[reddit] rewrite extractor logic (closes #551)

Handle images and videos hosted on Reddit "natively",
allowing them to use reddit-specific metadata to build directory
and file names.

gallery_dl/extractor/reddit.py

@@ -9,74 +9,88 @@
 """Extractors for https://www.reddit.com/"""
 
 from .common import Extractor, Message
-from .. import text, util, extractor, exception
+from .. import text, util, exception
 from ..cache import cache
 
 
 class RedditExtractor(Extractor):
     """Base class for reddit extractors"""
     category = "reddit"
+    directory_fmt = ("{category}", "{subreddit}")
+    filename_fmt = "{id} {title[:242]}.{extension}"
+    archive_fmt = "{url}"
     cookiedomain = None
 
     def __init__(self, match):
         Extractor.__init__(self, match)
         self.api = RedditAPI(self)
-        self.max_depth = int(self.config("recursion", 0))
-        self._visited = set()
+        self.max_depth = self.config("recursion", 0)
 
     def items(self):
-        subre = RedditSubmissionExtractor.pattern
+        match_submission = RedditSubmissionExtractor.pattern.match
+        match_subreddit = RedditSubredditExtractor.pattern.match
+        match_user = RedditUserExtractor.pattern.match
+
         submissions = self.submissions()
+        visited = set()
         depth = 0
 
         yield Message.Version, 1
-        with extractor.blacklist(
-                util.SPECIAL_EXTRACTORS,
-                [RedditSubredditExtractor, RedditUserExtractor]):
-            while True:
-                extra = []
-                for url, data in self._urls(submissions):
-                    if url[0] == "#":
+
+        while True:
+            extra = []
+
+            for submission, comments in submissions:
+                urls = []
+
+                if submission:
+                    yield Message.Directory, submission
+                    visited.add(submission["id"])
+                    url = submission["url"]
+
+                    if url.startswith("https://i.redd.it/"):
+                        text.nameext_from_url(url, submission)
+                        yield Message.Url, url, submission
+                    elif submission["is_video"]:
+                        submission["extension"] = None
+                        yield Message.Url, "ytdl:" + url, submission
+                    elif not submission["is_self"]:
+                        urls.append((url, submission))
+
+                if self.api.comments:
+                    if submission:
+                        for url in text.extract_iter(
+                                submission["selftext_html"] or "",
+                                ' href="', '"'):
+                            urls.append((url, submission))
+                    for comment in comments:
+                        for url in text.extract_iter(
+                                comment["body_html"] or "", ' href="', '"'):
+                            urls.append((url, comment))
+
+                for url, data in urls:
+                    if not url or url[0] == "#":
                         continue
                     if url[0] == "/":
                         url = "https://www.reddit.com" + url
 
-                    match = subre.match(url)
+                    match = match_submission(url)
                     if match:
                         extra.append(match.group(1))
-                    else:
+                    elif not match_user(url) and not match_subreddit(url):
                         yield Message.Queue, text.unescape(url), data
 
-                if not extra or depth == self.max_depth:
-                    return
-                depth += 1
-                submissions = (
-                    self.api.submission(sid) for sid in extra
-                    if sid not in self._visited
-                )
+            if not extra or depth == self.max_depth:
+                return
+            depth += 1
+            submissions = (
+                self.api.submission(sid) for sid in extra
+                if sid not in self._visited
+            )
 
     def submissions(self):
         """Return an iterable containing all (submission, comments) tuples"""
 
-    def _urls(self, submissions):
-        for submission, comments in submissions:
-
-            if submission:
-                self._visited.add(submission["id"])
-
-                if not submission["is_self"]:
-                    yield submission["url"], submission
-
-                for url in text.extract_iter(
-                        submission["selftext_html"] or "", ' href="', '"'):
-                    yield url, submission
-
-            if comments:
-                for comment in comments:
-                    for url in text.extract_iter(
-                            comment["body_html"] or "", ' href="', '"'):
-                        yield url, comment
-
 
 class RedditSubredditExtractor(RedditExtractor):
     """Extractor for URLs from subreddits on reddit.com"""
@@ -84,7 +98,10 @@ class RedditSubredditExtractor(RedditExtractor):
     pattern = (r"(?:https?://)?(?:\w+\.)?reddit\.com/r/"
                r"([^/?&#]+(?:/[a-z]+)?)/?(?:\?([^#]*))?(?:$|#)")
     test = (
-        ("https://www.reddit.com/r/lavaporn/"),
+        ("https://www.reddit.com/r/lavaporn/", {
+            "range": "1-20",
+            "count": ">= 20",
+        }),
         ("https://www.reddit.com/r/lavaporn/top/?sort=top&t=month"),
         ("https://old.reddit.com/r/lavaporn/"),
         ("https://np.reddit.com/r/lavaporn/"),
@@ -210,7 +227,7 @@ class RedditAPI():
         link_id = "t3_" + submission_id if self.morecomments else None
         submission, comments = self._call(endpoint, {"limit": self.comments})
         return (submission["data"]["children"][0]["data"],
-                self._flatten(comments, link_id) if self.comments else None)
+                self._flatten(comments, link_id) if self.comments else ())
 
     def submissions_subreddit(self, subreddit, params):
         """Collect all (submission, comments)-tuples of a subreddit"""
@@ -290,7 +307,8 @@ class RedditAPI():
                 raise exception.AuthorizationError()
             if data["error"] == 404:
                 raise exception.NotFoundError()
-            raise Exception(data["message"])
+            self.log.debug(data)
+            raise exception.StopExtraction(data.get("message"))
         return data
 
     def _pagination(self, endpoint, params):
@@ -315,7 +333,7 @@ class RedditAPI():
                             except exception.AuthorizationError:
                                 pass
                         else:
-                            yield post, None
+                            yield post, ()
 
                     elif kind == "t1" and self.comments:
                         yield None, (post,)