mark as deprecated

Enable `python setup.py test`

.gitignore

@@ -1,3 +1,5 @@
 celerybeat-schedule
 oyster_settings.py
+oyster.egg-info/
 *.pyc
+.tox

.travis.yml

@@ -2,8 +2,9 @@ language: python
 python:
     - "2.6"
     - "2.7"
-install: pip install scrapelib pymongo nose  --use-mirrors
+install: pip install scrapelib pymongo nose celery  --use-mirrors
 script: nosetests
+services: mongodb
 notifications:
     email:
         - jturk@sunlightfoundation.com

README.rst

@@ -1,3 +1,5 @@
+**DEPRECATED** - this project is abandoned & will not be seeing future updates
+
 ======
 oyster
 ======

changelog.rst

@@ -1,6 +1,14 @@
 oyster changelog
 ================
 
+0.4.0-dev
+---------
+    * S3 storage backend bugfix
+    * lots of improvements to signal script
+    * oyster.ext cloudsearch, elasticsearch, and superfastmatch
+    * use python logging w/ mongo handler
+    * add tox/python setup.py test (thanks Marc Abramowitz!)
+
 0.3.2
 -----
 **2012-03-29**

oyster/__init__.py

@@ -1,4 +1,4 @@
-__version__ = "0.3.2"
+__version__ = "0.4.0-dev"
 
 import os
 os.environ['CELERY_CONFIG_MODULE'] = 'oyster.celeryconfig'

oyster/celeryconfig.py

@@ -1,13 +1,3 @@
 from oyster.conf import settings
 
-CELERY_IMPORTS = ["oyster.tasks"] + list(settings.CELERY_TASK_MODULES)
+CELERY_IMPORTS = ['oyster.tasks'] + list(settings.CELERY_TASK_MODULES)
-
-BROKER_TRANSPORT = 'mongodb'
-BROKER_HOST = settings.MONGO_HOST
-BROKER_PORT = settings.MONGO_PORT
-
-CELERY_RESULT_BACKEND = 'mongodb'
-CELERY_MONGODB_BACKEND_SETTINGS = {
-    'host': settings.MONGO_HOST,
-    'port': settings.MONGO_PORT,
-}

oyster/core.py

@@ -1,4 +1,5 @@
 import datetime
+import logging
 import hashlib
 import random
 import sys
@@ -6,7 +7,9 @@ import sys
 import pymongo
 import scrapelib
 
+from .mongolog import MongoHandler
 from .storage import engines
+from celery.execute import send_task
 
 
 class Kernel(object):
@@ -22,21 +25,22 @@ class Kernel(object):
         configurable for ease of testing, only one should be instantiated
         """
 
-        # set up a capped log if it doesn't exist
+        # set up the log
         self.db = pymongo.Connection(mongo_host, mongo_port)[mongo_db]
-        try:
+
-            self.db.create_collection('logs', capped=True,
+        self.log = logging.getLogger('oyster')
-                                      size=mongo_log_maxsize)
+        self.log.setLevel(logging.DEBUG)
-        except pymongo.errors.CollectionInvalid:
+        self.log.addHandler(MongoHandler(mongo_db, host=mongo_host,
-            # cap collection if not capped?
+                                         port=mongo_port,
-            pass
+                                         capped_size=mongo_log_maxsize))
 
         # create status document if it doesn't exist
         if self.db.status.count() == 0:
             self.db.status.insert({'update_queue': 0})
 
         # ensure an index on _random
-        self.db.tracked.ensure_index([('_random', pymongo.ASCENDING)])
+        self.db.tracked.ensure_index('_random')
+        self.db.tracked.ensure_index('url')
 
         self.scraper = scrapelib.Scraper(user_agent=user_agent,
                                          requests_per_minute=rpm,
@@ -70,14 +74,6 @@ class Kernel(object):
         self.db.drop_collection('logs')
         self.db.drop_collection('status')
 
-    def log(self, action, url, error=False, **kwargs):
-        """ add an entry to the oyster log """
-        kwargs['action'] = action
-        kwargs['url'] = url
-        kwargs['error'] = error
-        kwargs['timestamp'] = datetime.datetime.utcnow()
-        self.db.logs.insert(kwargs)
-
     def _add_doc_class(self, doc_class, **properties):
         self.doc_classes[doc_class] = properties
 
@@ -93,17 +89,16 @@ class Kernel(object):
             any keyword args will be added to the document's metadata
         """
         if doc_class not in self.doc_classes:
-            error = 'unregistered doc_class %s' % doc_class
+            error = 'error tracking %s, unregistered doc_class %s'
-            self.log('track', url=url, error=error)
+            self.log.error(error, url, doc_class)
-            raise ValueError(error)
+            raise ValueError(error % (url, doc_class))
 
         # try and find an existing version of this document
         tracked = None
 
         if id:
             tracked = self.db.tracked.find_one({'_id': id})
-
+        else:
-        if not tracked:
             tracked = self.db.tracked.find_one({'url': url})
 
         # if id exists, ensure that URL and doc_class haven't changed
@@ -117,15 +112,14 @@ class Kernel(object):
                 return tracked['_id']
             else:
                 # id existed but with different URL
-                error = ('%s already exists with different data (tracked: '
+                message = ('%s already exists with different data (tracked: '
-                         '%s, %s) (new: %s, %s)' % (tracked['_id'],
+                           '%s, %s) (new: %s, %s)')
-                                                    tracked['url'],
+                args = (tracked['_id'], tracked['url'], tracked['doc_class'],
-                                                    tracked['doc_class'],
+                        url, doc_class)
-                                                    url, doc_class))
+                self.log.error(message, *args)
-                self.log('track', url=url, error=error)
+                raise ValueError(message % args)
-                raise ValueError(error)
 
-        self.log('track', url=url)
+        self.log.info('tracked %s [%s]', url, id)
 
         newdoc = dict(url=url, doc_class=doc_class,
                       _random=random.randint(0, sys.maxint),
@@ -187,8 +181,8 @@ class Kernel(object):
                                     'storage_type': storage.storage_type,
                                    })
             # fire off onchanged functions
-            for onchanged in doc_class['onchanged']:
+            for onchanged in doc_class.get('onchanged', []):
-                onchanged(doc, newdata)
+                send_task(onchanged, (doc['_id'],))
 
         if error:
             # if there's been an error, increment the consecutive_errors count
@@ -208,7 +202,11 @@ class Kernel(object):
         else:
             doc['next_update'] = None
 
-        self.log('update', url=url, new_doc=new_version, error=error)
+        if error:
+            self.log.warning('error updating %s [%s]', url, doc['_id'])
+        else:
+            new_version = ' (new)'
+            self.log.info('updated %s [%s]%s', url, doc['_id'], new_version)
 
         self.db.tracked.save(doc, safe=True)
 
@@ -257,6 +255,16 @@ class Kernel(object):
         storage = self.storage[doc_class['storage_engine']]
         return storage.get(doc['versions'][-1]['storage_key'])
 
+    def extract_text(self, doc):
+        version = self.get_last_version(doc)
+        doc_class = self.doc_classes[doc['doc_class']]
+        try:
+            extract_text = doc_class['extract_text']
+        except KeyError:
+            raise ValueError('doc_class %s missing extract_text' %
+                             doc['doc_class'])
+        return extract_text(doc, version)
+
 
 def _get_configured_kernel():
     """ factory, gets a connection configured with oyster.conf.settings """

