bt-decode

A python wrapper around the rust scale-codec crate for fast scale-decoding of Bittensor data structures.

Usage

DelegateInfo

get_delegates

import bittensor
from bt_decode import DelegateInfo

# Setup subtensor connection
subtensor = bittensor.subtensor()
# Grab result from RuntimeAPI
hex_bytes_result = sub.query_runtime_api(
    runtime_api="DelegateInfoRuntimeApi",
    method="get_delegates",
    params=[ ]
)
# Decode scale-encoded Vec<DelegateInfo>
delegates_info: List[DelegateInfo] = DelegateInfo.decode_vec(
    bytes.fromhex(
        hex_bytes_result
))

get_delegated

import bittensor
from bt_decode import DelegateInfo

validator_key = bittensor.Keypair(ss58_address="5E9fVY1jexCNVMjd2rdBsAxeamFGEMfzHcyTn2fHgdHeYc5p")

# Setup subtensor connection
subtensor = bittensor.subtensor()
# Grab result from RuntimeAPI
hex_bytes_result = sub.query_runtime_api(
    runtime_api="DelegateInfoRuntimeApi",
    method="get_delegated",
    params=[list( validator_key.public_key )]
)
# Decode scale-encoded Vec<(DelegateInfo, take)>
delegated_info: List[Tuple[DelegateInfo, int]] = DelegateInfo.decode_delegated(
    bytes.fromhex(
        hex_bytes_result
))

NeuronInfo

get_neuron

import bittensor
from bt_decode import NeuronInfo

# Setup subtensor connection
subtensor = bittensor.subtensor()
NETUID = 1
UID = 0
# Grab result from RuntimeAPI
hex_bytes_result = sub.query_runtime_api(
    runtime_api="NeuronInfoRuntimeApi",
    method="get_neuron",
    params=[NETUID, UID]
)
# Decode scale-encoded NeuronInfo
neuron: NeuronInfo = NeuronInfo.decode(
    bytes.fromhex(
        hex_bytes_result
))

get_neurons

import bittensor
from bt_decode import NeuronInfo

# Setup subtensor connection
subtensor = bittensor.subtensor()
NETUID = 1
# Grab result from RuntimeAPI
hex_bytes_result = sub.query_runtime_api(
    runtime_api="NeuronInfoRuntimeApi",
    method="get_neurons",
    params=[NETUID]
)
# Decode scale-encoded Vec<NeuronInfo>
neurons: List[NeuronInfo] = NeuronInfo.decode_vec(
    bytes.fromhex(
        hex_bytes_result
))

NeuronInfoLite

get_neuron_lite

import bittensor
from bt_decode import NeuronInfoLite

# Setup subtensor connection
subtensor = bittensor.subtensor()
NETUID = 1
UID = 0
# Grab result from RuntimeAPI
hex_bytes_result = sub.query_runtime_api(
    runtime_api="NeuronInfoRuntimeApi",
    method="get_neuron_lite",
    params=[NETUID, UID]
)
# Decode scale-encoded NeuronInfoLite
neuron_lite: NeuronInfoLite = NeuronInfoLite.decode(
    bytes.fromhex(
        hex_bytes_result
))

get_neurons_lite

import bittensor
from bt_decode import NeuronInfoLite

# Setup subtensor connection
subtensor = bittensor.subtensor()
NETUID = 1
# Grab result from RuntimeAPI
hex_bytes_result = sub.query_runtime_api(
    runtime_api="NeuronInfoRuntimeApi",
    method="get_neurons_lite",
    params=[NETUID]
)
# Decode scale-encoded Vec<NeuronInfoLite>
neurons_lite: List[NeuronInfoLite] = NeuronInfoLite.decode_vec(
    bytes.fromhex(
        hex_bytes_result
))

StakeInfo

get_stake_info_for_coldkey

import bittensor
from bt_decode import StakeInfo

validator_key = bittensor.Keypair(ss58_address="5HBtpwxuGNL1gwzwomwR7sjwUt8WXYSuWcLYN6f9KpTZkP4k")

# Setup subtensor connection
subtensor = bittensor.subtensor()
encoded_coldkey = list( validator_key.public_key )
# Grab result from RuntimeAPI
hex_bytes_result = sub.query_runtime_api(
    runtime_api="StakeInfoRuntimeApi",
    method="get_stake_info_for_coldkey",
    params=[encoded_coldkey]
)
# Decode scale-encoded StakeInfo
stake_info: List[StakeInfo] = StakeInfo.decode_vec(
    bytes.fromhex(
        hex_bytes_result
))

get_stake_info_for_coldkeys

import bittensor
from bt_decode import StakeInfo

validator_key_0 = bittensor.Keypair(ss58_address="5GcCZ2BPXBjgG88tXJCEtkbdg2hNrPbL4EFfbiVRvBZdSQDC")
validator_key_1 = bittensor.Keypair(ss58_address="5HBtpwxuGNL1gwzwomwR7sjwUt8WXYSuWcLYN6f9KpTZkP4k")

encoded_coldkeys = [
    list( validator_key_0.public_key ),
    list( validator_key_1.public_key )
]

