diff --git a/src/part1/constants.py b/src/part1/constants.py index 9abfbba2..960e4b82 100644 --- a/src/part1/constants.py +++ b/src/part1/constants.py @@ -1,38 +1,10 @@ -"""Bit layout and default epoch for the Twitter Snowflake ID generator. - -A Snowflake identifier is a 63-bit positive integer packed as follows, from -the most significant bit downwards: - - 1 bit unused sign bit, always zero - 41 bits milliseconds elapsed since a custom epoch - 10 bits node identifier - 12 bits per-millisecond sequence counter - -The shift and mask constants should be derived from the bit widths so that -changing a width keeps everything else consistent. -""" - -from typing import Final - - -# TODO: замените заглушки (0) на корректные значения, см. TASK.md. -# Все константы, зависящие от ёмкостей *_BITS, должны вычисляться из них. - -# Twitter's original Snowflake epoch: 2010-11-04 01:42:54.657 UTC. -EPOCH_MS_DEFAULT: Final[int] = 0 - -# Node identifier used when the caller does not supply one. -NODE_ID_DEFAULT: Final[int] = 0 - -# Width of each field, in bits. -TIMESTAMP_BITS: Final[int] = 0 -NODE_ID_BITS: Final[int] = 0 -SEQUENCE_ID_BITS: Final[int] = 0 - -# Largest value each field can hold. -TIMESTAMP_MS_MAX: Final[int] = 0 -NODE_ID_MAX: Final[int] = 0 -SEQUENCE_ID_MAX: Final[int] = 0 - -# Здесь можно добавить собственные вспомогательные константы -# (например, сдвиги полей при сборке идентификатора). +EPOCH_MS_DEFAULT: int = 1288834974657 +NODE_ID_DEFAULT: int = 1 +TIMESTAMP_BITS: int = 41 +NODE_ID_BITS: int = 10 +SEQUENCE_ID_BITS: int = 12 +TIMESTAMP_MS_MAX: int = 2**TIMESTAMP_BITS - 1 +NODE_ID_MAX: int = 2**NODE_ID_BITS - 1 +SEQUENCE_ID_MAX: int = 2**SEQUENCE_ID_BITS - 1 +NODE_ID_SHIFT: int = SEQUENCE_ID_BITS +TIMESTAMP_SHIFT: int = SEQUENCE_ID_BITS + NODE_ID_BITS diff --git a/src/part1/snowflake.py b/src/part1/snowflake.py index d3abf050..ae1e7ee5 100644 --- a/src/part1/snowflake.py +++ b/src/part1/snowflake.py @@ -1,114 +1,48 @@ -"""Twitter Snowflake identifier generator. +import time -Snowflake produces roughly time-ordered 64-bit integer identifiers without -coordination between nodes. Each identifier packs the milliseconds elapsed -since a custom epoch, a node identifier and a per-millisecond sequence counter -into a single 63-bit positive integer (see ``constants`` for the layout). - -The public entry point is :func:`generate_snowflake_id`. It is stateless: the -caller passes the sequence counter on every call and is responsible for -advancing it within a millisecond and resetting it when the clock ticks over. - -Each packed field can be read back on its own with :func:`decode_timestamp_ms`, -:func:`decode_node_id` and :func:`decode_sequence_id`. -""" - -import time # noqa: F401 - -from .constants import ( # noqa: F401 +from .constants import ( EPOCH_MS_DEFAULT, NODE_ID_DEFAULT, NODE_ID_MAX, + NODE_ID_SHIFT, SEQUENCE_ID_MAX, TIMESTAMP_MS_MAX, + TIMESTAMP_SHIFT, ) def read_current_millis(epoch_ms: int) -> int: - """Read the number of milliseconds elapsed since the custom epoch. - - Args: - epoch_ms: The custom epoch expressed as Unix milliseconds. - - Returns: - The count of whole milliseconds between ``epoch_ms`` and now. May be - negative if ``epoch_ms`` lies in the future. - """ - # TODO: реализуйте функцию - return 0 + return int(time.time() * 1000) - epoch_ms def decode_timestamp_ms(snowflake_id: int, epoch_ms: int = EPOCH_MS_DEFAULT) -> int: - """Read the timestamp field out of a Snowflake identifier. - - Args: - snowflake_id: An identifier produced against the same epoch. - epoch_ms: The epoch the identifier was generated against. Defaults to - the original Twitter epoch (2010-11-04 01:42:54.657 UTC). - - Returns: - The absolute Unix time in milliseconds at which the identifier was - generated. - """ - # TODO: реализуйте функцию - return 0 + return snowflake_id // (2**TIMESTAMP_SHIFT) + epoch_ms -def decode_node_id(snowflake_id: int) -> int: - """Read the node identifier field out of a Snowflake identifier. - - Args: - snowflake_id: An identifier produced by :func:`generate_snowflake_id`. - - Returns: - The node identifier packed into ``snowflake_id``, in the range - ``[0, NODE_ID_MAX]``. - """ - # TODO: реализуйте функцию - return 0 +def decode_node_id(snowflake_id: int, epoch_ms: int = EPOCH_MS_DEFAULT) -> int: + return ( + snowflake_id + - ( + (decode_timestamp_ms(snowflake_id, epoch_ms) - epoch_ms) + * 2**TIMESTAMP_SHIFT + ) + ) // (2**NODE_ID_SHIFT) def decode_sequence_id(snowflake_id: int) -> int: - """Read the sequence counter field out of a Snowflake identifier. - - Args: - snowflake_id: An identifier produced by :func:`generate_snowflake_id`. - - Returns: - The per-millisecond sequence counter packed into ``snowflake_id``, in - the range ``[0, SEQUENCE_ID_MAX]``. - """ - # TODO: реализуйте функцию - return 0 + return snowflake_id - ((snowflake_id // 2**NODE_ID_SHIFT) * (2**NODE_ID_SHIFT)) def generate_snowflake_id( - sequence_id: int, - node_id: int = NODE_ID_DEFAULT, - epoch_ms: int = EPOCH_MS_DEFAULT, + sequence_id: int, node_id: int = NODE_ID_DEFAULT, epoch_ms: int = EPOCH_MS_DEFAULT ) -> int | None: - """Build and return a Snowflake identifier for the current millisecond. - - The function is stateless: the caller passes the per-millisecond sequence - counter explicitly. It is the caller's responsibility to increment - ``sequence_id`` for identifiers minted within the same millisecond and to - reset it once the clock advances. - - Args: - sequence_id: The per-millisecond sequence counter, in the range - ``[0, SEQUENCE_ID_MAX]``. - node_id: The identifier of this node, in the range ``[0, NODE_ID_MAX]``. - Optional; defaults to ``NODE_ID_DEFAULT``. - epoch_ms: The start of the epoch as Unix milliseconds. Optional; - defaults to the original Twitter epoch (2010-11-04 01:42:54.657 - UTC). - - Returns: - The Snowflake identifier as a positive 63-bit integer, or ``None`` if - ``node_id`` is outside ``[0, NODE_ID_MAX]``, ``sequence_id`` is outside - ``[0, SEQUENCE_ID_MAX]``, or the elapsed time no longer fits in the - timestamp field (roughly 69 years after ``epoch_ms``). In each of those - cases an explanatory message is printed to stdout first. - """ - # TODO: реализуйте функцию - return 0 + current_millis = read_current_millis(epoch_ms) + if current_millis < 0 or current_millis > TIMESTAMP_MS_MAX: + raise ValueError("Превышен максимальный лимит времени") + if node_id < 0 or node_id > NODE_ID_MAX: + raise ValueError("Node Id выходит за допустимые пределы") + if sequence_id < 0 or sequence_id > SEQUENCE_ID_MAX: + raise ValueError("Sequence Id выходит за допустимые пределы") + return (current_millis * (2**TIMESTAMP_SHIFT)) + ( + node_id * (2**NODE_ID_SHIFT) + sequence_id + )