oyster/ext/cloudsearch.py

@@ -0,0 +1,32 @@
+# needed so we can import cloudsearch
+from __future__ import absolute_import
+
+from celery.task.base import Task
+
+from ..core import kernel
+from ..conf import settings
+
+from cloudsearch import CloudSearch
+
+cs = CloudSearch(settings.CLOUDSEARCH_DOMAIN, settings.CLOUDSEARCH_ID, 20)
+
+
+class CloudSearchPush(Task):
+    """ task that updates documents """
+    # results go straight to database
+    ignore_result = True
+
+    # a bit under 1MB
+    MAX_BYTES = 1048000
+
+    def run(self, doc_id):
+        doc = kernel.db.tracked.find_one({'_id': doc_id})
+        text = kernel.extract_text(doc)
+        pieces = [text[i:i+self.MAX_BYTES] for i in
+                  xrange(0, len(text), self.MAX_BYTES)]
+
+        self.get_logger().debug('adding {0} pieces for {1}'.format(
+            len(pieces), doc_id))
+        for i, piece in enumerate(pieces):
+            cloud_id = '%s_%s' % (doc_id.lower(), i)
+            cs.add_document(cloud_id, text=piece, **doc['metadata'])

oyster/ext/elasticsearch.py

@@ -0,0 +1,36 @@
+import logging
+from celery.task.base import Task
+
+from ..core import kernel
+from ..conf import settings
+
+from pyes import ES
+
+es = ES(settings.ELASTICSEARCH_HOST)
+log = logging.getLogger('oyster.ext.elasticsearch')
+
+class ElasticSearchPush(Task):
+    # results go straight to elasticsearch
+    ignore_result = True
+
+    def run(self, doc_id):
+        doc = kernel.db.tracked.find_one({'_id': doc_id})
+
+        try:
+            text = kernel.extract_text(doc)
+            if not text:
+                log.info('no text for %s', doc_id,
+                         extra={'doc_class':doc['doc_class']})
+                return
+
+            log.info('tracked %s', doc_id,
+                     extra={'doc_class':doc['doc_class']})
+
+            es.index(dict(doc['metadata'], text=text),
+                     settings.ELASTICSEARCH_INDEX,
+                     settings.ELASTICSEARCH_DOC_TYPE,
+                     id=doc_id)
+        except Exception as e:
+            log.warning('error tracking %s', doc_id,
+                        extra={'doc_class':doc['doc_class']}, exc_info=True)
+            raise

oyster/ext/superfastmatch.py

@@ -0,0 +1,23 @@
+# needed so we can import superfastmatch.Client
+from __future__ import absolute_import
+from celery.task.base import Task
+
+from ..core import kernel
+from ..conf import settings
+
+from superfastmatch import Client
+
+sfm = Client(settings.SUPERFASTMATCH_URL)
+
+
+class SuperFastMatchPush(Task):
+    """ task that pushes documents to SFM """
+
+    # results go straight to database
+    ignore_result = True
+
+    def run(self, doc_id):
+        doc = kernel.db.tracked.find_one({'_id': doc_id})
+        text = kernel.extract_text(doc)
+        doctype, docid = settings.SUPERFASTMATCH_ID_FUNC(doc_id)
+        sfm.add(doctype, docid, text, defer=True, **doc['metadata'])

