Fix: seemh server detection (#362)

* Fix: seemh server detection

* Fix: support tw

comiccrawler/mods/seemh.py

@@ -13,6 +13,7 @@
 from lzstring import LZString
 
 from ..core import Episode, grabhtml
+from ..util import balance
 
 domain = ["seemh.com", "ikanman.com", "manhuagui.com", "www.mhgui.com"]
 name = "看漫畫"
@@ -107,8 +108,9 @@ def get_images(html, url):
 	corejs = grabhtml(urljoin(url, corejs_url), referer=url)
 
 	# cache server list
-	servs = re.search(r"var servs=(.+)", configjs).group(1)
-	servs = eval(servs)
+	m = re.search(r"自动|自動", configjs)
+	s = balance(configjs, m.start(), "[", "]")
+	servs = eval(s)
 	servs = [host["h"] for category in servs for host in category["hosts"]]
 
 	global servers

comiccrawler/util.py

@@ -33,15 +33,15 @@ def create_safefilepath_table():
 		":": "：",
 		"\"": "＂",
 		"*": "＊"
-	})
+		})
 	table.update({
 		c: None for c in set(chr(i) for i in range(128)).difference(string.printable)
-	})
+		})
 	table.update({
 		chr(i): " " for i in range(32) if chr(i) not in table
-	})
+		})
 	return str.maketrans(table)
-	
+
 safefilepath_table = create_safefilepath_table()
 dot_table = str.maketrans({".": "．"})
 
@@ -71,8 +71,39 @@ def clean_tags(html):
 class MinimumAny:
 	def __le__(self, other):
 		return True
-		
+
 	def __eq__(self, other):
 		return self is other
 
 MIN = MinimumAny()
+
+def balance(s: str, index: int, left="(", right=")", skip=0):
+	"""Return the string inside (including) matched left and right brackets."""
+	# backward search
+	count = 0
+	for i in range(index, -1, -1):
+		if s[i] == right:
+			count += 1
+		elif s[i] == left:
+			if count == -skip:
+				break
+			count -= 1
+	else:
+		raise ValueError(f"Unbalanced brackets: {s}")
+	start = i
+
+	# forward search
+	count = 0
+	for j in range(index, len(s)):
+		if s[j] == left:
+			count += 1
+		elif s[j] == right:
+			if count == -skip:
+				break
+			count -= 1
+	else:
+		raise ValueError(f"Unbalanced brackets: {s}")
+	end = j + 1
+
+	return s[start:end]
+