SQLAlchemy/PostgreSQL instead of MongoDB

1 files changed, 50 insertions, 49 deletions

diff --git a/utils/db_init.py b/utils/db_init.py
index 98bbe0d..725ea74 100755
--- a/utils/db_init.py
+++ b/utils/db_init.py
@@ -1,17 +1,28 @@
 #!/usr/bin/env python
-import sys
-
-from datetime import datetime
+import os, sys
 
 from pkgcore.config import load_config
 from pkgcore.cache import metadata
 from pkgcore.ebuild import repository
 
-from pymongo import Connection
+from sqlalchemy import create_engine
+from sqlalchemy.orm import scoped_session, sessionmaker
+from sqlalchemy.orm.exc import NoResultFound
+
+path = os.path.join(os.path.dirname(__file__), os.path.pardir)
+sys.path.insert(0, path)
+
+from grumpy.models import Base, Developer, Ebuild, Herd, Package
 
 def main(path):
-    # Set up a link to our MongoDB collection
-    db = Connection('localhost', 27017)['grumpy']
+    engine = create_engine('postgresql://grumpy:grumpy@localhost/grumpy')
+    session = scoped_session(sessionmaker(autocommit=False,
+                                          autoflush=False,
+                                          bind=engine))
+    Base.query = session.query_property()
+
+    Base.metadata.drop_all(bind=engine)
+    Base.metadata.create_all(bind=engine)
 
     # pkgcore part to fetch all the ebuild information
     conf = load_config()
@@ -21,30 +32,38 @@ def main(path):
                                                eclass_cache=eclass_cache)
 
     for pkg in overlay_repo:
-        ver_id = "%s/%s-%s" % (pkg.category, pkg.package, pkg.version)
-        pkg_id = "%s/%s" % (pkg.category, pkg.package)
+        pkg_id = "%s/%s" % (pkg.package, pkg.category)
 
-        # Load package info from packages collection
-        package = db.ebuilds.find_one(dict(_id=pkg_id))
+        package = Package.query.filter_by(cp=pkg_id).first()
         if not package:
-            # ...or create data if pkg not in db
-            package = dict(
-                _id = pkg_id,
-                _ver = 0,                   # Internal schema version
-                cat = pkg.category,
-                pkg = pkg.package,
-                desc = pkg.description,
-                ldesc = pkg.longdescription,
-                homepage = pkg.homepage,
-                herds = list(pkg.herds) if pkg.herds else [],
-                maintainers = [m.email for m in pkg.maintainers] \
-                                        if pkg.maintainers else [],
-            )
-        else:
-            # TODO: Should we update it here?
-            pass
-        # Housekeeping
-        package['synced_on'] = datetime.now()
+
+            package = Package(pkg.category, pkg.package, pkg.description, \
+                              pkg.longdescription, pkg.homepage)
+            # Fetch devs and herds from db
+            devs = []
+            for dev in pkg.maintainers:
+                try:
+                    devs.append(Developer.query.filter_by(email=dev.email).one())
+                except NoResultFound:
+                    devs.append(Developer(dev.email))
+            session.add_all(devs)
+            herds = []
+            for herd in pkg.herds:
+                # Workaround for empty <herd></herd> tags
+                if herd is None:
+                    herd = 'fix-me'
+                herd = herd.strip()
+                try:
+                    herds.append(Herd.query.filter_by(name=herd).one())
+                except NoResultFound:
+                    herds.append(Herd(herd))
+            session.add_all(herds)
+
+            package.devs = devs
+            package.herds = herds
+
+            session.add(package)
+            session.commit()
 
         # Parse IUSE (iuse contains all flags, fiuse contains only '+use')
         iuse = list()
@@ -56,27 +75,9 @@ def main(path):
             else:
                 iuse.append(u)
 
-        # We don't care about existing ebuilds and just overwrite them
-        ebuild = dict(
-            _id = ver_id,
-            _ver = 0,
-            eapi = pkg.eapi,
-            cat = pkg.category,
-            pkg = pkg.package,
-            iuse = iuse,
-            fiuse = fiuse,
-            keywords = list(pkg.keywords) if pkg.keywords else [],
-            slot = pkg.slot,
-            version = pkg.version,
-            # TODO, need to figure out a proper queryable structure for these
-#           license = pkg.license,
-#           depends = pkg.depends,
-#           rdepends = pkg.rdepends
-
-            # Housekeeping
-            synced_on = datetime.now()
-        )
-        print "Adding: ", db.packages.save(package), db.ebuilds.save(ebuild)
+        session.add(Ebuild(package, pkg.version, pkg.revision, pkg.eapi, \
+                           pkg.slot, pkg.keywords, iuse, fiuse))
+        session.commit()
 
 if __name__ == '__main__':
     if len(sys.argv) != 2: