Merge branch 'benbusby:main' into main

app/filter.py

@@ -349,7 +349,12 @@ class Filter:
         if any(url in link['href'] for url in unsupported_g_pages):
             # FIXME: The "Shopping" tab requires further filtering (see #136)
             # Temporarily removing all links to that tab for now.
+            parent = link.parent
+            while parent:
+                p_cls = parent.attrs.get('class') or []
+                if parent.name == 'footer' or f'{GClasses.footer}' in p_cls:
                     link.decompose()
+                parent = parent.parent
             return
 
         # Replace href with only the intended destination (no "utm" type tags)

app/models/g_classes.py

@@ -11,6 +11,7 @@ class GClasses(Enum):
     """
     main_tbm_tab = 'KP7LCb'
     images_tbm_tab = 'n692Zd'
+    footer = 'TuS8Ad'
 
     def __str__(self):
         return self.value

app/request.py

@@ -1,6 +1,6 @@
 from app.models.config import Config
 from datetime import datetime
-import xml.etree.ElementTree as ET
+from defusedxml import ElementTree as ET
 import random
 import requests
 from requests import Response, ConnectionError

app/routes.py

@@ -304,8 +304,8 @@ def search():
     search_util = Search(request, g.user_config, g.session_key)
     query = search_util.new_search_query()
 
-    bang = resolve_bang(query=query, bangs_dict=bang_json)
+    bang = resolve_bang(query, bang_json, url_for('.index'))
-    if bang != '':
+    if bang:
         return redirect(bang)
 
     # Redirect to home if invalid/blank search

app/utils/bangs.py

@@ -38,7 +38,7 @@ def gen_bangs_json(bangs_file: str) -> None:
     print('* Finished creating ddg bangs json')
 
 
-def resolve_bang(query: str, bangs_dict: dict) -> str:
+def resolve_bang(query: str, bangs_dict: dict, fallback: str) -> str:
     """Transform's a user's query to a bang search, if an operator is found
 
     Args:
@@ -59,8 +59,14 @@ def resolve_bang(query: str, bangs_dict: dict) -> str:
         if operator not in split_query \
                 and operator[1:] + operator[0] not in split_query:
             continue
-        return bangs_dict[operator]['url'].replace(
+
-            '{}',
+        bang_query = query.replace(
-            query.replace(operator if operator in split_query
+            operator if operator in split_query else operator[1:] +
-                          else operator[1:] + operator[0], '').strip(), 1)
+            operator[0], ''
+        ).strip()
+
+        if bang_query:
+            return bangs_dict[operator]['url'].replace('{}', bang_query, 1)
+        else:
+            return fallback
     return ''

requirements.txt

@@ -6,6 +6,7 @@ cffi==1.15.0
 chardet==3.0.4
 click==8.0.3
 cryptography==3.3.2
+defusedxml==0.7.1
 Flask==1.1.1
 Flask-Session==0.4.0
 idna==2.9

setup.cfg

@@ -19,6 +19,7 @@ include_package_data = True
 install_requires=
     beautifulsoup4
     cryptography
+    defusedxml
     Flask
     Flask-Session
     python-dotenv

test/test_routes.py

@@ -47,6 +47,11 @@ def test_ddg_bang(client):
     assert rv._status_code == 302
     assert rv.headers.get('Location').startswith('https://github.com')
 
+    # Ensure bang without content doesn't redirect to the result
+    rv = client.get(f'/{Endpoint.search}?q=!gh')
+    assert rv._status_code == 302
+    assert not rv.headers.get('Location').startswith('https://github.com')
+
 
 def test_config(client):
     rv = client.post(f'/{Endpoint.config}', data=demo_config)