swh.lister.phabricator

Add a lister of all hosted repositories on a Phabricator instance
Closes T808

README.md

@@ -15,6 +15,7 @@ following Python modules:
 - `swh.lister.gitlab`
 - `swh.lister.pypi`
 - `swh.lister.npm`
+- `swh.lister.phabricator`
 
 Dependencies
 ------------
@@ -164,6 +165,18 @@ logging.basicConfig(level=logging.DEBUG)
 npm_lister()
 ```
 
+## lister-phabricator
+
+Once configured, you can execute a Phabricator lister using the following instructions in a `python3` script:
+
+```lang=python
+import logging
+from swh.lister.phabricator.tasks import incremental_phabricator_lister
+
+logging.basicConfig(level=logging.DEBUG)
+incremental_phabricator_lister(forge_url='https://forge.softwareheritage.org', api_token='XXXX')
+```
+
 Licensing
 ---------
 

swh/lister/cli.py

@@ -9,7 +9,8 @@ import click
 
 logger = logging.getLogger(__name__)
 
-SUPPORTED_LISTERS = ['github', 'gitlab', 'bitbucket', 'debian', 'pypi', 'npm']
+SUPPORTED_LISTERS = ['github', 'gitlab', 'bitbucket', 'debian', 'pypi',
+                     'npm', 'phabricator']
 
 
 @click.command()
@@ -96,6 +97,14 @@ def cli(db_url, listers, drop_tables):
                 NpmVisitModel.metadata.drop_all(_lister.db_engine)
             NpmVisitModel.metadata.create_all(_lister.db_engine)
 
+        elif lister == 'phabricator':
+            from .phabricator.models import IndexingModelBase as ModelBase
+            from .phabricator.lister import PhabricatorLister
+            _lister = PhabricatorLister(
+                forge_url='https://forge.softwareheritage.org',
+                api_token='',
+                override_config=override_conf)
+
         else:
             raise ValueError(
                 'Invalid lister %s: only supported listers are %s' %

swh/lister/core/tests/conftest.py

@@ -11,4 +11,5 @@ def celery_includes():
         'swh.lister.gitlab.tasks',
         'swh.lister.npm.tasks',
         'swh.lister.pypi.tasks',
+        'swh.lister.phabricator.tasks',
     ]

swh/lister/phabricator/lister.py

+# Copyright (C) 2019 the Software Heritage developers
+# License: GNU General Public License version 3, or any later version
+# See top-level LICENSE file for more information
+
+
+from swh.lister.core.indexing_lister import SWHIndexingHttpLister
+from swh.lister.phabricator.models import PhabricatorModel
+from collections import defaultdict
+
+
+class PhabricatorLister(SWHIndexingHttpLister):
+
+    PATH_TEMPLATE = '&order=oldest&attachments[uris]=1&after=%s'
+    MODEL = PhabricatorModel
+    LISTER_NAME = 'phabricator'
+
+    def __init__(self, forge_url, api_token, override_config=None):
+        if forge_url.endswith("/"):
+            forge_url = forge_url[:-1]
+        self.forge_url = forge_url
+        api_endpoint = ('api/diffusion.repository.'
+                        'search?api.token=%s') % api_token
+        api_baseurl = '%s/%s' % (forge_url, api_endpoint)
+        super().__init__(api_baseurl=api_baseurl,
+                         override_config=override_config)
+
+    def request_headers(self):
+        """
+        (Override) Set requests headers to send when querying the
+        Phabricator API
+        """
+        return {'User-Agent': 'Software Heritage phabricator lister',
+                'Accept': 'application/json'}
+
+    def get_model_from_repo(self, repo):
+        url = get_repo_url(repo['attachments']['uris']['uris'])
+        if url is None:
+            return None
+        return {
+            'uid': self.forge_url + str(repo['id']),
+            'indexable': repo['id'],
+            'name': repo['fields']['shortName'],
+            'full_name': repo['fields']['name'],
+            'html_url': url,
+            'origin_url': url,
+            'description': None,
+            'origin_type': repo['fields']['vcs']
+        }
+
+    def get_next_target_from_response(self, response):
+        body = response.json()['result']['cursor']
+        if body['after'] != 'null':
+            return body['after']
+        else:
+            return None
+
+    def transport_response_simplified(self, response):
+        repos = response.json()
+        if repos['result'] is None:
+            raise ValueError(
+                'Problem during information fetch: %s' % repos['error_code'])
+        repos = repos['result']['data']
+        return [self.get_model_from_repo(repo) for repo in repos]
+
+    def filter_before_inject(self, models_list):
+        """
+        (Overrides) SWHIndexingLister.filter_before_inject
+        Bounds query results by this Lister's set max_index.
