Updated the update_docs management command to also reindex.

django_website/docs/management/commands/update_docs.py

@@ -4,32 +4,43 @@
 """
 from __future__ import absolute_import
 
+import json
+import haystack
 import subprocess
 import sphinx.cmdline
 from django.conf import settings
 from django.core.management.base import NoArgsCommand
+from django.utils.html import strip_tags
 from unipath import FSPath as Path
-from ...models import DocumentRelease
+from ...models import DocumentRelease, Document
 
 class Command(NoArgsCommand):
     def handle_noargs(self, **kwargs):
         try:
-            verbose = int(kwargs['verbosity']) > 0
+            verbosity = int(kwargs['verbosity'])
         except (KeyError, TypeError, ValueError):
-            verbose = True
+            verbosity = 1
 
         for release in DocumentRelease.objects.all():
-            if verbose:
+            if verbosity >= 1:
                 print "Updating %s..." % release
-                
+
             destdir = Path(settings.DOCS_BUILD_ROOT).child(release.lang, release.version)
             if not destdir.exists():
                 destdir.mkdir(parents=True)
+
+            #
+            # Update the release from SCM.
+            #
 
             # Make an SCM checkout/update into the destination directory.
             # Do this dynamically in case we add other SCM later.
             getattr(self, 'update_%s' % release.scm)(release.scm_url, destdir)
 
+            #
+            # Use Sphinx to build the release docs into JSON documents.
+            #
+
             # Make the directory for the JSON files - sphinx-build doesn't
             # do it for us, apparently.
             json_build_dir = destdir.child('_build', 'json')
@@ -44,5 +55,61 @@ def handle_noargs(self, **kwargs):
                 json_build_dir,    # Destination directory
             ])
 
+            #
+            # Rebuild the imported document list and search index.
+            #
+
+            # Build a dict of {path_fragment: document_object}. We'll pop values
+            # out of this dict as we go which'll make sure we know which
+            # remaining documents need to be deleted (and unindexed) later on.
+            documents = dict((doc.path, doc) for doc in release.documents.all())
+
+            # Walk the tree we've just built looking for ".fjson" documents
+            # (just JSON, but Sphinx names them weirdly). Each one of those 
+            # documents gets a corresponding Document object created which
+            # we'll then ask Sphinx to reindex. 
+            #
+            # We have to be a bit careful to reverse-engineer the correct
+            # relative path component, especially for "index" documents,
+            # otherwise the search results will be incorrect. 
+            for built_doc in json_build_dir.walk():
+                if built_doc.isfile() and built_doc.ext == '.fjson':
+
+                    # Convert the built_doc path which is now an absolute
+                    # path (i.e. "/home/docs/en/1.2/_build/ref/models.json")
+                    # into a path component (i.e. "ref/models").
+                    path = json_build_dir.rel_path_to(built_doc) 
+                    if path.stem == 'index':
+                        path = path.parent
+                    path = str(path.parent.child(path.stem))
+
+                    # Read out the content and create a new Document object for
+                    # it. We'll strip the HTML tags here (for want of a better
+                    # place to do it).
+                    with open(built_doc) as fp:
+                        json_doc = json.load(fp)
+                        try:
+                            body = strip_tags(json_doc['body'])
+                            title = strip_tags(json_doc['title'])
+                        except KeyError, ex:
+                            if verbosity >= 2:
+                                print "Skipping: %s (no %s)" % (path, ex.args[0])
+                            continue
+
+                    doc = documents.pop(path, Document(path=path, release=release))
+                    doc.title = title
+                    doc.body = body
+                    if verbosity >= 2:
+                        print "Indexing:", doc
+                    doc.save()
+                    haystack.site.update_object(doc)
+
+            # Clean up any remaining documents.
+            for doc in documents.values():
+                if verbosity >= 2:
+                    print "Deleting:", doc
+                haystack.site.remove_object(doc)
+                doc.delete()
+
     def update_svn(self, url, destdir):
         subprocess.call(['svn', 'checkout', '-q', url, destdir])