관리-도구

편집 파일: history.py

# coding=utf-8
#
# Copyright © Cloud Linux GmbH & Cloud Linux Software, Inc 2010-2023 All Rights Reserved
#
# Licensed under CLOUD LINUX LICENSE AGREEMENT
# http://cloudlinux.com/docs/LICENSE.TXT
import enum
import itertools
from enum import StrEnum
from typing import Self, Iterator, Iterable, NamedTuple, Any, TypeAlias

from lvestats.orm import BurstingEventType

_InTimestamp: TypeAlias = int | float

@enum.unique
class IntervalType(StrEnum):
    OVERUSING = enum.auto()
    NORMAL = enum.auto()

class Interval(NamedTuple):
    start: float
    end: float

@property
    def duration(self) -> float:
        return self.end - self.start

def __str__(self) -> str:
        return _format_pair(round(self.start), round(self.end))

class _Intervals(tuple[Interval, ...]):
    def __str__(self) -> str:
        return _format_joined_intervals(self)

class LveHistory:
    # NOTE(vlebedev): Empty history is special case: `.first_interval_type` == `.ongoing_interval_type`
    __slots__ = ('_first_interval_type', '_timestamps', '_finished_overusing_duration')

def __init__(
        self,
        first_interval_type: IntervalType | None = None,
        timestamps: tuple[float, ...] = tuple(),
        closed_overusing_duration: float | None = None,
    ) -> None:
        if not (
            (first_interval_type, timestamps) == (None, tuple()) or
            first_interval_type is not None and len(timestamps) > 0
        ):
            raise ValueError('Either both `first_interval_type` and `timestamps` must be set or both unset!')

self._first_interval_type = first_interval_type
        self._timestamps = timestamps

if closed_overusing_duration is None:
            closed_overusing_duration = sum(i.duration for i in _overusing_intervals_iter(
                first_interval_type, timestamps,
            )) if first_interval_type is not None else 0.0
        self._finished_overusing_duration = closed_overusing_duration

@property
    def first_interval_type(self) -> IntervalType | None:
        return self._first_interval_type

@property
    def timestamps(self) -> tuple[float, ...]:
        return self._timestamps

@property
    def empty(self) -> bool:
        if len(self._timestamps) == 0:
            assert self._first_interval_type is None
            return True
        return False

@property
    def contains_overusing(self) -> bool:
        ts_len = len(self._timestamps)
        if ts_len == 0:
            return False
        if self._first_interval_type == IntervalType.OVERUSING:
            return True
        if ts_len > 1:
            return True
        return False

@property
    def ongoing_interval_type(self) -> IntervalType | None:
        ts_len = len(self._timestamps)
        if ts_len == 0:
            return None
        assert self._first_interval_type is not None
        return get_interval_type_after(self._first_interval_type, ts_len)

def trim(self, cutoff: _InTimestamp) -> Self:
        cutoff = float(cutoff)
        if len(self._timestamps) == 0 or cutoff <= self._timestamps[0]:
            return self
        cls = type(self)
        if cutoff >= self._timestamps[-1]:
            return cls(self.ongoing_interval_type, (cutoff,))
        assert self._first_interval_type is not None
        trimmed_duration, cutoff_pos = _get_trimmed_overusing_duration_and_position(
            self._first_interval_type,
            self._timestamps,
            cutoff,
        )
        new_first_interval_type = self._first_interval_type
        new_timestamps = self._timestamps[cutoff_pos:]
        if self._timestamps[cutoff_pos] == cutoff:
            switch_new_first_interval = cutoff_pos % 2 != 0
        else:
            switch_new_first_interval = cutoff_pos % 2 == 0
            new_timestamps = (cutoff, *new_timestamps)
        if switch_new_first_interval:
            new_first_interval_type = get_other_interval_type(new_first_interval_type)
        new_overusing_duration = self._finished_overusing_duration - trimmed_duration
        return cls(new_first_interval_type, new_timestamps, new_overusing_duration)

def append(self, timestamp: _InTimestamp, event_type: BurstingEventType) -> Self:
        timestamp = float(timestamp)
        cls = type(self)
        timestamps = self._timestamps

new_interval_type = {
            BurstingEventType.STARTED: IntervalType.OVERUSING,
            BurstingEventType.STOPPED: IntervalType.NORMAL,
        }[event_type]

try:
            if timestamp <= timestamps[-1]:
                raise ValueError('Timestamp must be greater than the latest known one!')
        except IndexError:
            return cls(new_interval_type, (timestamp,))

if len(timestamps) == 0:
            return cls(new_interval_type, (timestamp,))

if self.ongoing_interval_type == new_interval_type:
            # NOTE(vlebedev): Ongoing interval has not changed so history is not altered.
            return self

