Minting
This is the first of many tutorials in a series where you'll be creating a complete NFT smart contract from scratch that conforms with all the NEAR NFT standards.
Today you'll learn how to create the logic needed to mint NFTs and have them show up in your NEAR wallet. You will be filling a bare-bones skeleton smart contract to add minting functionalities.
You can find the skeleton contract in our Skeleton folder
A completed version of this tutorial can be found in the Basic NFT folder
Introduction
To get started, go to the nft-contract-skeleton
folder in our repo. If you haven't cloned the repository, refer to the Contract Architecture to get started.
cd nft-contract-skeleton/
If you wish to see the finished code of this step-by-step basic NFT contract tutorial, that can be found on the nft-contract-basic
folder.
Modifications to the skeleton contract
In order to implement the logic needed for minting, we should break it up into smaller tasks and handle those one-by-one. Let's step back and think about the best way to do this by asking ourselves a simple question: what does it mean to mint an NFT?
To mint a non-fungible token, in the most simple way possible, a contract needs to be able to associate a token with an owner on the blockchain. This means you'll need:
- A way to keep track of tokens and other information on the contract.
- A way to store information for each token such as
metadata
(more on that later). - A way to link a token with an owner.
That's it! We've now broken down the larger problem into some smaller, less daunting, subtasks. Let's start by tackling the first and work our way through the rest.
Storing information on the contract
Start by navigating to nft-contract-skeleton/src/lib.rs
and filling in some of the code blocks.
You need to be able to store important information on the contract such as the list of tokens that an account has.
Contract Struct
The first thing to do is modifying the contract struct
as follows:
Loading...
This allows you to get the information stored in these data structures from anywhere in the contract. The code above has created 3 token specific storages:
- tokens_per_owner: allows you to keep track of the tokens owned by any account
- tokens_by_id: returns all the information about a specific token
- token_metadata_by_id: returns just the metadata for a specific token
In addition, you'll keep track of the owner of the contract as well as the metadata for the contract.
You might be confused as to some of the types that are being used. In order to make the code more readable, we've introduced custom data types which we'll briefly outline below:
- AccountId: a string that ensures there are no special or unsupported characters.
- TokenId: simply a string.
As for the Token
, TokenMetadata
, and NFTContractMetadata
data types, those are structs that we'll define later in this tutorial.
Initialization Functions
Next, create what's called an initialization function; we will name it new
, but you can choose any name you prefer.
This function needs to be invoked when you first deploy the contract. It will initialize all the contract's fields that you've defined above with default values.
Don't forget to add the owner_id
and metadata
fields as parameters to the function, so only those can be customized.
This function will default all the collections to be empty and set the owner
and metadata
equal to what you pass in.
Loading...
More often than not when doing development, you'll need to deploy contracts several times. You can imagine that it might get tedious to have to pass in metadata every single time you want to initialize the contract. For this reason, let's create a function that can initialize the contract with a set of default metadata
. You can call it new_default_meta
and it'll only take the owner_id
as a parameter.
Loading...
This function is simply calling the previous new
function and passing in the owner that you specify and also passes in some default metadata.
Metadata and token information
Now that you've defined what information to store on the contract itself and you've defined some ways to initialize the contract, you need to define what information should go in the Token
, TokenMetadata
, and NFTContractMetadata
data types.
Let's switch over to the nft-contract-skeleton/src/metadata.rs
file as this is where that information will go.
If you look at the standards for metadata, you'll find all the necessary information that you need to store for both TokenMetadata
and NFTContractMetadata
. Simply fill in the following code.
Loading...
This now leaves you with the Token
struct and something called a JsonToken
. The Token
struct will hold all the information directly related to the token excluding the metadata. The metadata, if you remember, is stored in a map on the contract in a data structure called token_metadata_by_id
. This allows you to quickly get the metadata for any token by simply passing in the token's ID.
For the Token
struct, you'll just keep track of the owner for now.
Loading...
Since NEAR smart contracts receive and return data in JSON format, the purpose of the JsonToken
is to act as output when the user asks information for an NFT. This means you'll want to store the owner, token ID, and metadata.
Loading...
Some of you might be thinking "how come we don't just store all the information in the Token
struct?".
The reason behind this is that it's actually more efficient to construct the JSON token on the fly only when you need it rather than storing all the information in the token struct.