+        """
+        models_list = [m for m in models_list if m is not None]
+        return super().filter_before_inject(models_list)
+
+    def _bootstrap_repositories_listing(self):
+        """
+        Method called when no min_bound value has been provided
+        to the lister. Its purpose is to:
+
+            1. get the first repository data hosted on the Phabricator
+               instance
+
+            2. inject them into the lister database
+
+            3. return the first repository index to start the listing
+               after that value
+
+        Returns:
+             int: The first repository index
+        """
+        params = '&order=oldest&limit=1'
+        response = self.safely_issue_request(params)
+        models_list = self.transport_response_simplified(response)
+        self.max_index = models_list[0]['indexable']
+        models_list = self.filter_before_inject(models_list)
+        injected = self.inject_repo_data_into_db(models_list)
+        self.create_missing_origins_and_tasks(models_list, injected)
+        return self.max_index
+
+    def run(self, min_bound=None, max_bound=None):
+        """
+        (Override) Run the lister on the specified Phabricator instance
+
+        Args:
+            min_bound (int): Optional repository index to start the listing
+                after it
+            max_bound (int): Optional repository index to stop the listing
+                after it
+        """
+        # initial call to the lister, we need to bootstrap it in that case
+        if min_bound is None:
+            min_bound = self._bootstrap_repositories_listing()
+        super().run(min_bound, max_bound)
+
+
+def get_repo_url(attachments):
+    """
+    Return url for a hosted repository from its uris attachments according
+    to the following priority lists:
+    * protocol: https > http
+    * identifier: shortname > callsign > id
+    """
+    processed_urls = defaultdict(dict)
+    for uri in attachments:
+        protocol = uri['fields']['builtin']['protocol']
+        url = uri['fields']['uri']['effective']
+        identifier = uri['fields']['builtin']['identifier']
+        if protocol in ('http', 'https'):
+            processed_urls[protocol][identifier] = url
+        elif protocol is None:
+            for protocol in ('https', 'http'):
+                if url.startswith(protocol):
+                    processed_urls[protocol]['undefined'] = url
+                break
+    for protocol in ['https', 'http']:
+        for identifier in ['shortname', 'callsign', 'id', 'undefined']:
+            if (protocol in processed_urls and
+                    identifier in processed_urls[protocol]):
+                return processed_urls[protocol][identifier]
+    return None

swh/lister/phabricator/models.py

+# Copyright (C) 2019 the Software Heritage developers
+# License: GNU General Public License version 3, or any later version
+# See top-level LICENSE file for more information
+
+from sqlalchemy import Column, String, Integer
+
+from swh.lister.core.models import IndexingModelBase
+
+
+class PhabricatorModel(IndexingModelBase):
+    """a Phabricator repository"""
+    __tablename__ = 'phabricator_repos'
+
+    uid = Column(String, primary_key=True)
+    indexable = Column(Integer, index=True)

swh/lister/phabricator/tasks.py

+# Copyright (C) 2019 the Software Heritage developers
+# License: GNU General Public License version 3, or any later version
+# See top-level LICENSE file for more information
+
+from swh.scheduler.celery_backend.config import app
+from swh.lister.phabricator.lister import PhabricatorLister
+
+
+def new_lister(
+        forge_url='https://forge.softwareheritage.org', api_token='', **kw):
+    return PhabricatorLister(forge_url=forge_url, api_token=api_token, **kw)
+
+
+@app.task(name=__name__ + '.IncrementalPhabricatorLister')
+def incremental_phabricator_lister(**lister_args):
+    lister = new_lister(**lister_args)
+    lister.run(min_bound=lister.db_last_index())
+
+
+@app.task(name=__name__ + '.FullPhabricatorLister')
+def full_phabricator_lister(**lister_args):
+    lister = new_lister(**lister_args)
+    lister.run()
+
+
+@app.task(name=__name__ + '.ping')
+def ping():
+    return 'OK'

swh/lister/phabricator/tests/api_empty_response.json

+{
+  "result": {
+    "data": [],
+    "cursor": {
+      "after": null
+    }
+  }
+}

swh/lister/phabricator/tests/api_response_undefined_protocol.json

+{
+  "id": 8,
+  "type": "REPO",
+  "phid": "PHID-REPO-ge2icigfu5ijk2whqfbl",
+  "fields": {
+    "name": "Blender Libraries",
+    "vcs": "svn",
+    "callsign": "BL",
+    "shortName": null,
+    "status": "active",
+    "isImporting": false,
