|
| 1 | +"""Helpers for repository package catalog, metrics, and rebuild collapse.""" |
| 2 | + |
| 3 | +from collections import defaultdict |
| 4 | + |
| 5 | +from django.db.models import CharField, Func, Max, Min, Q, Value |
| 6 | +from django.db.models.functions import Coalesce |
| 7 | + |
| 8 | +from pulp_python.app.models import PythonPackageContent |
| 9 | +from pulp_python.app.versions import ( |
| 10 | + BUILD_SUFFIX_PATTERN, |
| 11 | + normalize_package_index_ordering, |
| 12 | + rebuild_release, |
| 13 | + version_sort_key, |
| 14 | +) |
| 15 | + |
| 16 | + |
| 17 | +def base_version_annotation(field_name="version"): |
| 18 | + """SQL expression that strips a trailing rebuild suffix from ``version``. |
| 19 | +
|
| 20 | + Uses ``versions.BUILD_SUFFIX_PATTERN`` (POSIX) so Python ``strip_build_suffix`` |
| 21 | + and this ``REGEXP_REPLACE`` stay aligned. Implemented with ``REGEXP_REPLACE`` |
| 22 | + so it does not depend on Django's ``RegexpReplace`` (not present in every |
| 23 | + Django 4.2/5.2 packaging Pulp uses). |
| 24 | + """ |
| 25 | + return Func( |
| 26 | + field_name, |
| 27 | + Value(BUILD_SUFFIX_PATTERN), |
| 28 | + Value(""), |
| 29 | + function="REGEXP_REPLACE", |
| 30 | + output_field=CharField(), |
| 31 | + ) |
| 32 | + |
| 33 | + |
| 34 | +def collapse_python_builds(queryset): |
| 35 | + """Keep one content unit per ``(name_normalized, base_version)``. |
| 36 | +
|
| 37 | + ``base_version`` is ``version`` with a trailing rebuild suffix stripped. |
| 38 | + The unit with the latest ``pulp_created`` is kept. Callers that want one |
| 39 | + row per logical version (not per wheel/sdist) should also filter |
| 40 | + ``packagetype``. |
| 41 | + """ |
| 42 | + return ( |
| 43 | + queryset.prefetch_related(None) |
| 44 | + .annotate(_collapse_base_version=base_version_annotation()) |
| 45 | + .order_by("name_normalized", "_collapse_base_version", "-pulp_created") |
| 46 | + .distinct("name_normalized", "_collapse_base_version") |
| 47 | + ) |
| 48 | + |
| 49 | + |
| 50 | +def python_packages_in_version(repository_version): |
| 51 | + """Python package content contained in ``repository_version``.""" |
| 52 | + if repository_version is None: |
| 53 | + return PythonPackageContent.objects.none() |
| 54 | + return PythonPackageContent.objects.filter(pk__in=repository_version.content) |
| 55 | + |
| 56 | + |
| 57 | +def apply_package_prefix_filters( |
| 58 | + queryset, |
| 59 | + name_normalized_prefix=None, |
| 60 | + name_prefix=None, |
| 61 | + name_normalized_contains=None, |
| 62 | +): |
| 63 | + """Apply case-insensitive name filters used by the package index.""" |
| 64 | + if name_normalized_prefix: |
| 65 | + queryset = queryset.filter(name_normalized__startswith=name_normalized_prefix) |
| 66 | + if name_normalized_contains: |
| 67 | + queryset = queryset.filter(name_normalized__contains=name_normalized_contains) |
| 68 | + if name_prefix: |
| 69 | + queryset = queryset.filter(name__istartswith=name_prefix) |
| 70 | + return queryset |
| 71 | + |
| 72 | + |
| 73 | +def membership_in_version_q(repository, repository_version): |
| 74 | + """Q-object matching RepositoryContent rows present in ``repository_version``.""" |
| 75 | + return Q( |
| 76 | + version_memberships__repository=repository, |
| 77 | + version_memberships__version_added__number__lte=repository_version.number, |
| 78 | + ) & ( |
| 79 | + Q(version_memberships__version_removed__isnull=True) |
| 80 | + | Q(version_memberships__version_removed__number__gt=repository_version.number) |
| 81 | + ) |
| 82 | + |
| 83 | + |
| 84 | +def last_updated_annotation(repository, repository_version): |
| 85 | + """Newest repository-membership time among all package units for a name. |
| 86 | +
|
| 87 | + Uses ``RepositoryContent.pulp_created`` (any rebuild/build), falling back to |
| 88 | + the content unit's ``pulp_created``. |
| 89 | + """ |
| 90 | + return Coalesce( |
| 91 | + Max( |
| 92 | + "version_memberships__pulp_created", |
| 93 | + filter=membership_in_version_q(repository, repository_version), |
| 94 | + ), |
| 95 | + Max("pulp_created"), |
| 96 | + ) |
| 97 | + |
| 98 | + |
| 99 | +def distinct_package_names_qs(content_qs, repository, repository_version, ordering=None): |
| 100 | + """One row per distinct ``name_normalized``, ordered for stable pagination.""" |
| 101 | + if ordering is None: |
| 102 | + ordering = normalize_package_index_ordering([]) |
| 103 | + qs = content_qs.order_by().values("name_normalized").annotate(name=Max("name")) |
| 104 | + if repository_version is None: |
| 105 | + qs = qs.annotate(last_updated=Max("pulp_created")) |
