8 @@ def _parse_latex_header_span(
             colspan = int(colspan[: colspan.find('"')])
             return f"\\multicolumn{{{colspan}}}{{{multicol_align}}}{{{display_val}}}"
         elif 'rowspan="' in attrs:
+            if multirow_align == "naive":
+                return display_val
             rowspan = attrs[attrs.find('rowspan="') + 9 :]
             rowspan = int(rowspan[: rowspan.find('"')])
             return f"\\multirow[{multirow_align}]{{{rowspan}}}{{*}}{{{display_val}}}"

pandas/tests/io/formats/style/test_to_latex.py

@@ -24,6 +24,13 @@ def df():
     return DataFrame({"A": [0, 1], "B": [-0.61, -1.22], "C": ["ab", "cd"]})
 
 
+@pytest.fixture
+def df_ext():
+    return DataFrame(
+        {"A": [0, 1, 2], "B": [-0.61, -1.22, -2.22], "C": ["ab", "cd", "de"]}
+    )
+
+
 @pytest.fixture
 def styler(df):
     return Styler(df, uuid_len=0, precision=2)
@@ -210,11 +217,9 @@ def test_multiindex_columns(df):
     assert expected == s.to_latex(sparse_columns=False)
 
 
-def test_multiindex_row(df):
+def test_multiindex_row(df_ext):
     ridx = MultiIndex.from_tuples([("A", "a"), ("A", "b"), ("B", "c")])
-    df.loc[2, :] = [2, -2.22, "de"]
-    df = df.astype({"A": int})
-    df.index = ridx
+    df_ext.index = ridx
     expected = dedent(
         """\
         \\begin{tabular}{llrrl}
@@ -225,8 +230,9 @@ def test_multiindex_row(df):
         \\end{tabular}
         """
     )
-    s = df.style.format(precision=2)
-    assert expected == s.to_latex()
+    styler = df_ext.style.format(precision=2)
+    result = styler.to_latex()
+    assert expected == result
 
     # non-sparse
     expected = dedent(
@@ -239,15 +245,32 @@ def test_multiindex_row(df):
         \\end{tabular}
         """
     )
-    assert expected == s.to_latex(sparse_index=False)
+    result = styler.to_latex(sparse_index=False)
+    assert expected == result
+
+
+def test_multirow_naive(df_ext):
+    ridx = MultiIndex.from_tuples([("X", "x"), ("X", "y"), ("Y", "z")])
+    df_ext.index = ridx
+    expected = dedent(
+        """\
+        \\begin{tabular}{llrrl}
+         &  & A & B & C \\\\
+        X & x & 0 & -0.61 & ab \\\\
+         & y & 1 & -1.22 & cd \\\\
+        Y & z & 2 & -2.22 & de \\\\
+        \\end{tabular}
+        """
+    )
+    styler = df_ext.style.format(precision=2)
+    result = styler.to_latex(multirow_align="naive")
+    assert expected == result
 
 
-def test_multiindex_row_and_col(df):
+def test_multiindex_row_and_col(df_ext):
     cidx = MultiIndex.from_tuples([("Z", "a"), ("Z", "b"), ("Y", "c")])
     ridx = MultiIndex.from_tuples([("A", "a"), ("A", "b"), ("B", "c")])
-    df.loc[2, :] = [2, -2.22, "de"]
-    df = df.astype({"A": int})
-    df.index, df.columns = ridx, cidx
+    df_ext.index, df_ext.columns = ridx, cidx
     expected = dedent(
         """\
         \\begin{tabular}{llrrl}
@@ -259,8 +282,9 @@ def test_multiindex_row_and_col(df):
         \\end{tabular}
         """
     )
-    s = df.style.format(precision=2)
-    assert s.to_latex(multirow_align="b", multicol_align="l") == expected
+    styler = df_ext.style.format(precision=2)
+    result = styler.to_latex(multirow_align="b", multicol_align="l")
+    assert result == expected
 
     # non-sparse
     expected = dedent(
@@ -274,16 +298,15 @@ def test_multiindex_row_and_col(df):
         \\end{tabular}
         """
     )
-    assert s.to_latex(sparse_index=False, sparse_columns=False) == expected
+    result = styler.to_latex(sparse_index=False, sparse_columns=False)
+    assert result == expected
 
 
-def test_multi_options(df):
+def test_multi_options(df_ext):
     cidx = MultiIndex.from_tuples([("Z", "a"), ("Z", "b"), ("Y", "c")])
     ridx = MultiIndex.from_tuples([("A", "a"), ("A", "b"), ("B", "c")])
-    df.loc[2, :] = [2, -2.22, "de"]
-    df = df.astype({"A": int})
-    df.index, df.columns = ridx, cidx
-    styler = df.style.format(precision=2)
+    df_ext.index, df_ext.columns = ridx, cidx
+    styler = df_ext.style.format(precision=2)
 
     expected = dedent(
         """\
@@ -292,7 +315,8 @@ def test_multi_options(df):
     \\multirow[c]{2}{*}{A} & a & 0 & -0.61 & ab \\\\
     """
     )
-    assert expected in styler.to_latex()
+    result = styler.to_latex()
+    assert expected in result
 
     with option_context("styler.latex.multicol_align", "l"):
         assert " &  & \\multicolumn{2}{l}{Z} & Y \\\\" in styler.to_latex()
@@ -311,30 +335,25 @@ def test_multiindex_columns_hidden():
     assert "{tabular}{lrrr}" in s.to_latex()
 