In addition, some operations might only need the metadata for a token and so having the metadata in a separate data structure is more optimal.
Function for querying contract metadata
Now that you've defined some of the types that were used in the previous section, let's move on and create the first view function nft_metadata
. This will allow users to query for the contract's metadata as per the metadata standard.
Loading...
This function will get the metadata
object from the contract which is of type NFTContractMetadata
and will return it.
Just like that, you've completed the first two tasks and are ready to move onto last part of the tutorial.
Minting Logic
Now that all the information and types are defined, let's start brainstorming how the minting logic will play out. In the end, you need to link a Token
and TokenId
to a specific owner. Let's look back at the lib.rs
file to see how you can accomplish this. There are a couple data structures that might be useful:
//keeps track of all the token IDs for a given account
pub tokens_per_owner: LookupMap<AccountId, UnorderedSet<TokenId>>,
//keeps track of the token struct for a given token ID
pub tokens_by_id: LookupMap<TokenId, Token>,
//keeps track of the token metadata for a given token ID
pub token_metadata_by_id: UnorderedMap<TokenId, TokenMetadata>,
Looking at these data structures, you could do the following:
- Add the token ID into the set of tokens that the receiver owns. This will be done on the
tokens_per_owner
field. - Create a token object and map the token ID to that token object in the
tokens_by_id
field. - Map the token ID to it's metadata using the
token_metadata_by_id
.
Storage Implications
With those steps outlined, it's important to take into consideration the storage costs of minting NFTs. Since you're adding bytes to the contract by creating entries in the data structures, the contract needs to cover the storage costs. If you just made it so any user could go and mint an NFT for free, that system could easily be abused and users could essentially "drain" the contract of all it's funds by minting thousands of NFTs. For this reason, you'll make it so that users need to attach a deposit to the call to cover the cost of storage. You'll measure the initial storage usage before anything was added and you'll measure the final storage usage after all the logic is finished. Then you'll make sure that the user has attached enough $NEAR to cover that cost and refund them if they've attached too much.
This is how we do it in code:
Loading...
You'll notice that we're using some internal methods such as refund_deposit
and internal_add_token_to_owner
. We've described the function of refund_deposit
and as for internal_add_token_to_owner
, this will add a token to the set of tokens an account owns for the contract's tokens_per_owner
data structure. You can create these functions in a file called internal.rs
. Go ahead and create the file. Your new contract architecture should look as follows:
nft-contract
├── Cargo.lock
├── Cargo.toml
├── README.md
├── build.sh
└── src
├── approval.rs
├── enumeration.rs
├── internal.rs
├── lib.rs
├── metadata.rs
├── mint.rs
├── nft_core.rs
├── events.rs
└── royalty.rs
Add the following to your newly created internal.rs
file.
Loading...
You may notice more functions in the internal.rs
file than we need for now. You may ignore them, we'll come back to them later.
Let's now quickly move to the lib.rs
file and make the functions we just created invokable in other files. We'll add the internal crates and mod the file as shown below:
Loading...
At this point, the core logic is all in place so that you can mint NFTs. You can use the function nft_mint
which takes the following parameters:
- token_id: the ID of the token you're minting (as a string).
- metadata: the metadata for the token that you're minting (of type
TokenMetadata
which is found in themetadata.rs
file). - receiver_id: specifies who the owner of the token will be.
Behind the scenes, the function will:
- Calculate the initial storage before adding anything to the contract
- Create a
Token
object with the owner ID - Link the token ID to the newly created token object by inserting them into the
tokens_by_id
field. - Link the token ID to the passed in metadata by inserting them into the
token_metadata_by_id
field. - Add the token ID to the list of tokens that the owner owns by calling the
internal_add_token_to_owner
function. - Calculate the final and net storage to make sure that the user has attached enough NEAR to the call in order to cover those costs.
Querying for token information
If you were to go ahead and deploy this contract, initialize it, and mint an NFT, you would have no way of knowing or querying for the information about the token you just minted. Let's quickly add a way to query for the information of a specific NFT. You'll move to the nft-contract-skeleton/src/nft_core.rs
file and edit the nft_token
function.
It will take a token ID as a parameter and return the information for that token. The JsonToken
contains the token ID, the owner ID, and the token's metadata.
Loading...