The only issue with this approach is that there is no guarantee that a timestamp corresponds to a single visit: we can make several passes over an origin with the same (virtual) visit timestamp, for instance:

• when we have downloaded the contents of an origin for further processing, the visit timestamp is the time we downloaded the data; if we fail to import for some reason, and we try again, two visits get the same timestamp.
• when we have listed the debian archive, the visit timestamp is the time we fetched the indices; we can then process the packages in several (partial) passes before succeeding.

Although the latest incremental visit for a given timestamp is probably the most interesting, in the general case we will need a way to disambiguate those visits.

As requested here is an example of an origin where we struggle to make the full visit on the first pass, and therefore have a bunch of visits with the same timestamp:

https://archive.softwareheritage.org/api/1/origin/65623577/

New swh-web version using intrinsic info is now deployed to moma.

In order to disambiguate origin visits with the same date, I set up a query parameter visit_id
to retrieve the adequate visit for such cases (see https://archive.softwareheritage.org/browse/origin/deb/url/deb://Debian/packages/linux/).
For a given date, the default behaviour is to fetch the latest visit with that date.