Merge pull request #301 from lbryio/query-available-blobs

Query available blobs and Upload to Lighthouse Scripts
This commit is contained in:
Job Evers‐Meltzer 2017-01-12 08:42:08 -06:00 committed by GitHub
commit 61702ba1e4
8 changed files with 540 additions and 2 deletions

View file

@ -194,7 +194,7 @@ class HashBlob(object):
return self.blob_hash[:16] return self.blob_hash[:16]
def __repr__(self): def __repr__(self):
return str(self) return '<{}({})>'.format(self.__class__.__name__, str(self))
class BlobFile(HashBlob): class BlobFile(HashBlob):

View file

@ -38,7 +38,7 @@ class Metadata(StructuredDict):
def __init__(self, metadata, migrate=True, target_version=None): def __init__(self, metadata, migrate=True, target_version=None):
if not isinstance(metadata, dict): if not isinstance(metadata, dict):
raise TypeError("metadata is not a dictionary") raise TypeError("{} is not a dictionary".format(metadata))
starting_version = metadata.get('ver', '0.0.1') starting_version = metadata.get('ver', '0.0.1')
StructuredDict.__init__(self, metadata, starting_version, migrate, target_version) StructuredDict.__init__(self, metadata, starting_version, migrate, target_version)

23
scripts/common.py Normal file
View file

@ -0,0 +1,23 @@
import logging
from twisted.internet import defer
log = logging.getLogger(__name__)
@defer.inlineCallbacks
def getNames(wallet, names=None):
if names:
defer.returnValue(names)
nametrie = yield wallet.get_nametrie()
defer.returnValue(list(getNameClaims(nametrie)))
def getNameClaims(trie):
for x in trie:
if 'txid' in x:
try:
yield str(x['name'])
except UnicodeError:
log.warning('Skippin name %s as it is not ascii', x['name'])

59
scripts/name.py Normal file
View file

@ -0,0 +1,59 @@
import logging
import random
from twisted.internet import defer
from twisted.internet import reactor
from lbrynet.core import Error
from lbrynet.core import StreamDescriptor
from lbrynet.metadata import Metadata
log = logging.getLogger(__name__)
class Name(object):
def __init__(self, name):
self.name = name
self.sd_hash = None
self.sd_blob = None
@defer.inlineCallbacks
def setSdHash(self, wallet):
try:
stream = yield wallet.get_stream_info_for_name(self.name)
metadata = Metadata.Metadata(stream)
self.sd_hash = _getSdHash(metadata)
except (Error.InvalidStreamInfoError, AssertionError):
pass
except Exception:
log.exception('What happened')
@defer.inlineCallbacks
def download_sd_blob(self, session):
print('Trying to get sd_blob for {} using {}'.format(self.name, self.sd_hash))
try:
blob = yield download_sd_blob_with_timeout(
session, self.sd_hash, session.payment_rate_manager)
self.sd_blob = blob
yield self._after_download(blob)
print('Downloaded sd_blob for {} using {}'.format(self.name, self.sd_hash))
except defer.TimeoutError:
print('Downloading sd_blob for {} timed-out'.format(self.name))
# swallow errors from the timeout
pass
except Exception:
log.exception('Failed to download {}'.format(self.name))
def _after_download(self, blob):
pass
def _getSdHash(metadata):
return metadata['sources']['lbry_sd_hash']
def download_sd_blob_with_timeout(session, sd_hash, payment_rate_manager):
d = StreamDescriptor.download_sd_blob(session, sd_hash, payment_rate_manager)
d.addTimeout(random.randint(10, 30), reactor)
return d

50
scripts/pool.py Normal file
View file

@ -0,0 +1,50 @@
import itertools
import logging
from twisted.internet import defer
log = logging.getLogger(__name__)
class DeferredPool(defer.Deferred):
def __init__(self, deferred_iter, pool_size):
self.deferred_iter = deferred_iter
self.pool_size = pool_size
# results are stored unordered
self.result_list = []
self.started_count = 0
self.total_count = None
defer.Deferred.__init__(self)
for deferred in itertools.islice(deferred_iter, pool_size):
self._start_one(deferred)
def _start_one(self, deferred):
deferred.addCallbacks(self._callback, self._callback,
callbackArgs=(self.started_count, defer.SUCCESS),
errbackArgs=(self.started_count, defer.FAILURE))
self.started_count += 1
def _callback(self, result, index, success):
self.result_list.append((index, success, result))
if self._done():
self._finish()
else:
self._process_next()
return result
def _done(self):
return self.total_count == len(self.result_list)
def _finish(self):
result_list = [(s, r) for i, s, r in sorted(self.result_list)]
self.callback(result_list)
def _process_next(self):
try:
deferred = next(self.deferred_iter)
except StopIteration:
self.total_count = self.started_count
else:
self._start_one(deferred)

View file

@ -0,0 +1,186 @@
from __future__ import print_function
from lbrynet.core import log_support
import argparse
import collections
import itertools
import logging
import os
import random
import shutil
import sys
import tempfile
import appdirs
from twisted.internet import defer
from twisted.internet import reactor
from twisted.internet import protocol
from twisted.internet import endpoints
from lbrynet import analytics
from lbrynet import conf
from lbrynet.core import Error
from lbrynet.core import Wallet
from lbrynet.core import BlobAvailability
from lbrynet.core import BlobManager
from lbrynet.core import HashAnnouncer
from lbrynet.core import PeerManager
from lbrynet.core import Session
from lbrynet.core import utils
from lbrynet.core.client import DHTPeerFinder
from lbrynet.dht import node
from lbrynet.metadata import Metadata
from lbrynet.core import StreamDescriptor as sd
import common
import name
import pool
import track
log = logging.getLogger()
def main(args=None):
conf.initialize_settings()
parser = argparse.ArgumentParser()
parser.add_argument('--limit', type=int)
parser.add_argument('--download', action='store_true',
help='Set flag to also download each sd_blob and report on success')
args = parser.parse_args(args)
log_support.configure_console()
log_support.configure_twisted()
# make a fresh dir or else we will include blobs that we've
# already downloaded but might not otherwise be available.
db_dir = tempfile.mkdtemp()
try:
blob_dir = os.path.join(db_dir, 'blobfiles')
os.makedirs(blob_dir)
storage = Wallet.InMemoryStorage()
wallet = Wallet.LBRYumWallet(storage)
session = Session.Session(
0,
db_dir=db_dir,
lbryid=utils.generate_id(),
blob_dir=blob_dir,
dht_node_port=4444,
known_dht_nodes=conf.settings.known_dht_nodes,
peer_port=3333,
use_upnp=False,
wallet=wallet
)
api = analytics.Api.new_instance()
run(args, session, api)
reactor.run()
finally:
shutil.rmtree(db_dir)
@defer.inlineCallbacks
def run(args, session, api):
try:
yield session.setup()
names = yield common.getNames(session.wallet)
if args.limit and len(names) > args.limit:
names = random.sample(list(names), args.limit)
names = [Name(n) for n in names]
blob_tracker = BlobAvailability.BlobAvailabilityTracker(
session.blob_manager, session.peer_finder, session.dht_node)
tracker = yield Tracker(session, names, blob_tracker)
yield tracker.processNameClaims(args.download)
event = makeEvent(tracker.stats)
if args.download and not args.limit:
api.track(event)
else:
# don't send event to analytics if it doesn't contain the full info
print(event)
except Exception:
log.exception('Something bad happened')
finally:
reactor.stop()
class Tracker(track.Tracker):
def __init__(self, session, names, blob_tracker):
track.Tracker.__init__(self, session, names)
self.blob_tracker = blob_tracker
@defer.inlineCallbacks
def processNameClaims(self, download=False):
try:
yield self._getSdHashes()
yield self._filterNames('sd_hash')
yield self._checkAvailability()
yield self._filterNames('is_available')
yield self.print_attempts_counter()
if download:
yield self._downloadAllBlobs()
yield self._filterNames('sd_blob')
except Exception:
log.exception('Something bad happened')
def print_attempts_counter(self):
print(self.attempts_counter)
def attempts_counter(self):
return collections.Counter([n.availability_attempts for n in self.names])
def _checkAvailability(self):
return pool.DeferredPool(
(n.check_availability(self.blob_tracker) for n in self.names),
10
)
class Name(name.Name):
# From experience, very few sd_blobs get found after the third attempt
MAX_ATTEMPTS = 6
def __init__(self, my_name):
name.Name.__init__(self, my_name)
self.is_available = None
self.availability_attempts = 0
@defer.inlineCallbacks
def _check_availability(self, blob_tracker):
b = yield blob_tracker.get_blob_availability(self.sd_hash)
peer_count = b[self.sd_hash]
self._setAvailable(peer_count)
@defer.inlineCallbacks
def check_availability(self, blob_tracker):
while not self.is_available and self.availability_attempts < self.MAX_ATTEMPTS:
self.availability_attempts += 1
log.info('Attempt %s to find %s', self.availability_attempts, self.name)
yield self._check_availability(blob_tracker)
def _setAvailable(self, peer_count):
self.is_available = peer_count > 0
def makeEvent(stats):
return {
'userId': 'lbry',
'event': 'Content Availability',
'properties': {
'total_published': stats['sd_hash'],
'sd_blob_available_on_dht': stats['is_available'],
'sd_blob_available_for_download': stats['sd_blob'],
},
'context': {
'app': {
'name': 'Availability Tracker',
'version': 1,
},
'library': {
'name': 'lbrynet-analytics',
'version': '1.0.0'
},
},
'timestamp': utils.isonow()
}
if __name__ == '__main__':
sys.exit(main())

View file

@ -0,0 +1,176 @@
from __future__ import print_function
from lbrynet.core import log_support
import argparse
import collections
import itertools
import logging
import os
import random
import sys
import appdirs
from twisted.internet import defer
from twisted.internet import reactor
from twisted.internet import protocol
from twisted.internet import endpoints
from lbrynet import conf
from lbrynet.core import Error
from lbrynet.core import Wallet
from lbrynet.core import BlobAvailability
from lbrynet.core import BlobManager
from lbrynet.core import HashAnnouncer
from lbrynet.core import PeerManager
from lbrynet.core import Session
from lbrynet.core import utils
from lbrynet.core.client import DHTPeerFinder
from lbrynet.dht import node
from lbrynet.metadata import Metadata
from lbrynet.core import StreamDescriptor as sd
from lbrynet import reflector
import common
import name
import pool
import track
log = logging.getLogger('main')
def main(args=None):
conf.initialize_settings()
parser = argparse.ArgumentParser()
parser.add_argument('destination', type=conf.server_port, nargs='+')
parser.add_argument('--names', nargs='*')
parser.add_argument('--limit', type=int)
args = parser.parse_args(args)
log_support.configure_console(level='INFO')
db_dir = appdirs.user_data_dir('lighthouse-uploader')
safe_makedirs(db_dir)
# no need to persist metadata info
storage = Wallet.InMemoryStorage()
wallet = Wallet.LBRYumWallet(storage)
blob_dir = os.path.join(db_dir, 'blobfiles')
safe_makedirs(blob_dir)
# Don't set a hash_announcer, we have no need to tell anyone we
# have these blobs
blob_manager = BlobManager.DiskBlobManager(None, blob_dir, db_dir)
# TODO: make it so that I can disable the BlobAvailabilityTracker
# or, in general, make the session more reusable for users
# that only want part of the functionality
session = Session.Session(
blob_data_payment_rate=0,
db_dir=db_dir,
lbryid=utils.generate_id(),
blob_dir=blob_dir,
dht_node_port=4444,
known_dht_nodes=conf.settings.known_dht_nodes,
peer_port=3333,
use_upnp=False,
wallet=wallet,
blob_manager=blob_manager,
)
assert session.wallet
run(session, args.destination, args.names, args.limit)
reactor.run()
def safe_makedirs(directory):
try:
os.makedirs(directory)
except OSError:
pass
@defer.inlineCallbacks
def run(session, destinations, names, limit):
try:
yield session.setup()
while not session.wallet.network.is_connected():
log.info('Retrying wallet startup')
try:
yield session.wallet.start()
except ValueError:
pass
names = yield getNames(session.wallet, names)
if limit and limit < len(names):
names = random.sample(names, limit)
log.info('Processing %s names', len(names))
names = [Name(n, session.blob_manager) for n in names]
t = Tracker(session, destinations, names)
yield t.processNameClaims()
except Exception:
log.exception('Something bad happened')
finally:
log.warning('We are stopping the reactor gracefully')
reactor.stop()
def logAndStop(err):
log_support.failure(err, log, 'This sucks: %s')
reactor.stop()
def logAndRaise(err):
log_support.failure(err, log, 'This still sucks: %s')
return err
class Tracker(track.Tracker):
def __init__(self, session, destinations, names):
self.destinations = destinations
track.Tracker.__init__(self, session, names)
@property
def blob_manager(self):
return self.session.blob_manager
@defer.inlineCallbacks
def processNameClaims(self):
yield super(Tracker, self).processNameClaims()
log.info('Sending the blobs')
yield self._sendSdBlobs()
@defer.inlineCallbacks
def _sendSdBlobs(self):
blobs = [n.sd_blob for n in self.names if n.sd_blob]
log.info('Sending %s blobs', len(blobs))
blob_hashes = [b.blob_hash for b in blobs]
for destination in self.destinations:
factory = reflector.BlobClientFactory(self.blob_manager, blob_hashes)
yield self._connect(destination, factory)
@defer.inlineCallbacks
def _connect(self, destination, factory):
url, port = destination
ip = yield reactor.resolve(url)
try:
print('Connecting to {}'.format(ip))
yield reactor.connectTCP(ip, port, factory)
#factory.finished_deferred.addTimeout(60, reactor)
value = yield factory.finished_deferred
if value:
print('Success!')
else:
print('Not success?', value)
except Exception:
log.exception('Somehow failed to send blobs')
class Name(name.Name):
def __init__(self, my_name, blob_manager):
name.Name.__init__(self, my_name)
self.blob_manager = blob_manager
def _after_download(self, blob):
# keep the blob for future runs
self.blob_manager.blob_completed(blob)
if __name__ == '__main__':
sys.exit(main())

44
scripts/track.py Normal file
View file

@ -0,0 +1,44 @@
import logging
import random
from twisted.internet import defer
from lbrynet.core import Error
import pool
log = logging.getLogger(__name__)
class Tracker(object):
def __init__(self, session, names):
self.session = session
self.names = names
self.stats = {}
@property
def wallet(self):
return self.session.wallet
@defer.inlineCallbacks
def processNameClaims(self):
try:
log.info('Starting to get name claims')
yield self._getSdHashes()
self._filterNames('sd_hash')
log.info('Downloading all of the blobs')
yield self._downloadAllBlobs()
except Exception:
log.exception('Something bad happened')
def _getSdHashes(self):
return pool.DeferredPool((n.setSdHash(self.wallet) for n in self.names), 10)
def _filterNames(self, attr):
self.names = [n for n in self.names if getattr(n, attr)]
self.stats[attr] = len(self.names)
print("We have {} names with attribute {}".format(len(self.names), attr))
def _downloadAllBlobs(self):
return pool.DeferredPool((n.download_sd_blob(self.session) for n in self.names), 10)