Fix regex to find throttling function name (#1282)

* Generalize regex to find throttle function name
* Extend tests and fixtures for throttling name
* Rename mock asset to follow similar naming style

pytube/cipher.py

@@ -270,7 +270,7 @@ def get_throttling_function_name(js: str) -> str:
         # Bpa.length || iha("")) }};
         # In the above case, `iha` is the relevant function name
         r'a\.[a-zA-Z]\s*&&\s*\([a-z]\s*=\s*a\.get\("n"\)\)\s*&&\s*'
-        r'\([a-z]\s*=\s*([a-zA-Z0-9$]{3})(\[\d+\])?\([a-z]\)',
+        r'\([a-z]\s*=\s*([a-zA-Z0-9$]+)(\[\d+\])?\([a-z]\)',
     ]
     logger.debug('Finding throttling function name')
     for pattern in function_patterns:
@@ -285,7 +285,7 @@ def get_throttling_function_name(js: str) -> str:
                 idx = idx.strip("[]")
                 array = re.search(
                     r'var {nfunc}\s*=\s*(\[.+?\]);'.format(
-                        nfunc=function_match.group(1)),
+                        nfunc=re.escape(function_match.group(1))),
                     js
                 )
                 if array:

tests/conftest.py

@@ -150,13 +150,17 @@ def channel_videos_html():
 
 @pytest.fixture
 def base_js():
-    """Youtube base.js retrieved on 2022-02-04 from
-    https://www.youtube.com/watch?v=vmzxpUsN0uA
+    """Youtube base.js files retrieved on 2022-02-04 and 2022-04-15
+    from https://www.youtube.com/watch?v=vmzxpUsN0uA and
+    https://www.youtube.com/watch?v=Y4-GSFKZmEg respectively
     """
-    file_path = os.path.join(
-        os.path.dirname(os.path.realpath(__file__)),
-        "mocks",
-        "base.js.gz",
-    )
-    with gzip.open(file_path, 'rb') as f:
-        return f.read().decode('utf-8')
+    base_js_files = []
+    for file in ["base.js-2022-02-04.gz", "base.js-2022-04-15.gz"]:
+        file_path = os.path.join(
+            os.path.dirname(os.path.realpath(__file__)),
+            "mocks",
+            file,
+        )
+        with gzip.open(file_path, 'rb') as f:
+            base_js_files.append(f.read().decode('utf-8'))
+    return base_js_files

tests/test_cipher.py

@@ -80,11 +80,24 @@ def test_js_splice():
 
 
 def test_get_throttling_function_name(base_js):
-    # Values expected as of 2022/02/04:
-    raw_var = r'var Apa=[hha]'
-    assert raw_var in base_js
-    raw_code = r'a.url="";a.C&&(b=a.get("n"))&&(b=Apa[0](b),a.set("n",b),'\
-               r'Apa.length||hha(""))}};'
-    assert raw_code in base_js
-    func_name = cipher.get_throttling_function_name(base_js)
-    assert func_name == "hha"
+    base_js_code_fragments = [
+        # Values expected as of 2022/02/04:
+        {
+            'raw_var' : r'var Apa=[hha]',
+            'raw_code': r'a.url="";a.C&&(b=a.get("n"))&&(b=Apa[0](b),a.set("n",b),'\
+                        r'Apa.length||hha(""))}};',
+            'nfunc_name': 'hha'
+        },
+        # Values expected as of 2022/04/15:
+        {
+            'raw_var' : r'var $x=[uq]',
+            'raw_code': r'a.url="";a.D&&(b=a.get("n"))&&(b=$x[0](b),a.set("n",b),'\
+                        r'$x.length||uq(""))',
+            'nfunc_name': 'uq'
+        }
+    ]
+    for code_fragment, base_js_file in zip(base_js_code_fragments, base_js):
+        assert code_fragment['raw_var'] in base_js_file
+        assert code_fragment['raw_code'] in base_js_file
+        func_name = cipher.get_throttling_function_name(base_js_file)
+        assert func_name == code_fragment['nfunc_name']