diff --git a/PKG-INFO b/PKG-INFO index c4d090d..4504fa9 100644 --- a/PKG-INFO +++ b/PKG-INFO @@ -1,10 +1,10 @@ Metadata-Version: 1.0 Name: swh.objstorage -Version: 0.0.20 +Version: 0.0.21 Summary: Software Heritage Object Storage Home-page: https://forge.softwareheritage.org/diffusion/DOBJS Author: Software Heritage developers Author-email: swh-devel@inria.fr License: UNKNOWN Description: UNKNOWN Platform: UNKNOWN diff --git a/swh.objstorage.egg-info/PKG-INFO b/swh.objstorage.egg-info/PKG-INFO index c4d090d..4504fa9 100644 --- a/swh.objstorage.egg-info/PKG-INFO +++ b/swh.objstorage.egg-info/PKG-INFO @@ -1,10 +1,10 @@ Metadata-Version: 1.0 Name: swh.objstorage -Version: 0.0.20 +Version: 0.0.21 Summary: Software Heritage Object Storage Home-page: https://forge.softwareheritage.org/diffusion/DOBJS Author: Software Heritage developers Author-email: swh-devel@inria.fr License: UNKNOWN Description: UNKNOWN Platform: UNKNOWN diff --git a/swh/objstorage/api/server.py b/swh/objstorage/api/server.py index 4a606f8..024fd69 100644 --- a/swh/objstorage/api/server.py +++ b/swh/objstorage/api/server.py @@ -1,143 +1,147 @@ # Copyright (C) 2015-2017 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 import asyncio import aiohttp.web import click from swh.core import config from swh.core.api_async import (SWHRemoteAPI, decode_request, encode_data_server as encode_data) from swh.model import hashutil from swh.objstorage import get_objstorage DEFAULT_CONFIG_PATH = 'objstorage/server' DEFAULT_CONFIG = { 'cls': ('str', 'pathslicing'), 'args': ('dict', { 'root': '/srv/softwareheritage/objects', 'slicing': '0:2/2:4/4:6', - }) + }), + 'client_max_size': ('int', 1024 * 1024 * 1024), } @asyncio.coroutine def index(request): return aiohttp.web.Response(body="SWH Objstorage API server") @asyncio.coroutine def check_config(request): req = yield from decode_request(request) return encode_data(request.app['objstorage'].check_config(**req)) @asyncio.coroutine def contains(request): req = yield from decode_request(request) return encode_data(request.app['objstorage'].__contains__(**req)) @asyncio.coroutine def add_bytes(request): req = yield from decode_request(request) return encode_data(request.app['objstorage'].add(**req)) @asyncio.coroutine def get_bytes(request): req = yield from decode_request(request) return encode_data(request.app['objstorage'].get(**req)) @asyncio.coroutine def get_batch(request): req = yield from decode_request(request) return encode_data(request.app['objstorage'].get_batch(**req)) @asyncio.coroutine def check(request): req = yield from decode_request(request) return encode_data(request.app['objstorage'].check(**req)) # Management methods @asyncio.coroutine def get_random_contents(request): req = yield from decode_request(request) return encode_data(request.app['objstorage'].get_random(**req)) # Streaming methods @asyncio.coroutine def add_stream(request): hex_id = request.match_info['hex_id'] obj_id = hashutil.hash_to_bytes(hex_id) check_pres = (request.query.get('check_presence', '').lower() == 'true') objstorage = request.app['objstorage'] if check_pres and obj_id in objstorage: return encode_data(obj_id) with objstorage.chunk_writer(obj_id) as write: # XXX (3.5): use 'async for chunk in request.content.iter_any()' while not request.content.at_eof(): chunk = yield from request.content.readany() write(chunk) return encode_data(obj_id) @asyncio.coroutine def get_stream(request): hex_id = request.match_info['hex_id'] obj_id = hashutil.hash_to_bytes(hex_id) response = aiohttp.web.StreamResponse() yield from response.prepare(request) for chunk in request.app['objstorage'].get_stream(obj_id, 2 << 20): response.write(chunk) yield from response.drain() return response def make_app(config, **kwargs): + if 'client_max_size' in config: + kwargs['client_max_size'] = config['client_max_size'] + app = SWHRemoteAPI(**kwargs) app.router.add_route('GET', '/', index) app.router.add_route('POST', '/check_config', check_config) app.router.add_route('POST', '/content/contains', contains) app.router.add_route('POST', '/content/add', add_bytes) app.router.add_route('POST', '/content/get', get_bytes) app.router.add_route('POST', '/content/get/batch', get_batch) app.router.add_route('POST', '/content/get/random', get_random_contents) app.router.add_route('POST', '/content/check', check) app.router.add_route('POST', '/content/add_stream/{hex_id}', add_stream) app.router.add_route('GET', '/content/get_stream/{hex_id}', get_stream) app.update(config) app['objstorage'] = get_objstorage(app['cls'], app['args']) return app def make_app_from_configfile(config_path=DEFAULT_CONFIG_PATH, **kwargs): return make_app(config.read(config_path, DEFAULT_CONFIG), **kwargs) @click.command() @click.argument('config-path', required=1) @click.option('--host', default='0.0.0.0', help="Host to run the server") @click.option('--port', default=5003, type=click.INT, help="Binding port of the server") @click.option('--debug/--nodebug', default=True, help="Indicates if the server should run in debug mode") def launch(config_path, host, port, debug): app = make_app_from_configfile(config_path, debug=bool(debug)) aiohttp.web.run_app(app, host=host, port=int(port)) if __name__ == '__main__': launch() diff --git a/swh/objstorage/tests/objstorage_testing.py b/swh/objstorage/tests/objstorage_testing.py index 3d9e067..47131ea 