browse: Support resolution of SWH persistent identifiers

Closes T926

docs/uri-scheme-browse-identifiers.rst

+SWH identifiers
+^^^^^^^^^^^^^^^
+
+A subset of Software Heritage objects (contents, directories, releases and revisions)
+can be browsed using :ref:`persistent-identifiers`. Those identifiers
+are guaranteed to remain stable (persistent) over time.
+
+.. http:get:: /browse/(swh_id)/
+
+    End point to browse SWH objects using their persistent identifiers.
+    A redirection to the adequate HTML view will be performed when
+    reaching it.
+
+    :param string swh_id: a persistent identifier for a SWH object
+        (see the :ref:`persistent-identifiers` section to learn more
+        about its syntax)
+    :resheader Location: the redirection URL for browsing the SWH object
+        associated to the provided identifier
+    :statuscode 302: no error
+    :statuscode 400: the provided identifier is malformed
+
+    **Examples:**
+
+    .. parsed-literal::
+
+        :swh_web_browse:`swh:1:cnt:0ffd12d85cdec70c88e852fc3f5ea9fd342213cd`
+        :swh_web_browse:`swh:1:dir:db990da9af15427455ce7836ce2b8a34b9bf67f5`
+        :swh_web_browse:`swh:1:rel:a9b7e3f1eada90250a6b2ab2ef3e0a846cb16831`
+        :swh_web_browse:`swh:1:rev:f1b94134a4b879bc55c3dacdb496690c8ebdc03f`
\ No newline at end of file

docs/uri-scheme-browse.rst

@@ -91,4 +91,6 @@ SWH Browse Urls
 
 .. include:: uri-scheme-browse-revision.rst
 
-.. include:: uri-scheme-browse-release.rst
\ No newline at end of file
+.. include:: uri-scheme-browse-release.rst
+
+.. include:: uri-scheme-browse-identifiers.rst

swh/web/browse/urls.py

@@ -8,6 +8,7 @@ from django.shortcuts import render
 
 import swh.web.browse.views.directory # noqa
 import swh.web.browse.views.content # noqa
+import swh.web.browse.views.identifiers # noqa
 import swh.web.browse.views.origin # noqa
 import swh.web.browse.views.person # noqa
 import swh.web.browse.views.release # noqa

swh/web/browse/views/identifiers.py

+# Copyright (C) 2017-2018  The Software Heritage developers
+# See the AUTHORS file at the top-level directory of this distribution
+# License: GNU General Public License version 3, or any later version
+# See top-level LICENSE file for more information
+
+from django.shortcuts import redirect
+
+from swh.model.identifiers import parse_persistent_identifier
+
+from swh.web.browse.browseurls import browse_route
+from swh.web.common.utils import reverse
+from swh.web.common.exc import BadInputExc, handle_view_exception
+
+
+@browse_route(r'(?P<swh_id>swh:[0-9]+:[a-z]+:[0-9a-f]+)/',
+              view_name='browse-swh-id')
+def swh_id_browse(request, swh_id):
+    """
+    Django view enabling to browse the SWH archive using
+    :ref:`persistent-identifiers`.
+
+    The url that points to it is :http:get:`/browse/(swh_id)/`.
+
+    Args:
+        swh_id: A SWH persistent identifier
+
+    Returns:
+        A redirection to the adequate view to browse the
+        requested SWH object
+    """
+    try:
+        swh_id_parsed = parse_persistent_identifier(swh_id)
+        object_type = swh_id_parsed['object_type']
+        object_id = swh_id_parsed['object_id']
+        view_url = None
+        if object_type == 'cnt':
+            query_string = 'sha1_git:' + object_id
+            view_url = reverse('browse-content',
+                               kwargs={'query_string': query_string},
+                               query_params=request.GET)
+        elif object_type == 'dir':
+            view_url = reverse('browse-directory',
+                               kwargs={'sha1_git': object_id},
+                               query_params=request.GET)
+        elif object_type == 'rel':
+            view_url = reverse('browse-release',
+                               kwargs={'sha1_git': object_id},
+                               query_params=request.GET)
+        elif object_type == 'rev':
+            view_url = reverse('browse-revision',
+                               kwargs={'sha1_git': object_id},
+                               query_params=request.GET)
+        else:
+            msg = '\'%s\' is not a valid SWH persistent identifier!' % swh_id
+            raise BadInputExc(msg)
+    except Exception as exc:
+        return handle_view_exception(request, exc)
+
+    return redirect(view_url)