 
-def test_sparse_options(df):
+@pytest.mark.parametrize(
+    "option, value",
+    [
+        ("styler.sparse.index", True),
+        ("styler.sparse.index", False),
+        ("styler.sparse.columns", True),
+        ("styler.sparse.columns", False),
+    ],
+)
+def test_sparse_options(df_ext, option, value):
     cidx = MultiIndex.from_tuples([("Z", "a"), ("Z", "b"), ("Y", "c")])
     ridx = MultiIndex.from_tuples([("A", "a"), ("A", "b"), ("B", "c")])
-    df.loc[2, :] = [2, -2.22, "de"]
-    df.index, df.columns = ridx, cidx
-    s = df.style
-
-    latex1 = s.to_latex()
-
-    with option_context("styler.sparse.index", True):
-        latex2 = s.to_latex()
-    assert latex1 == latex2
-
-    with option_context("styler.sparse.index", False):
-        latex2 = s.to_latex()
-    assert latex1 != latex2
+    df_ext.index, df_ext.columns = ridx, cidx
+    styler = df_ext.style
 
-    with option_context("styler.sparse.columns", True):
-        latex2 = s.to_latex()
-    assert latex1 == latex2
-
-    with option_context("styler.sparse.columns", False):
-        latex2 = s.to_latex()
-    assert latex1 != latex2
+    latex1 = styler.to_latex()
+    with option_context(option, value):
+        latex2 = styler.to_latex()
+    assert (latex1 == latex2) is value
 
 
 def test_hidden_index(styler):
@@ -352,16 +371,14 @@ def test_hidden_index(styler):
 
 
 @pytest.mark.parametrize("environment", ["table", "figure*", None])
-def test_comprehensive(df, environment):
+def test_comprehensive(df_ext, environment):
     # test as many low level features simultaneously as possible
     cidx = MultiIndex.from_tuples([("Z", "a"), ("Z", "b"), ("Y", "c")])
     ridx = MultiIndex.from_tuples([("A", "a"), ("A", "b"), ("B", "c")])
-    df.loc[2, :] = [2, -2.22, "de"]
-    df = df.astype({"A": int})
-    df.index, df.columns = ridx, cidx
-    s = df.style
-    s.set_caption("mycap")
-    s.set_table_styles(
+    df_ext.index, df_ext.columns = ridx, cidx
+    stlr = df_ext.style
+    stlr.set_caption("mycap")
+    stlr.set_table_styles(
         [
             {"selector": "label", "props": ":{fig§item}"},
             {"selector": "position", "props": ":h!"},
@@ -373,8 +390,8 @@ def test_comprehensive(df, environment):
             {"selector": "rowcolors", "props": ":{3}{pink}{}"},  # custom command
         ]
     )
-    s.highlight_max(axis=0, props="textbf:--rwrap;cellcolor:[rgb]{1,1,0.6}--rwrap")
-    s.highlight_max(axis=None, props="Huge:--wrap;", subset=[("Z", "a"), ("Z", "b")])
+    stlr.highlight_max(axis=0, props="textbf:--rwrap;cellcolor:[rgb]{1,1,0.6}--rwrap")
+    stlr.highlight_max(axis=None, props="Huge:--wrap;", subset=[("Z", "a"), ("Z", "b")])
 
     expected = (
         """\
@@ -398,7 +415,8 @@ def test_comprehensive(df, environment):
 \\end{table}
 """
     ).replace("table", environment if environment else "table")
-    assert s.format(precision=2).to_latex(environment=environment) == expected
+    result = stlr.format(precision=2).to_latex(environment=environment)
+    assert result == expected
 
 
 def test_environment_option(styler):
@@ -687,13 +705,11 @@ def test_longtable_caption_label(styler, caption, cap_exp, label, lab_exp):
         (False, False),
     ],
 )
-def test_apply_map_header_render_mi(df, index, columns, siunitx):
+def test_apply_map_header_render_mi(df_ext, index, columns, siunitx):
     cidx = MultiIndex.from_tuples([("Z", "a"), ("Z", "b"), ("Y", "c")])
     ridx = MultiIndex.from_tuples([("A", "a"), ("A", "b"), ("B", "c")])
-    df.loc[2, :] = [2, -2.22, "de"]
-    df = df.astype({"A": int})
-    df.index, df.columns = ridx, cidx
-    styler = df.style
+    df_ext.index, df_ext.columns = ridx, cidx
+    styler = df_ext.style
 
     func = lambda v: "bfseries: --rwrap" if "A" in v or "Z" in v or "c" in v else None