FIX - Proper escaping of special characters for data ops tooltips (#1764)

Co-authored-by: Jérôme Dockès <jerome@dockes.org>

Author: rcap107

CHANGES.rst

@@ -74,6 +74,9 @@ Bugfixes
 - Fixed an issue where :class:`TableReport` would fail when computing associations
   for Polars dataframes if PyArrow was not installed.
   :pr:`1742` by :user:`Riccardo Cappuzzo <rcap107>`.
+- Fixed an issue in the Data Ops report generation in cases where the DataOp
+  contained escape characters or were spanning multiple lines.
+  :pr:`1764` by :user:`Riccardo Cappuzzo <rcap107>`.
 - Added :meth:`get_feature_names_out` to :class:`Cleaner` for consistency with the
   :class:`TableVectorizer` and other transformers. :pr:`1762` by
   :user:`Riccardo Cappuzzo <rcap107>`.
@@ -84,7 +87,6 @@ Bugfixes
   ``.skb.mark_as_y()`` used to raise an error, this has been fixed in :pr:`1782`
   by :user:`Jérôme Dockès <jeromedockes>`.
 
-
 Release 0.6.2
 =============
 

pyproject.toml

@@ -261,6 +261,7 @@ exclude = [
   "doc/_build",
   "doc/auto_examples",
   "build",
+  "pixi.lock"
 ]
 
 [tool.ruff.lint]

skrub/_data_ops/_inspection.py

@@ -309,6 +309,9 @@ def _node_kwargs(data_op, url=None):
     tooltip = html.escape(data_op._skrub_impl.creation_stack_last_line())
     if description := data_op._skrub_impl.description:
         tooltip = f"{tooltip}\n\n{html.escape(description)}"
+    # Also escape backslahses inserted in the .dot file
+    # otherwise they are interpreted as escape sequences by graphviz
+    tooltip = tooltip.replace("\\", "\\\\")
     kwargs["tooltip"] = tooltip
     if isinstance(data_op._skrub_impl, (Var, Value)):
         kwargs["peripheries"] = 2

skrub/tests/test_data_ops_stack_description.py

@@ -5,6 +5,7 @@
 from sklearn.preprocessing import FunctionTransformer
 
 import skrub
+from skrub._data_ops import _inspection
 
 # Those tests have to be outside of the `_data_ops` module (and thus out of
 # _data_ops/tests/, as we place the tests inside the package), because to
@@ -78,3 +79,17 @@ def test_apply_eval_failure(eval_data_op):
         r'^.*d = c\.skb\.apply\(FunctionTransformer\(lambda x: x \+ "string"\)\)',
     ):
         eval_data_op(e, {"a": 1.0, "b": 2.0})
+
+
+@pytest.mark.skipif(not _inspection._has_graphviz(), reason="report requires graphviz")
+def test_escaping_characters():
+    # fmt: off
+    out = skrub.as_data_op("") + "\n" + '"' + "\t" +'\\ ' \
+        '\t'
+    # fmt: on
+    g = out.skb.draw_graph()
+    result = g.svg.decode("utf-8")
+    assert (
+        """out = skrub.as_data_op("") + "\\n" """
+        """+ '"' + "\\t" +'\\ ' \\""" in result
+    )