Inapt: SRPM dependencies report

For now, we just load src/pkglists and build a simple
report on what packages you have to have in order
to satisfy SRPM's build requires. Everything is
very approxymate, of course.

port_stats/inapt.py

@@ -0,0 +1,74 @@
+
+from __future__ import print_function
+
+import collections
+import logging
+
+import rpm
+
+import lists
+
+
+LOG = logging.getLogger(__name__)
+
+
+class Dependency(collections.namedtuple(
+        'Dependency', ['name', 'flags', 'version'])):
+
+    HEADER_TRIPLETS = {
+        'require': (rpm.RPMTAG_REQUIRENAME,
+                    rpm.RPMTAG_REQUIREFLAGS,
+                    rpm.RPMTAG_REQUIREVERSION),
+        'provide': (rpm.RPMTAG_PROVIDENAME,
+                    rpm.RPMTAG_PROVIDEFLAGS,
+                    rpm.RPMTAG_PROVIDEVERSION)
+    }
+
+    @classmethod
+    def from_header(cls, header, kind):
+        tags = cls.HEADER_TRIPLETS[kind]
+        triplets = zip(header[tags[0]], header[tags[1]], header[tags[2]])
+        for name, flags, version in triplets:
+            if not name.startswith('rpmlib('):
+                yield cls(name, flags, version or None)
+        if kind == 'provide':
+            for name in header[rpm.RPMTAG_FILENAMES]:
+                yield cls(name, 0, None)
+
+
+class DependencyInfo(collections.namedtuple(
+        'DependencyInfo', ['src_req', 'srcrpms',
+                           'bin_req', 'reverse_prov'])):
+
+    @classmethod
+    def load(cls, repo):
+        src_list, bin_list = lists.read_pkglist_heders_for_repo(
+            repo['path'], repo['arch'], components=['classic'])
+
+        def deps(*args):
+            return frozenset(Dependency.from_header(*args))
+
+        src_req = dict((h.name, deps(h, 'require')) for h in src_list)
+        srcrpms = dict((h.name, h[rpm.RPMTAG_SOURCERPM].rsplit('-', 2)[0])
+                       for h in bin_list)
+        bin_req = dict((h.name, deps(h, 'require')) for h in bin_list)
+        bin_prov = dict((h.name, deps(h, 'provide')) for h in bin_list)
+        reverse_prov = {}
+        for name, provs in bin_prov.iteritems():
+            for p in provs:
+                reverse_prov.setdefault(p.name, []).append((name, p))
+        return cls(src_req, srcrpms, bin_req, reverse_prov)
+
+    def simple_srpm_report(self, srpm):
+        result = collections.defaultdict(lambda: collections.defaultdict(set))
+        for dep in self.src_req[srpm]:
+            for p in self.reverse_prov[dep.name]:
+                bin_name = p[0]
+                src_name = self.srcrpms[bin_name]
+                result[src_name][p[0]].add(dep.name)
+
+        for srpm in sorted(result):
+            line = ['%-30s' % srpm]
+            for k, v in result[srpm].iteritems():
+                line.append('%s [%s]' % (k, ' '.join(sorted(v))))
+            print('  '.join(line))

port_stats/lists.py

@@ -80,6 +80,22 @@ def read_pkglist_headers_rpm(path):
         input_file.close()
 
 
+def read_pkglist_heders_for_repo(repo_path, arches, components=None):
+    bin_headers = []
+    src_headers = []
+    for arch in arches:
+        basedir = os.path.join(repo_path, arch, 'base')
+        for pkglist in os.listdir(basedir):
+            parts = pkglist.split('.', 3)
+            if parts[0] not in ('pkglist', 'srclist'):
+                continue
+            if components is not None and parts[1] not in components:
+                continue
+            (src_headers if parts[0] == 'srclist' else bin_headers).extend(
+                read_pkglist_headers_rpm(os.path.join(basedir, pkglist)))
+    return src_headers, bin_headers
+
+
 def _read_pkglist_rpm(path):
     return (NEVR.from_header(h) for h in read_pkglist_headers_rpm(path))