+    "almanacServicePHID": null,
+    "spacePHID": null,
+    "dateCreated": 1385564674,
+    "dateModified": 1468574079,
+    "policy": {
+      "view": "public",
+      "edit": "admin",
+      "diffusion.push": "PHID-PROJ-hclk7tvd6pmpjmqastjl"
+    }
+  },
+  "attachments": {
+    "uris": {
+      "uris": [
+        {
+          "id": "70",
+          "type": "RURI",
+          "phid": "PHID-RURI-h7zdbkud6why4xrb2s2e",
+          "fields": {
+            "repositoryPHID": "PHID-REPO-ge2icigfu5ijk2whqfbl",
+            "uri": {
+              "raw": "https://svn.blender.org/svnroot/bf-blender/",
+              "display": "https://svn.blender.org/svnroot/bf-blender/",
+              "effective": "https://svn.blender.org/svnroot/bf-blender/",
+              "normalized": "svn.blender.org/svnroot/bf-blender"
+            },
+            "io": {
+              "raw": "observe",
+              "default": "none",
+              "effective": "observe"
+            },
+            "display": {
+              "raw": "always",
+              "default": "never",
+              "effective": "always"
+            },
+            "credentialPHID": null,
+            "disabled": false,
+            "builtin": {
+              "protocol": null,
+              "identifier": null
+            },
+            "dateCreated": "1467894515",
+            "dateModified": "1468574079"
+          }
+        }
+      ]
+    }
+  }
+}
\ No newline at end of file

swh/lister/phabricator/tests/conftest.py

+from swh.lister.core.tests.conftest import *  # noqa

swh/lister/phabricator/tests/test_lister.py

+# Copyright (C) 2019 the Software Heritage developers
+# License: GNU General Public License version 3, or any later version
+# See top-level LICENSE file for more information
+
+import re
+import json
+import unittest
+from swh.lister.core.tests.test_lister import HttpListerTester
+from swh.lister.phabricator.lister import PhabricatorLister
+from swh.lister.phabricator.lister import get_repo_url
+
+
+class PhabricatorListerTester(HttpListerTester, unittest.TestCase):
+    Lister = PhabricatorLister
+    test_re = re.compile(r'\&after=([^?&]+)')
+    lister_subdir = 'phabricator'
+    good_api_response_file = 'api_response.json'
+    good_api_response_undefined_protocol = 'api_response_undefined_'\
+                                           'protocol.json'
+    bad_api_response_file = 'api_empty_response.json'
+    first_index = 1
+    last_index = 12
+    entries_per_page = 10
+
+    def get_fl(self, override_config=None):
+        """(Override) Retrieve an instance of fake lister (fl).
+        """
+        if override_config or self.fl is None:
+            self.fl = self.Lister(forge_url='https://fakeurl', api_token='a-1',
+                                  override_config=override_config)
+            self.fl.INITIAL_BACKOFF = 1
+
+        self.fl.reset_backoff()
+        return self.fl
+
+    def test_get_repo_url(self):
+        f = open('swh/lister/%s/tests/%s' % (self.lister_subdir,
+                                             self.good_api_response_file))
+        api_response = json.load(f)
+        repos = api_response['result']['data']
+        for repo in repos:
+            self.assertEqual(
+                'https://forge.softwareheritage.org/source/%s.git' %
+                (repo['fields']['shortName']),
+                get_repo_url(repo['attachments']['uris']['uris']))
+
+        f = open('swh/lister/%s/tests/%s' %
+                 (self.lister_subdir,
+                  self.good_api_response_undefined_protocol))
+        repo = json.load(f)
+        self.assertEqual(
+                'https://svn.blender.org/svnroot/bf-blender/',
+                get_repo_url(repo['attachments']['uris']['uris']))

swh/lister/phabricator/tests/test_tasks.py

+from unittest.mock import patch
+
+
+def test_ping(swh_app, celery_session_worker):
+    res = swh_app.send_task(
+        'swh.lister.phabricator.tasks.ping')
+    assert res
+    res.wait()
+    assert res.successful()
+    assert res.result == 'OK'
+
+
+@patch('swh.lister.phabricator.tasks.PhabricatorLister')
+def test_incremental(lister, swh_app, celery_session_worker):
+    # setup the mocked PhabricatorLister
+    lister.return_value = lister
+    lister.db_last_index.return_value = 42
+    lister.run.return_value = None
+
+    res = swh_app.send_task(
+        'swh.lister.phabricator.tasks.IncrementalPhabricatorLister')
+    assert res
+    res.wait()
+    assert res.successful()
+
+    lister.assert_called_once_with(
+        api_token='', forge_url='https://forge.softwareheritage.org')
+    lister.db_last_index.assert_called_once_with()
+    lister.run.assert_called_once_with(min_bound=42)