Account (Base type) #
By deriving from the base Account
type, you can make your own program accounts.
Usage:
class MyAccount(Account):
data: u64
@instruction
def use_my_account(my_account: MyAccount):
# Gets the pubkey from this account.
key = my_account.key()
# Sets the value of `data` in this account.
my_account.data = 1
Right now, accounts (and other classes in Seahorse) can only define their underlying data type by using type-annotated fields with no default value. You’re also allowed to define enums to use in your accounts, like so:
class MyAccount(Account):
enum: MyEnum
data: u64
class MyEnum(Enum):
# Enums are the only place where you need to use this syntax, which should be
# familiar if you've used Python enums before. The numbers don't actually do
# anything here, they're just there for parsing.
OPTION_ONE = 1
OPTION_TWO = 2
OPTION_N = 3
Built-in account types #
Signer #
Wallet that signed the transaction with this instruction call. Often used as an account payer or seed.
Usage:
@instruction
def use_signer(signer: Signer):
# Gets the pubkey of the signer (as a Pubkey type).
key = signer.key()
Empty #
Account that will be initialized by this instruction. These accounts also save the bump
used to create them, if seeds were provided in their initialization. If you create an account without seeds and try to access the bump
, your program will error at runtime!
Seeds may be strings, integer numbers, lists of bytes, or other accounts. Importantly, seeds must either be literals (like a quoted string or a raw number) or instruction parameters. In order to become a seed, each of these types must be converted to bytes. Strings are converted into the bytes of their UTF-8 representation, integers are converted into their little-endian representation, and accounts are converted into their key.
Usage:
@instruction
def use_empty(empty: Empty[MyAccount], signer: Signer):
# Initializes the empty account. `signer` will pay for the cost of creating
# the account (rent).
# Because the seeds uniquely identify this account among all other accounts
# created by this program, if the same signer calls this instruction again,
# it will fail due to trying to re-initialize an existing Solana account.
my_account = empty.init(
payer = signer,
seeds = ['MyAccount', signer]
)
# Retrieve the bump seed used to create the PDA.
bump = empty.bump()
Padding and space #
Since accounts in Solana need to have a static size, it’s difficult to know how to store data with a variable size (like a string). Seahorse makes this easy by providing two extra options when initializing an account: padding
and space
.
Initializing an account with padding
gives you extra bytes to work with, which are used to store any heap-stored objects like the data of a string. You can use it like this:
@instruction
def use_empty(empty: Empty[MyAccount], signer: Signer):
my_account = empty.init(
payer = signer,
seeds = ['MyAccount', signer],
# 64 bytes of padding, letting you store a string up to 64 chars long!
padding = 64
)
Padding adds to the necessary size of the account’s struct, meaning the underlying struct that Seahorse generates and stores on-chain. For example, a Rust u64
takes up 8 bytes and a String
takes up 24 bytes. If you have an account with a u64
, a String
, and padding = 32
, then the entire account will be 8 + 24 + 32 = 64 bytes large.
If you want to get the size of a string, you should use Seahorse’s prelude size
function (padding = size(string)
) to get the size of the string’s data in bytes.
Alternatively, you can provide an absolute amount for the total account size with space
:
@instruction
def use_empty(empty: Empty[MyAccount], signer: Signer):
my_account = empty.init(
payer = signer,
seeds = ['MyAccount', signer],
# 512 bytes of account space total - you can store a string up to
# (512 - n) chars long, if n is the size of the account struct
space = 512
)
padding
and space
are mutually exclusive, specifying both will cause an error.
To summarize: the size of an account in Seahorse is either the size of its underlying Rust struct + padding
, or just space
. If you don’t want to have to figure out the size of your account’s struct, you should use padding
.
UncheckedAccount #
Maps directly to an Anchor UncheckedAccount
. An account that goes through no checks to determine its type/owning program.
Usage:
@instruction
def use_unchecked(unchecked: UncheckedAccount):
# Pretty much all you can do with these is get the key and use them in CPIs.
key = unchecked.key()
Program #
Account that allows you to invoke CPI calls! More on this in Other CPIs.
Usage:
@instruction
def use_program(program: Program):
accounts = ...
data = ...
program.invoke(
accounts,
data
)
Clock #
Solana’s Clock
sysvar. The API is nearly identical to what you would expect from using the clock in a Rust program.
Usage:
@instruction
def use_clock(clock: Clock):
slot: u64 = clock.slot()
epoch: u64 = clock.epoch()
time: i64 = clock.unix_timestamp()
Finding PDAs #
Program-derived addresses can be found using the Pubkey.find_program_address()
function, just like in Solana for Rust.
If you don’t supply an argument for the second parameter program
, then the current program’s address (as defined by declare_id
, see ) will be used.
# Finds a PDA owned by this program with the seeds ['seed', signer.key()].
pda1 = Pubkey.find_program_address(['seed', signer])
# Finds a PDA with the same seeds as above, but owned by `prog`.
pda2 = Pubkey.find_program_address(['seed', signer], prog.key())