Further cleanup

* Removes (partly) the ArchivingOrchestrator
* Removes the cli_feeder module, and makes it the 'default', allowing you to pass URLs directly on the command line, without having to use the cumbersome --cli_feeder.urls. Just do auto-archiver https://my.url.com
* More unit tests
* Improved error handling
This commit is contained in:
Patrick Robertson
2025-01-30 16:43:09 +01:00
parent 953011f368
commit d6b4b7a932
27 changed files with 417 additions and 191 deletions

View File

@@ -23,6 +23,6 @@ class SSLEnricher(Enricher):
logger.debug(f"fetching SSL certificate for {domain=} in {url=}")
cert = ssl.get_server_certificate((domain, 443))
cert_fn = os.path.join(ArchivingContext.get_tmp_dir(), f"{slugify(domain)}.pem")
cert_fn = os.path.join(self.tmp_dir, f"{slugify(domain)}.pem")
with open(cert_fn, "w") as f: f.write(cert)
to_enrich.add_media(Media(filename=cert_fn), id="ssl_certificate")