| 106 | + else: |
| 107 | + qs = qs.annotate(last_updated=last_updated_annotation(repository, repository_version)) |
| 108 | + return qs.order_by(*ordering) |
| 109 | + |
| 110 | + |
| 111 | +def assemble_package_index(content_qs, name_rows, repository, repository_version): |
| 112 | + """Build package-index dicts for ``name_rows``. |
| 113 | +
|
| 114 | + ``versions`` are distinct logical versions, newest first (PEP 440). |
| 115 | + ``latest_releases`` keeps the newest rebuild (latest ``pulp_created``) |
| 116 | + per base version in the same order. ``created_at`` is that unit's |
| 117 | + repository-membership time (``RepositoryContent.pulp_created``), falling |
| 118 | + back to the content unit's ``pulp_created``. ``last_updated`` is the newest |
| 119 | + membership among all units for the package (any rebuild), taken from |
| 120 | + ``name_rows`` when annotated. |
| 121 | + """ |
| 122 | + if not name_rows or repository_version is None: |
| 123 | + return [] |
| 124 | + |
| 125 | + names = [row["name_normalized"] for row in name_rows] |
| 126 | + name_by_normalized = {row["name_normalized"]: row["name"] for row in name_rows} |
| 127 | + |
| 128 | + in_this_version = membership_in_version_q(repository, repository_version) |
| 129 | + |
| 130 | + newest_units = list( |
| 131 | + content_qs.filter(name_normalized__in=names) |
| 132 | + .prefetch_related(None) |
| 133 | + .annotate(_base_version=base_version_annotation()) |
| 134 | + .order_by("name_normalized", "_base_version", "-pulp_created") |
| 135 | + .distinct("name_normalized", "_base_version") |
| 136 | + ) |
| 137 | + newest = [ |
| 138 | + { |
| 139 | + "pk": unit.pk, |
| 140 | + "name_normalized": unit.name_normalized, |
| 141 | + "version": unit.version, |
| 142 | + "_base_version": unit._base_version, |
| 143 | + "pulp_created": unit.pulp_created, |
| 144 | + } |
| 145 | + for unit in newest_units |
| 146 | + ] |
| 147 | + |
| 148 | + memberships = {} |
| 149 | + if newest: |
| 150 | + memberships = dict( |
| 151 | + PythonPackageContent.objects.filter(pk__in=[row["pk"] for row in newest]) |
| 152 | + .annotate( |
| 153 | + membership_created=Min( |
| 154 | + "version_memberships__pulp_created", |
| 155 | + filter=in_this_version, |
| 156 | + ) |
| 157 | + ) |
| 158 | + .values_list("pk", "membership_created") |
| 159 | + ) |
| 160 | + |
| 161 | + releases_by_name = defaultdict(list) |
| 162 | + for rel in newest: |
| 163 | + releases_by_name[rel["name_normalized"]].append(rel) |
| 164 | + |
| 165 | + result = [] |
| 166 | + for row in name_rows: |
| 167 | + normalized = row["name_normalized"] |
| 168 | + rels = sorted( |
| 169 | + releases_by_name.get(normalized, []), |
| 170 | + key=lambda item: version_sort_key(item["_base_version"]), |
| 171 | + reverse=True, |
| 172 | + ) |
| 173 | + versions = [item["_base_version"] for item in rels] |
| 174 | + latest_releases = [ |
| 175 | + { |
| 176 | + "version": item["_base_version"], |
| 177 | + "release": rebuild_release(item["version"]), |
| 178 | + "created_at": memberships.get(item["pk"]) or item["pulp_created"], |
| 179 | + } |
| 180 | + for item in rels |
| 181 | + ] |
| 182 | + result.append( |
| 183 | + { |
| 184 | + "name": name_by_normalized[normalized], |
| 185 | + "name_normalized": normalized, |
| 186 | + "last_updated": row.get("last_updated"), |
| 187 | + "versions": versions, |
| 188 | + "latest_releases": latest_releases, |
| 189 | + } |
| 190 | + ) |
| 191 | + return result |
| 192 | + |
| 193 | + |
| 194 | +def repository_metrics(content_qs): |
| 195 | + """Distinct package / logical-version / build counts for package content. |
| 196 | +
|
| 197 | + Identity is always ``PythonPackageContent`` (not filtered by packagetype): |
| 198 | +
|
| 199 | + * ``package_count``: distinct ``name_normalized`` |
| 200 | + * ``version_count``: distinct ``(name_normalized, base_version)`` |
| 201 | + * ``build_count``: distinct ``(name_normalized, version)`` |
| 202 | +
|
| 203 | + Until rebuild suffixes exist, ``version_count`` equals ``build_count``. |
| 204 | + """ |
| 205 | + content_qs = content_qs.order_by() |
| 206 | + return { |
| 207 | + "package_count": content_qs.values("name_normalized").distinct().count(), |
| 208 | + "version_count": ( |
| 209 | + content_qs.annotate(_base_version=base_version_annotation()) |
| 210 | + .values("name_normalized", "_base_version") |
| 211 | + .distinct() |
| 212 | + .count() |
| 213 | + ), |
| 214 | + "build_count": content_qs.values("name_normalized", "version").distinct().count(), |
| 215 | + } |
0 commit comments