oyster/mongolog.py

@@ -0,0 +1,54 @@
+"""
+    MongoDB handler for Python Logging
+
+    inspired by https://github.com/andreisavu/mongodb-log
+"""
+
+import logging
+import datetime
+import socket
+import pymongo
+
+
+class MongoFormatter(logging.Formatter):
+
+    def format(self, record):
+        """ turn a LogRecord into something mongo can store """
+        data = record.__dict__.copy()
+
+        data.update(
+            # format message
+            message=record.getMessage(),
+            # overwrite created (float) w/ a mongo-compatible datetime
+            created=datetime.datetime.utcnow(),
+            host=socket.gethostname(),
+            args=tuple(unicode(arg) for arg in record.args)
+        )
+        data.pop('msecs')   # not needed, stored in created
+
+        # TODO: ensure everything in 'extra' is MongoDB-ready
+        exc_info = data.get('exc_info')
+        if exc_info:
+            data['exc_info'] = self.formatException(exc_info)
+        return data
+
+
+class MongoHandler(logging.Handler):
+    def __init__(self, db, collection='logs', host='localhost', port=None,
+                 capped_size=100000000, level=logging.NOTSET, async=True):
+        db = pymongo.connection.Connection(host, port)[db]
+        # try and create the capped log collection
+        if capped_size:
+            try:
+                db.create_collection(collection, capped=True, size=capped_size)
+            except pymongo.errors.CollectionInvalid:
+                pass
+        self.collection = db[collection]
+        self.async = async
+        logging.Handler.__init__(self, level)
+        self.formatter = MongoFormatter()
+
+    def emit(self, record):
+        # explicitly set safe=False to get async insert
+        # TODO: what to do if an error occurs? not safe to log-- ignore?
+        self.collection.save(self.format(record), safe=not self.async)

oyster/scripts/signal.py

@@ -1,30 +1,53 @@
 #!/usr/bin/env python
 import argparse
+import traceback
+import random
+from celery.execute import send_task
+from celery import current_app
 
 from oyster.core import kernel
 
-
 def main():
     parser = argparse.ArgumentParser(
         description='do a task for all documents in a doc_class',
     )
 
-    parser.add_argument('function', type=str, help='path to function to apply')
+    parser.add_argument('task', type=str, help='task name to apply')
     parser.add_argument('doc_class', type=str,
                         help='doc_class to apply function to')
+    parser.add_argument('--sample', action='store_true')
+    parser.add_argument('--immediate', action='store_true')
 
     args = parser.parse_args()
 
-    docs = kernel.db.tracked.find({'doc_class': args.doc_class})
+    docs = kernel.db.tracked.find({'doc_class': args.doc_class,
-    print '%s docs in %s' % (docs.count(), args.doc_class)
+                                   'versions': {'$ne': []}
+                                  }, timeout=False)
+    total = docs.count()
+    print '{0} docs in {1}'.format(total, args.doc_class)
 
-    path, func = args.function.rsplit('.', 1)
+    if args.sample:
-    mod = __import__(path, fromlist=[func])
+        limit = 100
-    func = getattr(mod, func)
+        print 'sampling {0} documents'.format(limit)
+        docs = docs.limit(limit).skip(random.randint(0, total-limit))
+        args.immediate = True
 
-    for doc in docs:
+    errors = 0
-        func(doc, kernel.get_last_version(doc))
+
-        # optionally save doc?
+    if args.immediate:
+        module, name = args.task.rsplit('.', 1)
+        task = getattr(__import__(module, fromlist=[name]), name)
+        for doc in docs:
+            try:
+                task.apply((doc['_id'],), throw=True)
+            except Exception:
+                errors += 1
+                traceback.print_exc()
+        print '{0} errors in {1} documents'.format(errors, total)
+
+    else:
+        for doc in docs:
+            send_task(args.task, (doc['_id'], ))
 
 if __name__ == '__main__':
     main()

oyster/storage/s3.py

@@ -14,19 +14,19 @@ class S3Storage(object):
     @property
     def bucket(self):
         if not self._bucket:
-            self._bucket = self.s3conn.create_bucket(settings.AWS_BUCKET)
+            self._bucket = self.s3conn.get_bucket(settings.AWS_BUCKET)
         return self._bucket
 
     def _get_opt(self, doc_class, setting, default=None):
         """ doc_class first, then setting, then default """
-        self.kernel.doc_classes[doc_class].get(setting,
+        return self.kernel.doc_classes[doc_class].get(setting,
-           getattr(settings, setting, default))
+                                          getattr(settings, setting, default))
 