new_overusing_duration = self._finished_overusing_duration
        if len(timestamps) > 0 and new_interval_type == IntervalType.NORMAL:
            # NOTE(vlebedev): Previous interval was of overusing type so add it to finished overusing duration.
            new_overusing_duration += timestamp - timestamps[-1]
        return cls(self._first_interval_type, (*timestamps, timestamp), new_overusing_duration)

def get_overusing_duration(self, now: int | float) -> float:
        duration = self._finished_overusing_duration
        if len(self._timestamps) > 0 and self.ongoing_interval_type == IntervalType.OVERUSING:
            # NOTE(vlebedev): Overusing interval is still ongoing so add its current duration to the total one.
            duration += float(now) - self._timestamps[-1]
        return duration

def get_overusing_intervals(self, now: int | float) -> _Intervals:
        now = float(now)
        return _Intervals(self.get_intervals_iter(now, IntervalType.OVERUSING))

def get_intervals(self, now: int | float, intervals_type: IntervalType) -> _Intervals:
        return _Intervals(self.get_intervals_iter(now, intervals_type))

def get_intervals_iter(
        self,
        now: int | float,
        intervals_type: IntervalType = IntervalType.OVERUSING,
    ) -> Iterator[Interval]:
        now = float(now)
        if len(self._timestamps) > 0 and now < self._timestamps[0]:
            raise ValueError('Final timestamp must be greater than the latest known one!')
        for start, end in self._pairs_iter(intervals_type, now):
            yield Interval(start, end)

def _pairs_iter(self, intervals_type: IntervalType, final_item=None) -> Iterator[tuple[Any, Any]]:
        if len(self._timestamps) == 0:
            return
        assert self._first_interval_type is not None
        for start, type_, end in _typed_pairs_iter(
            self._first_interval_type,
            self._timestamps,
            final_item,
        ):
            if type_ != intervals_type:
                continue
            yield start, end

def __eq__(self, other: object) -> bool:
        cls = type(self)

if not isinstance(other, cls):
            return False

if (self._first_interval_type, self._timestamps) != (other._first_interval_type, other._timestamps):
            return False

assert self._finished_overusing_duration == other._finished_overusing_duration

return True

def __repr__(self) -> str:
        if self.empty:
            return f'{type(self).__name__}()'
        return f'{type(self).__name__}({self._first_interval_type.name}, {self._timestamps!r})'

def __str__(self) -> str:
        return _format_joined_intervals(
            _format_pair(*p)
            for p in self._pairs_iter(IntervalType.OVERUSING, "-")
        )

def get_interval_type_after(starts_with: IntervalType, ts_num: int) -> IntervalType:
    if ts_num in {0, 1} or ts_num % 2 != 0:
        return starts_with
    return get_other_interval_type(starts_with)

def get_other_interval_type(current_event_type: IntervalType) -> IntervalType:
    return IntervalType.OVERUSING if current_event_type == IntervalType.NORMAL else IntervalType.NORMAL

class _TrimmedDurationAndPosition(NamedTuple):
    duration: float
    position: int

def _get_trimmed_overusing_duration_and_position(
    first_interval_type: IntervalType,
    timestamps: Iterable[float],
    cutoff: float,
) -> _TrimmedDurationAndPosition:
    position = 0

def cutted_timestamps_iter():
        nonlocal position

for position, ts in enumerate(timestamps):
            ts = min(ts, cutoff)
            yield ts
            if ts >= cutoff:
                break

duration = sum(i.duration for i in _overusing_intervals_iter(first_interval_type, cutted_timestamps_iter()))

return _TrimmedDurationAndPosition(duration, position)

def _overusing_intervals_iter(
    first_interval_type: IntervalType,
    timestamps: Iterable[float],
) -> Iterable[Interval]:
    for start, interval_type, end in _typed_pairs_iter(first_interval_type, timestamps):
        if interval_type != IntervalType.OVERUSING:
            continue
        yield Interval(start, end)

def _typed_pairs_iter(
    first_interval_type: IntervalType,
    timestamps: Iterable[Any],
    final_item: Any | None = None,
) -> Iterable[tuple[Any, IntervalType, Any]]:
    if final_item is not None:
        timestamps = itertools.chain(timestamps, (final_item,))
    intervals_iter = itertools.pairwise(timestamps)
    intervals_types = itertools.cycle([
        first_interval_type,
        get_other_interval_type(first_interval_type),
    ])
    for (start, end), type_ in zip(intervals_iter, intervals_types):
        yield start, type_, end

def _format_pair(formatted_start, formatted_stop) -> str:
    return f'({formatted_start}-{formatted_stop})'

def _format_joined_intervals(formatted_intervals: Iterable) -> str:
    return 'U'.join(str(i) for i in formatted_intervals) or "()"