release initial code
Co-authored-by: Ying Sheng <sqy1415@gmail.com> Co-authored-by: Liangsheng Yin <hnyls2002@gmail.com> Co-authored-by: Zhiqiang Xie <xiezhq@stanford.edu> Co-authored-by: parasol-aser <3848358+parasol-aser@users.noreply.github.com> Co-authored-by: LiviaSun <33578456+ChuyueSun@users.noreply.github.com> Co-authored-by: Cody Yu <hao.yu.cody@gmail.com>
This commit is contained in:
co-authored by
Ying Sheng
Liangsheng Yin
Zhiqiang Xie
parasol-aser
LiviaSun
Cody Yu
parent
f6d40df0ee
commit
22085081bb
@@ -0,0 +1,220 @@
|
||||
import heapq
|
||||
import time
|
||||
from collections import defaultdict
|
||||
from dataclasses import dataclass
|
||||
from typing import Tuple
|
||||
|
||||
import torch
|
||||
|
||||
|
||||
class TreeNode:
|
||||
def __init__(self):
|
||||
self.children = defaultdict(TreeNode)
|
||||
self.parent = None
|
||||
self.value = None
|
||||
self.ref_counter = 0
|
||||
self.last_access_time = time.time()
|
||||
|
||||
def __lt__(self, other):
|
||||
return self.last_access_time < other.last_access_time
|
||||
|
||||
|
||||
def match(key, seq):
|
||||
i = 0
|
||||
for k, w in zip(key, seq):
|
||||
if k != w:
|
||||
break
|
||||
i += 1
|
||||
return i
|
||||
|
||||
|
||||
class RadixCache:
|
||||
def __init__(self, disable=False):
|
||||
self.root_node = TreeNode()
|
||||
self.root_node.value = []
|
||||
self.root_node.ref_counter = 1
|
||||
self.evictable_size_ = 0
|
||||
|
||||
self.disable = disable
|
||||
|
||||
##### Public API #####
|
||||
def match_prefix(self, key):
|
||||
if self.disable:
|
||||
return [], self.root_node
|
||||
|
||||
value = []
|
||||
last_node = [self.root_node]
|
||||
self._match_prefix_helper(self.root_node, key, value, last_node)
|
||||
if value:
|
||||
value = torch.concat(value)
|
||||
return value, last_node[0]
|
||||
|
||||
def insert(self, key, value=None):
|
||||
if self.disable:
|
||||
return len(key)
|
||||
|
||||
if value is None:
|
||||
value = [x for x in key]
|
||||
return self._insert_helper(self.root_node, key, value)
|
||||
|
||||
def pretty_print(self):
|
||||
self._print_helper(self.root_node, 0)
|
||||
print(f"#tokens: {self.total_size()}")
|
||||
|
||||
def total_size(self):
|
||||
return self._total_size_helper(self.root_node)
|
||||
|
||||
def evict(self, num_tokens, evict_callback):
|
||||
if self.disable:
|
||||
raise RuntimeError()
|
||||
|
||||
leaves = self._collect_leaves()
|
||||
heapq.heapify(leaves)
|
||||
|
||||
num_evicted = 0
|
||||
while num_evicted < num_tokens and len(leaves):
|
||||
x = heapq.heappop(leaves)
|
||||
|
||||
if x == self.root_node:
|
||||
break
|
||||
if x.ref_counter > 0:
|
||||
continue
|
||||
|
||||
num_evicted += evict_callback(x.value)
|
||||
self._delete_leaf(x)
|
||||
|
||||
if len(x.parent.children) == 0:
|
||||
heapq.heappush(leaves, x.parent)
|
||||
|
||||
def inc_ref_counter(self, node):
|
||||
delta = 0
|
||||
while node != self.root_node:
|
||||
if node.ref_counter == 0:
|
||||
self.evictable_size_ -= len(node.value)
|
||||
delta -= len(node.value)
|
||||
node.ref_counter += 1
|
||||
node = node.parent
|
||||
return delta
|
||||
|
||||
def dec_ref_counter(self, node):
|
||||
delta = 0
|
||||
while node != self.root_node:
|
||||
if node.ref_counter == 1:
|
||||
self.evictable_size_ += len(node.value)
|
||||
delta += len(node.value)
|
||||
node.ref_counter -= 1
|
||||
node = node.parent
|
||||
return delta
|
||||
|
||||
def evictable_size(self):
|
||||
return self.evictable_size_
|
||||
|
||||