     def put(self, tracked_doc, data, content_type):
         """ upload the document to S3 """
-        k = boto.s3.key.Key(self.bucket)
         aws_prefix = self._get_opt(tracked_doc['doc_class'], 'AWS_PREFIX', '')
         aws_bucket = self._get_opt(tracked_doc['doc_class'], 'AWS_BUCKET')
+        k = boto.s3.key.Key(self.bucket)
         key_name = aws_prefix + tracked_doc['_id']
         k.key = key_name
         headers = {'x-amz-acl': 'public-read',

oyster/tasks.py

@@ -11,10 +11,8 @@ class UpdateTask(Task):
 
     def run(self, doc_id):
         doc = kernel.db.tracked.find_one({'_id': doc_id})
-        kernel.update(doc)
-        for task in doc.get('post_update_tasks', []):
-            send_task(task, (doc_id,))
         kernel.db.status.update({}, {'$inc': {'update_queue': -1}})
+        kernel.update(doc)
         # don't sit on a connection
         kernel.db.connection.end_request()
 
@@ -24,15 +22,24 @@ class UpdateTaskScheduler(PeriodicTask):
 
     # 60s tick
     run_every = 60
+    ignore_result = True
 
     def run(self):
         # if the update queue isn't empty, wait to add more
         # (currently the only way we avoid duplicates)
         # alternate option would be to set a _queued flag on documents
-        if kernel.db.status.find_one()['update_queue']:
+        update_queue_size = kernel.db.status.find_one()['update_queue']
+        if update_queue_size:
+            self.get_logger().debug('waiting, update_queue_size={0}'.format(
+                                    update_queue_size))
             return
 
         next_set = kernel.get_update_queue()
+        if next_set:
+            self.get_logger().debug('repopulating update_queue')
+        else:
+            self.get_logger().debug('kernel.update_queue empty')
+
         for doc in next_set:
             UpdateTask.delay(doc['_id'])
             kernel.db.status.update({}, {'$inc': {'update_queue': 1}})

oyster/templates/log_row.html

@@ -1,6 +1,6 @@
-    <tr{% if log.error %} class="error" {% endif %}>
+    <tr class="{{log.levelname.lower}}">
-        <td>{{log.action}}</td>
+        <td>{{log.name}}</td>
-        <td><a href="{{request.script_root}}/tracked/{{log.url}}">{{log.url}}</td>
+        <td>{{log.message}}</td>
-        <td>{{log.timestamp.strftime("%Y-%m-%d %H:%M:%S")}}</td>
+        <td>{{log.created.strftime("%Y-%m-%d %H:%M:%S")}}</td>
-        <td>{% if log.error %}{{log.error}}{% endif %}</td>
+        <td>{% if log.exc_info %}{{log.exc_info}}{% endif %}</td>
     </tr>

oyster/tests/test_kernel.py

@@ -2,7 +2,7 @@ import time
 import datetime
 from unittest import TestCase
 
-from nose.tools import assert_raises
+from nose.tools import assert_raises, assert_equal
 
 from oyster.core import Kernel
 
@@ -44,8 +44,6 @@ class KernelTests(TestCase):
                    retry_attempts=7, retry_wait_minutes=8)
         assert c.db.connection.host == '127.0.0.1'
         assert c.db.connection.port == 27017
-        assert c.db.logs.options()['capped'] == True
-        assert c.db.logs.options()['size'] == 5000
         assert c.retry_wait_minutes == 8
         # TODO: test retry_attempts
         assert c.scraper.user_agent == 'test-ua'
@@ -55,15 +53,6 @@ class KernelTests(TestCase):
         # ensure that a bad document class raises an error
         assert_raises(ValueError, Kernel, doc_classes={'bad-doc': {}})
 
-    def test_log(self):
-        self.kernel.log('action1', 'http://example.com')
-        self.kernel.log('action2', 'http://test.com', error=True, pi=3)
-        assert self.kernel.db.logs.count() == 2
-        x = self.kernel.db.logs.find_one({'error': True})
-        assert x['action'] == 'action2'
-        assert x['url'] == 'http://test.com'
-        assert x['pi'] == 3
-
     def test_track_url(self):
         # basic insert
         id1 = self.kernel.track_url('http://example.com', 'default', pi=3)
@@ -73,11 +62,6 @@ class KernelTests(TestCase):
         assert obj['metadata'] == {'pi': 3}
         assert obj['versions'] == []
 
