lbry-sdk/lbrynet/dht/kbucket.py
Jack Robison 23c202b5e4
refactor Contact class, DHT RPCs, and Contact addition/removal
-track contact failures, last replied, and last requested. use this to provide a 'contact_is_good' property on Contact objects
-ensure no duplicate contact objects are created
-remove confusing conflation of node id strings with Contact objects, update docstrings
-move RPC failure tracking to a callback/errback pair in sendRPC (so the contact is only updated once)
-handle seed nodes during the join sequence by setting their node ids after they initially reply to our ping
-name all of the kademlia RPC keyword args, remove confusing **kwargs and dictionary parsing
-add host ip/port to DHT send/receive logging to make the results comprehensible when running many nodes at once
2018-06-07 15:02:47 -04:00

143 lines
5 KiB
Python

import constants
from error import BucketFull
class KBucket(object):
""" Description - later
"""
def __init__(self, rangeMin, rangeMax, node_id):
"""
@param rangeMin: The lower boundary for the range in the n-bit ID
space covered by this k-bucket
@param rangeMax: The upper boundary for the range in the ID space
covered by this k-bucket
"""
self.lastAccessed = 0
self.rangeMin = rangeMin
self.rangeMax = rangeMax
self._contacts = list()
self._node_id = node_id
def addContact(self, contact):
""" Add contact to _contact list in the right order. This will move the
contact to the end of the k-bucket if it is already present.
@raise kademlia.kbucket.BucketFull: Raised when the bucket is full and
the contact isn't in the bucket
already
@param contact: The contact to add
@type contact: kademlia.contact.Contact
"""
if contact in self._contacts:
# Move the existing contact to the end of the list
# - using the new contact to allow add-on data
# (e.g. optimization-specific stuff) to pe updated as well
self._contacts.remove(contact)
self._contacts.append(contact)
elif len(self._contacts) < constants.k:
self._contacts.append(contact)
else:
raise BucketFull("No space in bucket to insert contact")
def getContact(self, contactID):
"""Get the contact specified node ID
@raise IndexError: raised if the contact is not in the bucket
@param contactID: the node id of the contact to retrieve
@type contactID: str
@rtype: dht.contact._Contact
"""
for contact in self._contacts:
if contact.id == contactID:
return contact
raise IndexError(contactID)
def getContacts(self, count=-1, excludeContact=None):
""" Returns a list containing up to the first count number of contacts
@param count: The amount of contacts to return (if 0 or less, return
all contacts)
@type count: int
@param excludeContact: A contact to exclude; if this contact is in
the list of returned values, it will be
discarded before returning. If a C{str} is
passed as this argument, it must be the
contact's ID.
@type excludeContact: kademlia.contact.Contact or str
@raise IndexError: If the number of requested contacts is too large
@return: Return up to the first count number of contacts in a list
If no contacts are present an empty is returned
@rtype: list
"""
# Return all contacts in bucket
if count <= 0:
count = len(self._contacts)
# Get current contact number
currentLen = len(self._contacts)
# If count greater than k - return only k contacts
if count > constants.k:
count = constants.k
# Check if count value in range and,
# if count number of contacts are available
if not currentLen:
contactList = list()
# length of list less than requested amount
elif currentLen < count:
contactList = self._contacts[0:currentLen]
# enough contacts in list
else:
contactList = self._contacts[0:count]
if excludeContact in contactList:
contactList.remove(excludeContact)
def getBadOrUnknownContacts(self):
contacts = self.getContacts(sort_distance_to=False)
results = [contact for contact in contacts if contact.contact_is_good is False]
results.extend(contact for contact in contacts if contact.contact_is_good is None)
return results
return contactList
def removeContact(self, contact):
""" Remove the contact from the bucket
@param contact: The contact to remove
@type contact: dht.contact._Contact
@raise ValueError: The specified contact is not in this bucket
"""
self._contacts.remove(contact)
def keyInRange(self, key):
""" Tests whether the specified key (i.e. node ID) is in the range
of the n-bit ID space covered by this k-bucket (in otherwords, it
returns whether or not the specified key should be placed in this
k-bucket)
@param key: The key to test
@type key: str or int
@return: C{True} if the key is in this k-bucket's range, or C{False}
if not.
@rtype: bool
"""
if isinstance(key, str):
key = long(key.encode('hex'), 16)
return self.rangeMin <= key < self.rangeMax
def __len__(self):
return len(self._contacts)
def __contains__(self, item):
return item in self._contacts