Add limit to unassociate [RHELDST-20725]

We want to perform garbage collection in batches to reduce the amount of
resources it consumes and avoid out of memory exceptions. Unit
association criteria have limit and skip fields which we could use for
batching requests.

pubtools/pulplib/_impl/client/client.py

@@ -775,7 +775,7 @@ def _do_associate(self, src_repo_id, dest_repo_id, criteria=None, raw_options=No
             self._do_request, method="POST", url=url, json=body
         )
 
-    def _do_unassociate(self, repo_id, criteria=None):
+    def _do_unassociate(self, repo_id, criteria=None, limit=None, skip=None):
         url = os.path.join(
             self._url, "pulp/api/v2/repositories/%s/actions/unassociate/" % repo_id
         )
@@ -800,6 +800,12 @@ def _do_unassociate(self, repo_id, criteria=None):
             else:
                 body["criteria"]["filters"] = {"unit": pulp_search.filters}
 
+        if limit:
+            body["criteria"]["limit"] = limit
+
+        if skip:
+            body["criteria"]["skip"] = skip
+
         LOG.debug("Submitting %s unassociate: %s", url, body)
 
         return self._task_executor.submit(

pubtools/pulplib/_impl/model/repository/base.py

@@ -696,7 +696,14 @@ def remove_content(self, criteria=None, **kwargs):
                     Matcher.in_(type_ids),  # Criteria.with_field_in is deprecated
                 )
 
-        return f_proxy(self._client._do_unassociate(self.id, criteria=criteria))
+        return f_proxy(
+            self._client._do_unassociate(
+                self.id,
+                criteria=criteria,
+                limit=kwargs.get("limit"),
+                skip=kwargs.get("skip"),
+            )
+        )
 
     @classmethod
     def from_data(cls, data):