##### Internal Helper Functions #####
|
||||
def _match_prefix_helper(self, node, key, value, last_node):
|
||||
node.last_access_time = time.time()
|
||||
|
||||
for c_key, child in node.children.items():
|
||||
prefix_len = match(c_key, key)
|
||||
if prefix_len != 0:
|
||||
if prefix_len == len(key) and prefix_len != len(c_key):
|
||||
new_node = self._split_node(c_key, child, prefix_len)
|
||||
value.append(new_node.value)
|
||||
last_node[0] = new_node
|
||||
else:
|
||||
value.append(child.value[:prefix_len])
|
||||
last_node[0] = child
|
||||
self._match_prefix_helper(child, key[prefix_len:], value, last_node)
|
||||
break
|
||||
|
||||
def _split_node(self, key, child, split_len):
|
||||
# new_node -> child
|
||||
new_node = TreeNode()
|
||||
new_node.children = {key[split_len:]: child}
|
||||
new_node.parent = child.parent
|
||||
new_node.ref_counter = child.ref_counter
|
||||
new_node.value = child.value[:split_len]
|
||||
child.parent = new_node
|
||||
child.value = child.value[split_len:]
|
||||
new_node.parent.children[key[:split_len]] = new_node
|
||||
del new_node.parent.children[key]
|
||||
return new_node
|
||||
|
||||
def _insert_helper(self, node, key, value):
|
||||
node.last_access_time = time.time()
|
||||
|
||||
for c_key, child in node.children.items():
|
||||
prefix_len = match(c_key, key)
|
||||
|
||||
if prefix_len == len(c_key):
|
||||
if prefix_len == len(key):
|
||||
return prefix_len
|
||||
else:
|
||||
key = key[prefix_len:]
|
||||
value = value[prefix_len:]
|
||||
return prefix_len + self._insert_helper(child, key, value)
|
||||
|
||||
if prefix_len:
|
||||
new_node = self._split_node(c_key, child, prefix_len)
|
||||
return prefix_len + self._insert_helper(
|
||||
new_node, key[prefix_len:], value[prefix_len:]
|
||||
)
|
||||
|
||||
if len(key):
|
||||
new_node = TreeNode()
|
||||
new_node.parent = node
|
||||
new_node.value = value
|
||||
node.children[key] = new_node
|
||||
self.evictable_size_ += len(value)
|
||||
return 0
|
||||
|
||||
def _print_helper(self, node, indent):
|
||||
for key, child in node.children.items():
|
||||
print(" " * indent, len(key), key[:10], f"r={child.ref_counter}")
|
||||
self._print_helper(child, indent=indent + 2)
|
||||
|
||||
def _delete_leaf(self, node):
|
||||
for k, v in node.parent.children.items():
|
||||
if v == node:
|
||||
break
|
||||
del node.parent.children[k]
|
||||
self.evictable_size_ -= len(k)
|
||||
|
||||
def _total_size_helper(self, node):
|
||||
x = len(node.value)
|
||||
for child in node.children.values():
|
||||
x += self._total_size_helper(child)
|
||||
return x
|
||||
|
||||
def _collect_leaves(self):
|
||||
ret_list = []
|
||||
|
||||
def dfs_(cur_node):
|
||||
if len(cur_node.children) == 0:
|
||||
ret_list.append(cur_node)
|
||||
|
||||
for x in cur_node.children.values():
|
||||
dfs_(x)
|
||||
|
||||
dfs_(self.root_node)
|
||||
return ret_list
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
tree = RadixCache(disable=False)
|
||||
|
||||
tree.insert("Hello")
|
||||
tree.insert("Hello")
|
||||
tree.insert("Hello_L.A.!")
|
||||
# tree.insert("Hello_world! Happy")
|
||||
# tree.insert("I love you!")
|
||||
tree.pretty_print()
|
||||
|
||||
# print(tree.match_prefix("I love you! aha"))
|
||||
|
||||
# def evict_callback(x):
|
||||
# print("evict", x)
|
||||
# return len(x)
|
||||
|
||||
# tree.evict(5, evict_callback)
|
||||
# tree.evict(10, evict_callback)
|
||||
# tree.pretty_print()
|
||||
Reference in New Issue
Block a user