thumbnails enricher

src/enrichers/__init__.py

@@ -1,4 +1,5 @@
 from .enricher import Enricher
 from .screenshot_enricher import ScreenshotEnricher 
 from .wayback_enricher import WaybackEnricher
-from .hash_enricher import HashEnricher
+from .hash_enricher import HashEnricher
+from .thumbnail_enricher import ThumbnailEnricher

src/enrichers/thumbnail_enricher.py

@@ -0,0 +1,46 @@
+import uuid
+from media import Media
+from . import Enricher
+from metadata import Metadata
+from loguru import logger
+import ffmpeg, os
+
+
+class ThumbnailEnricher(Enricher):
+    """
+    Generates thumbnails for all the media
+    """
+    name = "thumbnail_enricher"
+
+    def __init__(self, config: dict) -> None:
+        # without this STEP.__init__ is not called
+        super().__init__(config)
+
+    @staticmethod
+    def configs() -> dict:
+        return {}
+
+    def enrich(self, to_enrich: Metadata) -> None:
+        logger.debug(f"generating thumbnails")
+        folder = os.path.join(to_enrich.get_tmp_dir(), str(uuid.uuid4()))
+        os.makedirs(folder, exist_ok=True)
+        for i, m in enumerate(to_enrich.media[::]):
+            if m.is_video():
+                logger.debug(f"generating thumbnails for {m.filename}")
+                fps, duration = 0.5, m.get("duration")
+                if duration is not None:
+                    duration = float(duration)
+                    if duration < 60: fps = 10.0 / duration
+                    elif duration < 120: fps = 20.0 / duration
+                    else: fps = 40.0 / duration
+
+                stream = ffmpeg.input(m.filename)
+                stream = ffmpeg.filter(stream, 'fps', fps=fps).filter('scale', 512, -1)
+                stream.output(os.path.join(folder, 'out%d.jpg')).run()
+
+                thumbnails = os.listdir(folder)
+                thumbnails_media = []
+                for t, fname in enumerate(thumbnails):
+                    if fname[-3:] == 'jpg':
+                        thumbnails_media.append(Media(filename=os.path.join(folder, fname)).set("id", f"thumbnail_{t}"))
+                to_enrich.media[i].set("thumbnails", thumbnails_media)

src/enrichers/wayback_enricher.py

@@ -37,7 +37,7 @@ class WaybackEnricher(Enricher):
         r = requests.post('https://web.archive.org/save/', headers=ia_headers, data={'url': url})
 
         if r.status_code != 200:
-            logger.error(em:=f"Internet archive failed with status of {r.status_code}: {r.json()}")
+            logger.error(em := f"Internet archive failed with status of {r.status_code}: {r.json()}")
             to_enrich.set("wayback", em)
             return
 
@@ -66,3 +66,4 @@ class WaybackEnricher(Enricher):
             to_enrich.set("wayback", wayback_url)
         else:
             to_enrich.set("wayback", {"job_id": job_id, "check_status": f'https://web.archive.org/save/status/{job_id}'})
+        to_enrich.set("wayback lookup", f"https://web.archive.org/web/*/{url}")