-        # logging
-        log = self.kernel.db.logs.find_one()
-        assert log['action'] == 'track'
-        assert log['url'] == 'http://example.com'
-
         # track same url again with same metadata returns id
         id2 = self.kernel.track_url('http://example.com', 'default', pi=3)
         assert id1 == id2
@@ -88,16 +72,12 @@ class KernelTests(TestCase):
         assert out == 'fixed-id'
 
         # can't pass track same id twice with different url
-        self.kernel.db.logs.drop()
         assert_raises(ValueError, self.kernel.track_url,
                       'http://example.com/3', 'default', 'fixed-id')
-        assert 'already exists' in self.kernel.db.logs.find_one()['error']
 
         # ... or different doc class
-        self.kernel.db.logs.drop()
         assert_raises(ValueError, self.kernel.track_url,
                       'http://example.com/2', 'change-hook', 'fixed-id')
-        assert 'already exists' in self.kernel.db.logs.find_one()['error']
 
         # different metadata is ok, but it should be updated
         self.kernel.track_url('http://example.com/2', 'default', 'fixed-id',
@@ -135,9 +115,6 @@ class KernelTests(TestCase):
 
         assert len(newobj['versions']) == 1
 
-        # check logs
-        assert self.kernel.db.logs.find({'action': 'update'}).count() == 1
-
         # and do another update..
         self.kernel.update(obj)
 
@@ -148,9 +125,6 @@ class KernelTests(TestCase):
         newobj = self.kernel.db.tracked.find_one()
         assert first_update < newobj['last_update']
 
-        # check that logs updated
-        assert self.kernel.db.logs.find({'action': 'update'}).count() == 2
-
     def test_update_failure(self):
         # track a non-existent URL
         self.kernel.track_url('http://not_a_url', 'default')
@@ -160,41 +134,37 @@ class KernelTests(TestCase):
         obj = self.kernel.db.tracked.find_one()
         assert obj['consecutive_errors'] == 1
 
-        # we should have logged an error too
-        assert self.kernel.db.logs.find({'action': 'update',
-                                         'error': {'$ne': False}}).count() == 1
-
         # update again
         self.kernel.update(obj)
 
         obj = self.kernel.db.tracked.find_one()
         assert obj['consecutive_errors'] == 2
 
-    def test_update_onchanged_fire_only_on_change(self):
+    #def test_update_onchanged_fire_only_on_change(self):
-        self.kernel.track_url('http://example.com', 'change-hook')
+    #    self.kernel.track_url('http://example.com', 'change-hook')
-        obj = self.kernel.db.tracked.find_one()
+    #    obj = self.kernel.db.tracked.find_one()
-        self.kernel.update(obj)
+    #    self.kernel.update(obj)
 
-        doc = self.kernel.db.tracked.find_one()
+    #    doc = self.kernel.db.tracked.find_one()
-        assert doc['hook_fired'] == 1
+    #    assert doc['hook_fired'] == 1
 
-        # again, we rely on example.com not updating
+    #    # again, we rely on example.com not updating
-        self.kernel.update(obj)
+    #    self.kernel.update(obj)
-        doc = self.kernel.db.tracked.find_one()
+    #    doc = self.kernel.db.tracked.find_one()
-        assert doc['hook_fired'] == 1
+    #    assert doc['hook_fired'] == 1
 
-    def test_update_onchanged_fire_again_on_change(self):
+    #def test_update_onchanged_fire_again_on_change(self):
-        self.kernel.track_url(RANDOM_URL, 'change-hook')
+    #    self.kernel.track_url(RANDOM_URL, 'change-hook')
-        obj = self.kernel.db.tracked.find_one()
+    #    obj = self.kernel.db.tracked.find_one()
-        self.kernel.update(obj)
+    #    self.kernel.update(obj)
 
