maintenance: simplify code (#79)

* simplify code

* delete line

courlan/cli.py

@@ -151,33 +151,31 @@ def _cli_process(args: Any) -> None:
                     break
                 batches.append(line_batch)
 
-            if batches:
-                futures = (
-                    executor.submit(
-                        _cli_check_urls,
-                        batch,
-                        strict=args.strict,
-                        with_redirects=args.redirects,
-                        language=args.language,
-                    )
-                    for batch in batches
-                )
-
-                for future in as_completed(futures):
-                    for valid, url in future.result():
-                        if valid:
-                            outputfh.write(url + "\n")
-                        # proceed with discarded URLs. to be rewritten
-                        elif args.discardedfile is not None:
-                            with open(
-                                args.discardedfile, "a", encoding="utf-8"
-                            ) as discardfh:
-                                discardfh.write(url)
-
-                batches = []
-            else:
+            if not batches:
                 break
 
+            futures = (
+                executor.submit(
+                    _cli_check_urls,
+                    batch,
+                    strict=args.strict,
+                    with_redirects=args.redirects,
+                    language=args.language,
+                )
+                for batch in batches
+            )
+
+            for future in as_completed(futures):
+                for valid, url in future.result():
+                    if valid:
+                        outputfh.write(url + "\n")
+                    # proceed with discarded URLs. to be rewritten
+                    elif args.discardedfile is not None:
+                        with open(
+                            args.discardedfile, "a", encoding="utf-8"
+                        ) as discardfh:
+                            discardfh.write(url)
+
 
 def process_args(args: Any) -> None:
     """Start processing according to the arguments"""

courlan/filters.py

@@ -167,17 +167,12 @@ def domain_filter(domain: str) -> bool:
         return False
 
     # unsuitable content
-    if UNSUITABLE_DOMAIN.match(domain):
-        return False
-
-    if FILE_TYPE.search(domain):
+    if UNSUITABLE_DOMAIN.match(domain) or FILE_TYPE.search(domain):
         return False
 
+    # extensions
     extension_match = EXTENSION_REGEX.search(domain)
-    if extension_match and extension_match[0] in WHITELISTED_EXTENSIONS:
-        return False
-
-    return True
+    return not extension_match or extension_match[0] not in WHITELISTED_EXTENSIONS
 
 
 def extension_filter(urlpath: str) -> bool:

courlan/urlutils.py

@@ -58,14 +58,12 @@ def extract_domain(
         blacklist = set()
     # new code: Python >= 3.6 with tld module
     domain, full_domain = get_tldinfo(url, fast=fast)
-    # invalid input
-    if full_domain is None:
-        return None
-    # blacklisting
-    if domain in blacklist or full_domain in blacklist:
-        return None
-    # return domain
-    return full_domain
+
+    return (
+        full_domain
+        if full_domain and not domain in blacklist and not full_domain in blacklist
+        else None
+    )
 
 
 def _parse(url: Any) -> SplitResult: