Fix S3 storage to media in whisper_enricher.py.

src/auto_archiver/modules/whisper_enricher/__manifest__.py

@@ -1,4 +1,4 @@
-{
+a={
     "name": "Whisper Enricher",
     "type": ["enricher"],
     "requires_setup": True,
@@ -12,7 +12,9 @@
                     "help": "WhisperApi api key for authentication"},
         "include_srt": {"default": False, "help": "Whether to include a subtitle SRT (SubRip Subtitle file) for the video (can be used in video players)."},
         "timeout": {"default": 90, "help": "How many seconds to wait at most for a successful job completion."},
-        "action": {"default": "translate", "help": "which Whisper operation to execute", "choices": ["transcribe", "translate", "language_detection"]},
+        "action": {"default": "translate",
+                   "help": "which Whisper operation to execute",
+                   "choices": ["transcribe", "translate", "language_detection"]},
     },
     "description": """
     Integrates with a Whisper API service to transcribe, translate, or detect the language of audio and video files.
@@ -27,6 +29,7 @@
     ### Notes
     - Requires a Whisper API endpoint and API key for authentication.
     - Only compatible with S3-compatible storage systems for media file accessibility.
+    - ** This stores the media files in S3 prior to enriching them as Whisper requires public URLs to access the media files.
     - Handles multiple jobs and retries for failed or incomplete processing.
     """
 }

src/auto_archiver/modules/whisper_enricher/whisper_enricher.py

@@ -15,17 +15,21 @@ class WhisperEnricher(Enricher):
     """
 
     def enrich(self, to_enrich: Metadata) -> None:
-        if not self._get_s3_storage():
+        storages = self.config['steps']['storages']
+        if not "s3_storage" in storages:
             logger.error("WhisperEnricher: To use the WhisperEnricher you need to use S3Storage so files are accessible publicly to the whisper service being called.")
             return
 
+        self.s3 = get_module("s3_storage", self.config)
         url = to_enrich.get_url()
         logger.debug(f"WHISPER[{self.action}]: iterating media items for {url=}.")
 
         job_results = {}
         for i, m in enumerate(to_enrich.media):
             if m.is_video() or m.is_audio():
-                m.store(url=url, metadata=to_enrich, storages=self.storages)
+                # TODO: this used to pass all storage items to store now
+                # Now only passing S3, the rest will get added later in the usual order (?)
+                m.store(url=url, metadata=to_enrich, storages=[self.s3])
                 try:
                     job_id = self.submit_job(m)
                     job_results[job_id] = False
@@ -53,8 +57,8 @@ class WhisperEnricher(Enricher):
                             to_enrich.set_content(f"\n[automatic video transcript]: {v}")
 
     def submit_job(self, media: Media):
-        s3 = get_module("s3_storage", self.config)
+
-        s3_url = s3.get_cdn_url(media)
+        s3_url = self.s3.get_cdn_url(media)
         assert s3_url in media.urls, f"Could not find S3 url ({s3_url}) in list of stored media urls "
         payload = {
             "url": s3_url,
@@ -107,10 +111,3 @@ class WhisperEnricher(Enricher):
             logger.debug(f"DELETE whisper {job_id=} result: {r_del.status_code}")
             return result
         return False
-
-    def _get_s3_storage(self) -> S3Storage:
-        try:
-            return next(s for s in self.config['steps']['storages'] if s == 's3_storage')
-        except:
-            logger.warning("No S3Storage instance found in storages")
-            return