-        doc = self.kernel.db.tracked.find_one()
+    #    doc = self.kernel.db.tracked.find_one()
-        assert doc['hook_fired'] == 1
+    #    assert doc['hook_fired'] == 1
 
-        # we rely on this URL updating
+    #    # we rely on this URL updating
-        self.kernel.update(obj)
+    #    self.kernel.update(obj)
-        doc = self.kernel.db.tracked.find_one()
+    #    doc = self.kernel.db.tracked.find_one()
-        assert doc['hook_fired'] == 2
+    #    assert doc['hook_fired'] == 2
 
     def test_get_update_queue(self):
         self.kernel.track_url('never-updates', 'fast-update')

oyster/tests/test_mongolog.py

@@ -0,0 +1,53 @@
+import unittest
+import logging
+import datetime
+
+import pymongo
+from ..mongolog import MongoHandler
+
+class TestMongoLog(unittest.TestCase):
+
+    DB_NAME = 'oyster_test'
+
+    def setUp(self):
+        pymongo.Connection().drop_database(self.DB_NAME)
+        self.log = logging.getLogger('mongotest')
+        self.log.setLevel(logging.DEBUG)
+        self.logs = pymongo.Connection()[self.DB_NAME]['logs']
+        # clear handlers upon each setup 
+        self.log.handlers = []
+        # async = False for testing
+        self.log.addHandler(MongoHandler(self.DB_NAME, capped_size=4000,
+                                         async=False))
+
+    def tearDown(self):
+        pymongo.Connection().drop_database(self.DB_NAME)
+
+    def test_basic_write(self):
+        self.log.debug('test')
+        self.assertEqual(self.logs.count(), 1)
+        self.log.debug('test')
+        self.assertEqual(self.logs.count(), 2)
+        # capped_size will limit these
+        self.log.debug('test'*200)
+        self.log.debug('test'*200)
+        self.assertEqual(self.logs.count(), 1)
+
+    def test_attributes(self):
+        self.log.debug('pi=%s', 3.14, extra={'pie':'pizza'})
+        logged = self.logs.find_one()
+        self.assertEqual(logged['message'], 'pi=3.14')
+        self.assertTrue(isinstance(logged['created'], datetime.datetime))
+        self.assertTrue('host' in logged)
+        self.assertEqual(logged['name'], 'mongotest')
+        self.assertEqual(logged['levelname'], 'DEBUG')
+        self.assertEqual(logged['pie'], 'pizza')
+
+        # and exc_info
+        try:
+            raise Exception('error!')
+        except:
+            self.log.warning('oh no!', exc_info=True)
+            logged = self.logs.find_one(sort=[('$natural', -1)])
+        self.assertEqual(logged['levelname'], 'WARNING')
+        self.assertTrue('error!' in logged['exc_info'])

oyster/tests/test_storage.py

@@ -8,14 +8,14 @@ from oyster.storage.dummy import DummyStorage
 
 def _simple_storage_test(StorageCls):
     kernel = Kernel(mongo_db='oyster_test')
+    kernel.doc_classes['default'] = {}
     storage = StorageCls(kernel)
 
     # ensure the class has a storage_type attribute
     assert hasattr(storage, 'storage_type')
 
     doc = {'_id': 'aabbccddeeff', 'url': 'http://localhost:8000/#test',
-           'metadata': {}
+           'doc_class': 'default', 'metadata': {} }
-          }
     storage_id = storage.put(doc, 'hello oyster', 'text/plain')
     assert storage_id
 

