Merge pull request #328 from renatodvc/fix/markup-clean-steps-default

Fix get_citations(markup_text=...) crashing without an "html" clean step

Author: grossir

CHANGES.md

@@ -11,7 +11,9 @@ Changes:
 -
 
 Fixes:
--
+- Fix 2 `TypeError`s raised in `get_citations(markup_text=...)` when
+  `clean_steps` was omitted or lacked `"html"`. The documented fallback that
+  prepends the `html` step was both unreachable and broken.
 
 ## Current
 

eyecite/models.py

@@ -930,11 +930,13 @@ def __post_init__(self):
         elif self.markup_text and not self.plain_text:
             self.source_text = self.markup_text
 
-            if "html" not in self.clean_steps:
-                self.clean_steps.insert("html", 0)
+            clean_steps = list(self.clean_steps or [])
+            if "html" not in clean_steps:
+                clean_steps.insert(0, "html")
                 logger.warning(
                     "`html` has been added to `markup_text` clean_steps list"
                 )
+            self.clean_steps = clean_steps
 
             self.plain_text = clean_text(self.markup_text, self.clean_steps)
 

tests/test_FindTest.py

@@ -1406,6 +1406,33 @@ def test_reference_filtering(self):
                 any(isinstance(cite, ReferenceCitation) for cite in citations)
             )
 
+    def test_markup_text_without_html_clean_step(self) -> None:
+        """Does passing `markup_text` without an `html` clean step auto-add
+        it (with a warning) instead of crashing?
+
+        """
+        markup = "<p>Lissner v. Test, <i>1 U.S. 1</i> (1982)</p>"
+
+        # clean_steps omitted entirely (get_citations defaults it to None)
+        with self.assertLogs("eyecite.models", level="WARNING") as logs:
+            cites = get_citations(markup_text=markup)
+        self.assertEqual([c.matched_text() for c in cites], ["1 U.S. 1"])
+        self.assertTrue(
+            any("`html` has been added" in line for line in logs.output)
+        )
+
+        # clean_steps provided but missing "html", must prepend it
+        with self.assertLogs("eyecite.models", level="WARNING"):
+            cites = get_citations(
+                markup_text=markup, clean_steps=["all_whitespace"]
+            )
+        self.assertEqual([c.matched_text() for c in cites], ["1 U.S. 1"])
+
+        # clean_steps already containing "html" stays untouched and quiet
+        with self.assertNoLogs("eyecite.models", level="WARNING"):
+            cites = get_citations(markup_text=markup, clean_steps=["html"])
+        self.assertEqual([c.matched_text() for c in cites], ["1 U.S. 1"])
+
     def test_markup_plaintiff_and_antecedent_guesses(self) -> None:
         # Can we identify full case names in markup text
         test_pairs = (