100644 --- a/swh/objstorage/tests/objstorage_testing.py +++ b/swh/objstorage/tests/objstorage_testing.py @@ -1,137 +1,144 @@ # Copyright (C) 2015-2017 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 import time from nose.tools import istest from swh.model import hashutil from swh.objstorage import exc class ObjStorageTestFixture(): def setUp(self): super().setUp() def hash_content(self, content): obj_id = hashutil.hash_data(content)['sha1'] return content, obj_id def assertContentMatch(self, obj_id, expected_content): # noqa content = self.storage.get(obj_id) self.assertEqual(content, expected_content) @istest def check_config(self): self.assertTrue(self.storage.check_config(check_write=False)) self.assertTrue(self.storage.check_config(check_write=True)) @istest def contains(self): content_p, obj_id_p = self.hash_content(b'contains_present') content_m, obj_id_m = self.hash_content(b'contains_missing') self.storage.add(content_p, obj_id=obj_id_p) self.assertIn(obj_id_p, self.storage) self.assertNotIn(obj_id_m, self.storage) @istest def add_get_w_id(self): content, obj_id = self.hash_content(b'add_get_w_id') r = self.storage.add(content, obj_id=obj_id) self.assertEqual(obj_id, r) self.assertContentMatch(obj_id, content) + @istest + def add_big(self): + content, obj_id = self.hash_content(b'add_big' * 1024 * 1024) + r = self.storage.add(content, obj_id=obj_id) + self.assertEqual(obj_id, r) + self.assertContentMatch(obj_id, content) + @istest def add_get_wo_id(self): content, obj_id = self.hash_content(b'add_get_wo_id') r = self.storage.add(content) self.assertEqual(obj_id, r) self.assertContentMatch(obj_id, content) @istest def add_get_batch(self): content1, obj_id1 = self.hash_content(b'add_get_batch_1') content2, obj_id2 = self.hash_content(b'add_get_batch_2') self.storage.add(content1, obj_id1) self.storage.add(content2, obj_id2) cr1, cr2 = self.storage.get_batch([obj_id1, obj_id2]) self.assertEqual(cr1, content1) self.assertEqual(cr2, content2) @istest def get_batch_unexisting_content(self): content, obj_id = self.hash_content(b'get_batch_unexisting_content') result = list(self.storage.get_batch([obj_id])) self.assertTrue(len(result) == 1) self.assertIsNone(result[0]) @istest def restore_content(self): valid_content, valid_obj_id = self.hash_content(b'restore_content') invalid_content = b'unexpected content' id_adding = self.storage.add(invalid_content, valid_obj_id) id_restore = self.storage.restore(valid_content) # Adding a false content then restore it to the right one and # then perform a verification should result in a successful check. self.assertEqual(id_adding, valid_obj_id) self.assertEqual(id_restore, valid_obj_id) self.assertContentMatch(valid_obj_id, valid_content) @istest def get_missing(self): content, obj_id = self.hash_content(b'get_missing') with self.assertRaises(exc.Error): self.storage.get(obj_id) @istest def check_missing(self): content, obj_id = self.hash_content(b'check_missing') with self.assertRaises(exc.Error): self.storage.check(obj_id) @istest def check_present(self): content, obj_id = self.hash_content(b'check_missing') self.storage.add(content) try: self.storage.check(obj_id) except: self.fail('Integrity check failed') @istest def add_stream(self): content = [b'chunk1', b'chunk2'] _, obj_id = self.hash_content(b''.join(content)) try: self.storage.add_stream(iter(content), obj_id=obj_id) except NotImplementedError: return self.assertContentMatch(obj_id, b''.join(content)) @istest def add_stream_sleep(self): def gen_content(): yield b'chunk1' time.sleep(0.5) yield b'chunk2' _, obj_id = self.hash_content(b'placeholder_id') try: self.storage.add_stream(gen_content(), obj_id=obj_id) except NotImplementedError: return self.assertContentMatch(obj_id, b'chunk1chunk2') @istest def get_stream(self): content_l = [b'1', b'2', b'3', b'4', b'5', b'6', b'7', b'8', b'9'] content = b''.join(content_l) _, obj_id = self.hash_content(content) self.storage.add(content, obj_id=obj_id) try: r = list(self.storage.get_stream(obj_id, chunk_size=1)) except NotImplementedError: return self.assertEqual(r, content_l) diff --git a/swh/objstorage/tests/test_objstorage_api.py b/swh/objstorage/tests/test_objstorage_api.py index 801284b..0070e51 100644 --- a/swh/objstorage/tests/test_objstorage_api.py +++ b/swh/objstorage/tests/test_objstorage_api.py @@ -1,36 +1,37 @@ # Copyright (C) 2015 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 import tempfile import unittest from nose.plugins.attrib import attr from swh.objstorage import get_objstorage from swh.objstorage.tests.objstorage_testing import ObjStorageTestFixture from swh.objstorage.tests.server_testing import ServerTestFixture from swh.objstorage.api.server import make_app @attr('db') class TestRemoteObjStorage(ServerTestFixture, ObjStorageTestFixture, unittest.TestCase): """ Test the remote archive API. """ def setUp(self): self.config = { 'cls': 'pathslicing', 'args': { 'root': tempfile.mkdtemp(), 'slicing': '0:1/0:5', - } + }, + 'client_max_size': 8 * 1024 * 1024, } self.app = make_app(self.config) super().setUp() self.storage = get_objstorage('remote', { 'url': self.url() }) diff --git a/version.txt b/version.txt index c8c3bdd..8bc3a77 100644 --- a/version.txt +++ b/version.txt @@ -1 +1 @@ -v0.0.20-0-g7418ec0 \ No newline at end of file +v0.0.21-0-ge70f532 \ No newline at end of file