Improve visualization of constituency tree

hankcs · hankcs · commit dd18590d4e18 · 2022-04-01T18:40:39.000-04:00
diff --git a/plugins/hanlp_common/hanlp_common/document.py b/plugins/hanlp_common/hanlp_common/document.py
@@ -316,10 +316,8 @@ def condense(block_, extras_=None):
                     while '  ─' in row[1]:
                         row[1] = row[1].replace('  ─', ' ──')
                     row[1] = row[1].replace('─ ─', '───')
-                    row[1] = row[1].replace('─  │', '───┤')
-                    row[1] = row[1].replace('─  ├', '───┼')
-                    row[1] = re.sub(r'►([\w-]+)(\s+)([│├])', lambda
-                        m: f'►{m.group(1)}{"─" * len(m.group(2))}{"┤" if m.group(3) == "│" else "┼"}', row[1])
+                    row[1] = re.sub(r'([►─])([\w-]*)(\s+)([│├])', lambda
+                        m: f'{m.group(1)}{m.group(2)}{"─" * len(m.group(3))}{"┤" if m.group(4) == "│" else "┼"}', row[1])
                     row[1] = re.sub(r'►(─+)►', r'─\1►', row[1])
                 for r, s in zip(extras, text):
                     r.extend(s)
diff --git a/plugins/hanlp_common/setup.py b/plugins/hanlp_common/setup.py
@@ -10,7 +10,7 @@
 
 setup(
     name='hanlp_common',
-    version='0.0.14',
+    version='0.0.15',
     description='HanLP: Han Language Processing',
     long_description=long_description,
     long_description_content_type="text/markdown",
