module to read package info

pkgreader.py

@@ -0,0 +1,44 @@
+import tarfile
+import logging
+
+logger = logging.getLogger(__name__)
+
+multikeys = {'depend', 'makepkgopt', 'optdepend', 'replaces', 'conflict',
+             'provides', 'license', 'backup', 'group'}
+
+def _add_to_dict(d, key, value):
+  if key in multikeys:
+    if key in d:
+      d[key].append(value)
+    else:
+      d[key] = [value]
+  else:
+    assert key not in d, 'unexpected multi-value key "%s"' % key
+    d[key] = value
+
+def readpkg(file):
+  tar = tarfile.open(file)
+  info = tar.next()
+  if not info or info.name != '.PKGINFO':
+    logger.warn('%s is not a nice package!', file)
+    info = '.PKGINFO' # have to look further
+  f = tar.extractfile(info)
+  data = f.read().decode()
+  tar.close()
+
+  d = {}
+  key = None
+  for l in data.split('\n'):
+    if l.startswith('#'):
+      continue
+    if not l:
+      continue
+    if '=' not in l:
+      value += l
+    else:
+      if key is not None:
+        _add_to_dict(d, key, value)
+      key, value = l.split(' = ', 1)
+  _add_to_dict(d, key, value)
+
+  return d

test_readpkg.py

@@ -0,0 +1,11 @@
+#!/usr/bin/env python3
+# vim:fileencoding=utf-8
+
+from subprocess import getoutput
+allpkgs = getoutput(r"locate -be --regex '\.pkg\.tar\.xz$'").split('\n')
+
+from pkgreader import readpkg
+for p in allpkgs:
+  print('reading package:', p)
+  d = readpkg(p)
+  print('desc:', d.get('pkgdesc', '(nothing)'))