swh/web/tests/browse/views/test_identifiers.py

+# Copyright (C) 2018  The Software Heritage developers
+# See the AUTHORS file at the top-level directory of this distribution
+# License: GNU General Public License version 3, or any later version
+# See top-level LICENSE file for more information
+
+# flake8: noqa
+
+from unittest.mock import patch
+from nose.tools import istest
+from django.test import TestCase
+
+from swh.web.common.exc import BadInputExc
+from swh.web.common.utils import reverse
+from swh.web.tests.testbase import SWHWebTestBase
+
+from .data.content_test_data import stub_content_text_data
+
+from .data.directory_test_data import stub_root_directory_sha1
+
+from .data.revision_test_data import revision_id_test
+
+from .data.release_test_data import stub_release
+
+swh_id_prefix = 'swh:1:'
+
+
+class SwhBrowseIdTest(SWHWebTestBase, TestCase):
+
+    @istest
+    def content_id_browse(self):
+        cnt_sha1_git = stub_content_text_data['checksums']['sha1_git']
+        swh_id = swh_id_prefix + 'cnt:' + cnt_sha1_git
+        url = reverse('browse-swh-id',
+                      kwargs={'swh_id': swh_id})
+
+        query_string = 'sha1_git:' + cnt_sha1_git
+        content_browse_url = reverse('browse-content',
+                                     kwargs={'query_string': query_string})
+
+        resp = self.client.get(url)
+
+        self.assertEquals(resp.status_code, 302)
+        self.assertEqual(resp['location'], content_browse_url)
+
+    @istest
+    def directory_id_browse(self):
+        swh_id = swh_id_prefix + 'dir:' + stub_root_directory_sha1
+        url = reverse('browse-swh-id',
+                      kwargs={'swh_id': swh_id})
+
+        directory_browse_url = reverse('browse-directory',
+                                       kwargs={'sha1_git': stub_root_directory_sha1})
+
+        resp = self.client.get(url)
+
+        self.assertEquals(resp.status_code, 302)
+        self.assertEqual(resp['location'], directory_browse_url)
+
+    @istest
+    def revision_id_browse(self):
+        swh_id = swh_id_prefix + 'rev:' + revision_id_test
+        url = reverse('browse-swh-id',
+                      kwargs={'swh_id': swh_id})
+
+        revision_browse_url = reverse('browse-revision',
+                                       kwargs={'sha1_git': revision_id_test})
+
+        resp = self.client.get(url)
+
+        self.assertEquals(resp.status_code, 302)
+        self.assertEqual(resp['location'], revision_browse_url)
+
+        query_params = {'origin_type': 'git',
+                        'origin_url': 'https://github.com/webpack/webpack'}
+
+        url = reverse('browse-swh-id',
+                      kwargs={'swh_id': swh_id},
+                      query_params=query_params)
+
+        revision_browse_url = reverse('browse-revision',
+                                       kwargs={'sha1_git': revision_id_test},
+                                       query_params=query_params)
+
+        resp = self.client.get(url)
+        self.assertEquals(resp.status_code, 302)
+        self.assertEqual(resp['location'], revision_browse_url)
+
+    @istest
+    def release_id_browse(self):
+        swh_id = swh_id_prefix + 'rel:' + stub_release['id']
+        url = reverse('browse-swh-id',
+                      kwargs={'swh_id': swh_id})
+
+        release_browse_url = reverse('browse-release',
+                                     kwargs={'sha1_git': stub_release['id']})
+
+        resp = self.client.get(url)
+
+        self.assertEquals(resp.status_code, 302)
+        self.assertEqual(resp['location'], release_browse_url)
+
+        query_params = {'origin_type': 'git',
+                        'origin_url': 'https://github.com/python/cpython'}
+
+        url = reverse('browse-swh-id',
+                      kwargs={'swh_id': swh_id},
+                      query_params=query_params)
+
+        release_browse_url = reverse('browse-release',
+                                     kwargs={'sha1_git': stub_release['id']},
+                                     query_params=query_params)
+
+        resp = self.client.get(url)
+        self.assertEquals(resp.status_code, 302)
+        self.assertEqual(resp['location'], release_browse_url)
+
+    @istest
+    def bad_id_browse(self):
+        swh_id = swh_id_prefix + 'foo:' + stub_release['id']
+        url = reverse('browse-swh-id',
+                      kwargs={'swh_id': swh_id})
+
+        resp = self.client.get(url)
+        self.assertEquals(resp.status_code, 400)
+

[Phabricator Migration user]

Some references in the commit message have been migrated:

• T926 is now #926 (closed)