Try to improve speed for marking unfurl

Signed-off-by: Tushar Goel <tushar.goel.dav@gmail.com>

Author: TG1999

CHANGELOG.rst

@@ -5,9 +5,11 @@ next release
 ---------------------
 
 - WARNING: Vulnerablecode V1 API and UI has stopped supporting Ubuntu OVAL advisories, please shift to V3 API for new Ubuntu advisories.
-- WARNING: We will deprecate improver pipelines for calculating package version rank and grouping advisories for packages in the next release, we are doing it at advisory import time instead of as separate pipelines, this will improve the performance and consistency of the data.
-- Calculate package verion rank, group advisories for packages and package risk score and advisory risk score during import of advisories.
+- WARNING: We will deprecate improver pipelines for calculating package version rank and grouping advisories for packages.
 - Add attribute ``pipeline_id`` to AdvisoryV2 to track the pipeline that created the advisory, also rename existing ``datasource_id`` and AVIDs.
+- We will group advisories for packages and calculate package risk score and advisory risk score only when an advisory is completely unfurled,
+  this will improve consistency of the data. In future if we change our grouping approach we have to make a migration to mark all advisories 
+  as not unfurled, so they can be goruped again with the new algo.
 
 
 Version v38.6.0

vulnerabilities/pipelines/v2_improvers/mark_unfurl_version_range.py

@@ -44,7 +44,7 @@ def mark_all_impacts_unfurled(self):
             impacted_packages=impacted_packages
         )
 
-        batch_size = 100
+        batch_size = 1000
         batch = []
 
         successful_qs = latest_advisories_with_all_impacts_unfurled_successfully(
@@ -106,7 +106,7 @@ def latest_advisories_with_all_impacts_unfurled_attempted(
         )
         .annotate(has_unattempted_impacts=Exists(impacts_not_attempted))
         .filter(has_unattempted_impacts=False)
-        .order_by("_all_impacts_unfurled", "id")
+        .order_by("_all_impacts_unfurled", "datasource_id")
         .values_list("id", flat=True)
     )