Bump version to v0.6.4 for release

Pipfile.lock

@@ -1643,11 +1643,11 @@
         },
         "vk-url-scraper": {
             "hashes": [
-                "sha256:dc1c5cebc2919620935be0b669bef899dca3b8a3ef6419a81c97d03903b012a5",
-                "sha256:edf155fde3e86950f715a4de09cbe975cf220cb1fcc2fa239dd27868d991ffb3"
+                "sha256:b81caa8738b0154aca25ab33012e8e4d8dc11cdc9fd371968afbbadd57e2e1ab",
+                "sha256:d12aec2828e8e47c766991207fd809e402ab832ea449129b94ac7eadb0d1a27c"
             ],
             "index": "pypi",
-            "version": "==0.3.24"
+            "version": "==0.3.26"
         },
         "warcio": {
             "hashes": [

src/auto_archiver/enrichers/wayback_enricher.py

@@ -23,6 +23,7 @@ class WaybackArchiverEnricher(Enricher, Archiver):
     def configs() -> dict:
         return {
             "timeout": {"default": 15, "help": "seconds to wait for successful archive confirmation from wayback, if more than this passes the result contains the job_id so the status can later be checked manually."},
+            "if_not_archived_within": {"default": None, "help": "only tell wayback to archive if no archive is available before the number of seconds specified, use None to ignore this option. For more information: https://docs.google.com/document/d/1Nsv52MvSjbLb2PCpHlat0gkzw0EvtSgpKHu4mk0MnrA"},
             "key": {"default": None, "help": "wayback API key. to get credentials visit https://archive.org/account/s3.php"},
             "secret": {"default": None, "help": "wayback API secret. to get credentials visit https://archive.org/account/s3.php"}
         }
@@ -50,7 +51,11 @@ class WaybackArchiverEnricher(Enricher, Archiver):
             "Accept": "application/json",
             "Authorization": f"LOW {self.key}:{self.secret}"
         }
-        r = requests.post('https://web.archive.org/save/', headers=ia_headers, data={'url': url})
+        post_data = {'url': url}
+        if self.if_not_archived_within:
+            post_data["if_not_archived_within"] = self.if_not_archived_within
+        # see https://docs.google.com/document/d/1Nsv52MvSjbLb2PCpHlat0gkzw0EvtSgpKHu4mk0MnrA for more options
+        r = requests.post('https://web.archive.org/save/', headers=ia_headers, data=post_data)
 
         if r.status_code != 200:
             logger.error(em := f"Internet archive failed with status of {r.status_code}: {r.json()}")

src/auto_archiver/enrichers/whisper_enricher.py

@@ -57,9 +57,12 @@ class WhisperEnricher(Enricher):
 
         for i, m in enumerate(to_enrich.media):
             if m.is_video() or m.is_audio():
-                job_id = to_enrich.media[i].get("whisper_model")["job_id"]
+                job_id = to_enrich.media[i].get("whisper_model", {}).get("job_id")
+                if not job_id: continue
                 to_enrich.media[i].set("whisper_model", {
                     "job_id": job_id,
+                    "job_status_check": f"{self.api_endpoint}/jobs/{job_id}",
+                    "job_artifacts_check": f"{self.api_endpoint}/jobs/{job_id}/artifacts",
                     **(job_results[job_id] if job_results[job_id] else {"result": "incomplete or failed job"})
                 })
                 # append the extracted text to the content of the post so it gets written to the DBs like gsheets text column

src/auto_archiver/utils/url.py

@@ -57,6 +57,14 @@ class UrlUtil:
         if "https://yt3.ggpht.com" in url and "default-user=" in url: return False
         if "https://www.youtube.com/s/search/audio/" in url: return False
 
+        # ok
+        if " https://ok.ru/res/i/" in url: return False
+
+        # vk
+        if "https://vk.com/emoji/" in url: return False
+        if "vk.com/images/" in url: return False
+        if "vk.com/images/reaction/" in url: return False
+
         return True
 
     @staticmethod

src/auto_archiver/version.py

@@ -3,7 +3,7 @@ _MAJOR = "0"
 _MINOR = "6"
 # On main and in a nightly release the patch should be one ahead of the last
 # released build.
-_PATCH = "3"
+_PATCH = "4"
 # This is mainly for nightly builds which have the suffix ".dev$DATE". See
 # https://semver.org/#is-v123-a-semantic-version for the semantics.
 _SUFFIX = ""