[bluesky] extract 'hashtags', 'mentions', and 'uris' metadata (#4438)

gallery_dl/extractor/bluesky.py

@@ -45,6 +45,21 @@ class BlueskyExtractor(Extractor):
                 if "images" in media:
                     images = media["images"]
 
+            if "facets" in post:
+                post["hashtags"] = tags = []
+                post["mentions"] = dids = []
+                post["uris"] = uris = []
+                for facet in post["facets"]:
+                    features = facet["features"][0]
+                    if "tag" in features:
+                        tags.append(features["tag"])
+                    elif "did" in features:
+                        dids.append(features["did"])
+                    elif "uri" in features:
+                        uris.append(features["uri"])
+            else:
+                post["hashtags"] = post["mentions"] = post["uris"] = ()
+
             post["post_id"] = post["uri"].rpartition("/")[2]
             post["count"] = len(images)
             post["date"] = text.parse_datetime(

test/results/bluesky.py

@@ -39,6 +39,51 @@ __tests__ = (
     "repostCount": int,
     "uri"        : "at://did:plc:z72i7hdynmk6r22z27h6tvur/app.bsky.feed.post/3kh5rarr3gn2n",
     "width"      : 1200,
+    "hashtags"   : [],
+    "mentions"   : [],
+    "uris"       : ["https://blueskyweb.xyz/blog/12-21-2023-butterfly"],
+},
+
+{
+    "#url"     : "https://bsky.app/profile/mikf.bsky.social/post/3kkzc3xaf5m2w",
+    "#category": ("", "bluesky", "post"),
+    "#class"   : bluesky.BlueskyPostExtractor,
+    "#urls"        : "https://bsky.social/xrpc/com.atproto.sync.getBlob?did=did:plc:cslxjqkeexku6elp5xowxkq7&cid=bafkreib7ydpe3xxo4cq7nn32w7eqhcanfaanz6caepd2z4kzplxtx2ctgi",
+    "#sha1_content": "9cf5748f6d00aae83fbb3cc2c6eb3caa832b90f4",
+
+    "author": {
+        "did"        : "did:plc:cslxjqkeexku6elp5xowxkq7",
+        "displayName": "mikf",
+        "handle"     : "mikf.bsky.social",
+        "labels"     : [],
+    },
+    "cid"        : "bafyreihtck7clocti2qshaiounadof74pxqhz7gnvbstxujqzhlodigqru",
+    "count"      : 1,
+    "createdAt"  : "2024-02-09T21:57:31.917Z",
+    "date"       : "dt:2024-02-09 21:57:31",
+    "description": "reading lewd books",
+    "extension"  : "jpeg",
+    "filename"   : "bafkreib7ydpe3xxo4cq7nn32w7eqhcanfaanz6caepd2z4kzplxtx2ctgi",
+    "hashtags"   : [
+        "patchouli",
+        "patchy",
+    ],
+    "mentions"   : [
+        "did:plc:cslxjqkeexku6elp5xowxkq7",
+    ],
+    "uris"       : [
+        "https://seiga.nicovideo.jp/seiga/im5977527",
+    ],
+    "width"      : 1024,
+    "height"     : 768,
+    "langs"      : ["en"],
+    "likeCount"  : int,
+    "num"        : 1,
+    "post_id"    : "3kkzc3xaf5m2w",
+    "replyCount" : int,
+    "repostCount": int,
+    "text"       : "testing \"facets\"\n\nsource: seiga.nicovideo.jp/seiga/im5977...\n#patchouli #patchy\n@mikf.bsky.social",
+    "uri"        : "at://did:plc:cslxjqkeexku6elp5xowxkq7/app.bsky.feed.post/3kkzc3xaf5m2w",
 },
 
 )