# Setup subtensor connection
subtensor = bittensor.subtensor()
# Grab result from RuntimeAPI
hex_bytes_result = sub.query_runtime_api(
    runtime_api="StakeInfoRuntimeApi",
    method="get_stake_info_for_coldkeys",
    params=[encoded_coldkeys]
)
# Decode scale-encoded Vec<(AccountId, StakeInfo)>
stake_info: List[Tuple[bytes, List["StakeInfo"]]] = StakeInfo.decode_vec_tuple_vec(
    bytes.fromhex(
        hex_bytes_result
))

SubnetInfo

get_subnet_info

import bittensor
from bt_decode import SubnetInfo

# Setup subtensor connection
subtensor = bittensor.subtensor()
NETUID = 1
# Grab result from RuntimeAPI
hex_bytes_result = sub.query_runtime_api(
    runtime_api="SubnetInfoRuntimeApi",
    method="get_subnet_info",
    params=[NETUID]
)
# Decode scale-encoded Option<SubnetInfo>
subnet_info: SubnetInfo = SubnetInfo.decode_option(
    bytes.fromhex(
        hex_bytes_result
))

get_subnets_info

import bittensor
from bt_decode import SubnetInfo

# Setup subtensor connection
subtensor = bittensor.subtensor()
# Grab result from RuntimeAPI
hex_bytes_result = sub.query_runtime_api(
    runtime_api="SubnetInfoRuntimeApi",
    method="get_subnets_info",
    params=[ ]
)
# Decode scale-encoded Vec<Option<SubnetInfo>>
subnets_info: List[Optional[SubnetInfo]] = SubnetInfo.decode_vec_option(
    bytes.fromhex(
        hex_bytes_result
))

SubnetHyperparameters

get_subnet_hyperparams

import bittensor
from bt_decode import SubnetHyperparameters

# Setup subtensor connection
subtensor = bittensor.subtensor()
NETUID = 1
# Grab result from RuntimeAPI
hex_bytes_result = sub.query_runtime_api(
    runtime_api="SubnetInfoRuntimeApi",
    method="get_subnet_hyperparams",
    params=[NETUID]
)
# Decode scale-encoded Option<SubnetHyperparameters>
subnet_hyper_params: Optional[SubnetHyperparameters] = SubnetHyperparameters.decode_option(
    bytes.fromhex(
        hex_bytes_result
))

decode by type string

Note: This feature is unstable, but working for multiple types.

You may also decode using a type-string formed from existing types by passing the metadata as pulled from a node (or formed manually).

import bittensor, bt_decode, scalecodec
# Get subtensor connection
sub = bittensor.subtensor()
# Create a param for the RPC call, using v15 metadata
v15_int = scalecodec.U32()
v15_int.value = 15
# Make the RPC call to grab the metadata
metadata_rpc_result = sub.substrate.rpc_request("state_call", [
    "Metadata_metadata_at_version",
    v15_int.encode().to_hex(),
    sub.substrate.get_chain_finalised_head()
])
# Decode the metadata into a PortableRegistry type
metadata_option_hex_str = metadata_rpc_result['result']
metadata_option_bytes = bytes.fromhex(metadata_option_hex_str[2:])
metadata_v15 = bt_decode.MetadataV15.decode_from_metadata_option(metadata_option_bytes)
registry = bt_decode.PortableRegistry.from_metadata_v15( metadata_v15 )

# Decode by type-string
NETUID = 1
## Grab result from RuntimeAPI
hex_bytes_result = sub.query_runtime_api(
    runtime_api="NeuronInfoRuntimeApi",
    method="get_neurons_lite",
    params=[NETUID]
)
## Decode scale-encoded NeuronInfoLite by type-string
neurons_lite: List[NeuronInfoLite] = bt_decode.decode(
    "Vec<NeuronInfoLite>", # type-string
    registry, # registry as above
    bytes.fromhex(
        hex_bytes_result # bytes to decode
    )
)

encode by type string

Note: This feature is unstable, but working for multiple types.

You may also encode using a type-string formed from existing types by passing the metadata as pulled from a node (or formed manually).

import bittensor, bt_decode, scalecodec
# Get subtensor connection
sub = bittensor.subtensor()
# Create a param for the RPC call, using v15 metadata
v15_int = scalecodec.U32()
v15_int.value = 15
# Make the RPC call to grab the metadata
metadata_rpc_result = sub.substrate.rpc_request("state_call", [
    "Metadata_metadata_at_version",
    v15_int.encode().to_hex(),
    sub.substrate.get_chain_finalised_head()
])
# Decode the metadata into a PortableRegistry type
metadata_option_hex_str = metadata_rpc_result['result']
metadata_option_bytes = bytes.fromhex(metadata_option_hex_str[2:])
metadata_v15 = bt_decode.MetadataV15.decode_from_metadata_option(metadata_option_bytes)
registry = bt_decode.PortableRegistry.from_metadata_v15( metadata_v15 )


## Encode an integer as a compact u16
compact_u16: list[int] = bt_decode.encode(
    "Compact<u16>", # type-string,
    2**16-1,
    registry
)
# [254, 255, 3, 0]
compact_u16_py_scale_codec = scalecodec.Compact()
compact_u16_py_scale_codec.value = 2**16-1
compact_u16_py_scale_codec.encode()

assert list(compact_u16_py_scale_codec.data.data) == compact_u16