Skip unmatched main capture groups in page searches

CHANGELOG.md

@@ -5,6 +5,7 @@ All notable changes to this project will be documented in this file. The format
 ## [Unreleased]
 
 ### Fixed
+- Skip unmatched optional capture groups in `Page.search()` when `main_group` selects them.
 - Initialize PDFium's form environment in `get_page_image` so that filled AcroForm field content is included when rendering pages via `Page.to_image()`. ([#1367](https://github.com/jsvine/pdfplumber/issues/1367))
 - Fix `make venv`, which created the virtual environment at `venv/` but then installed into `${VENV}` (default `.venv/`), causing the target to fail on a fresh checkout (h/t @soodoku). ([ec96f72](https://github.com/jsvine/pdfplumber/commit/ec96f72))
 - Include the wrapped exception's class name in `PdfminerException`'s message when `pdfminer.six` raises an exception without one (e.g. `PDFPasswordIncorrect`), which previously surfaced as a blank error message.

README.md

@@ -577,6 +577,7 @@ Many thanks to the following users who've contributed ideas, features, and fixes
 - [Sebastian Cao](https://github.com/cycsmail)
 - [Kaspar Naraghi](https://github.com/kaninaba94)
 - [Siddharth Gaur](https://github.com/siddharthgaur1)
+- [Jake Wang](https://github.com/jakezwang)
 
 ## Contributing
 

pdfplumber/utils/text.py

@@ -198,7 +198,7 @@ def search(
         gen = re.finditer(compiled, self.as_string)
         # Remove zero-length matches (can happen, e.g., with optional
         # patterns in regexes) and whitespace-only matches
-        filtered = filter(lambda m: bool(m.group(main_group).strip()), gen)
+        filtered = filter(lambda m: bool((m.group(main_group) or "").strip()), gen)
         return [
             self.match_to_dict(
                 m,

tests/test_utils.py

@@ -470,6 +470,22 @@ def test_search_string(self):
         results = page.search(r"10 Tuesday", layout=True)
         assert len(results) == 0
 
+    def test_search_optional_main_group(self):
+        page = self.pdf_scotus.pages[0]
+        assert page.search(r"(sdfsd)?", main_group=1) == []
+
+    def test_search_alternative_main_group(self):
+        page = self.pdf_scotus.pages[0]
+        results = page.search(
+            r"Tuesday|(supreme\s+court)",
+            main_group=1,
+            case=False,
+            return_groups=False,
+        )
+        expected = page.search(r"supreme\s+court", case=False, return_groups=False)
+        assert len(expected) == 2
+        assert results == expected
+
     def test_extract_text_lines(self):
         page = self.pdf_scotus.pages[0]
         results = page.extract_text_lines()