oyster/web.py

@@ -4,7 +4,7 @@ import datetime
 import functools
 
 import flask
-import pymongo.objectid
+import bson.objectid
 
 from oyster.conf import settings
 from oyster.core import kernel
@@ -14,7 +14,7 @@ class JSONEncoder(json.JSONEncoder):
     def default(self, obj):
         if isinstance(obj, datetime.datetime):
             return obj.isoformat()
-        elif isinstance(obj, pymongo.objectid.ObjectId):
+        elif isinstance(obj, bson.objectid.ObjectId):
             return str(obj)
         else:
             return super(JSONEncoder, self).default(obj)
@@ -51,22 +51,12 @@ def index():
     status = {
         'tracking': kernel.db.tracked.count(),
         'need_update': kernel.get_update_queue_size(),
-        'logs': list(kernel.db.logs.find().sort('$natural', -1).limit(20)),
+        'logs': list(kernel.db.logs.find().sort('$natural', -1).limit(100)),
         'mongo_host': settings.MONGO_HOST,
     }
     return status
 
 
-@app.route('/status/')
-@api_wrapper()
-def doc_list():
-    status = {
-        'tracking': kernel.db.tracked.count(),
-        'need_update': kernel.get_update_queue_size(),
-    }
-    return status
-
-
 @app.route('/log/')
 @api_wrapper('logs.html')
 def log_view():
@@ -86,23 +76,11 @@ def tracked():
     return json.dumps(tracked, cls=JSONEncoder)
 
 
-@app.route('/tracked/<path:url>')
+@app.route('/tracked/<id>')
-def tracked_view(url):
+def tracked_view(id):
-    url = _path_fixer(url)
+    doc = kernel.db.tracked.find_one({'_id': id})
-    doc = kernel.db.tracked.find_one({'url': url})
     return json.dumps(doc, cls=JSONEncoder)
 
 
-@app.route('/doc/<path:url>/<version>')
-def show_doc(url, version):
-    url = _path_fixer(url)
-    if version == 'latest':
-        version = -1
-    doc = kernel.get_version(url, version)
-    resp = flask.make_response(doc.read())
-    resp.headers['content-type'] = doc.content_type
-    return resp
-
-
 if __name__ == '__main__':
     app.run(debug=True)

requirements.txt

@@ -2,4 +2,4 @@ scrapelib
 pymongo>=2.0
 flask
 nose
-celery
+celery==2.5.3

setup.py

@@ -1,11 +1,21 @@
 #!/usr/bin/env python
+
+import os
 from setuptools import setup
-from oyster import __version__
+
+# Hack to prevent stupid "TypeError: 'NoneType' object is not callable" error
+# in multiprocessing/util.py _exit_function when running `python
+# setup.py test` (see
+# http://www.eby-sarna.com/pipermail/peak/2010-May/003357.html)
+try:
+    import multiprocessing
+except ImportError:
+    pass
 
 long_description = open('README.rst').read()
 
 setup(name="oyster",
-      version=__version__,
+      version='0.4.0-dev',
       py_modules=['oyster'],
       author="James Turk",
       author_email='jturk@sunlightfoundation.com',
@@ -21,8 +31,10 @@ setup(name="oyster",
                    "Operating System :: OS Independent",
                    "Programming Language :: Python",
                    ],
-      install_requires=["httplib2 >= 0.6.0", "scrapelib >= 0.5.4",
+      install_requires=["httplib2 >= 0.6.0", "scrapelib >= 0.7.2",
                         "pymongo >= 1.11", "flask", "celery"],
+      tests_require=["nose"],
+      test_suite='nose.collector',
       entry_points="""
 [console_scripts]
 scrapeshell = scrapelib:scrapeshell

tox.ini

@@ -0,0 +1,10 @@
+# Tox (http://codespeak.net/~hpk/tox/) is a tool for running tests
+# in multiple virtualenvs. This configuration file will run the
+# test suite on all supported python versions. To use it, "pip install tox"
+# and then run "tox" from this directory.
+
+[tox]
+envlist = py26, py27, pypy
+
+[testenv]
+commands = python setup.py test