Build a blog

In this tutorial, you create a blockchain with a module that lets you write to and read data from the blockchain. This module implements create and read functionalities for a blog-like application. The end user will be able to submit new blog posts and show a list of blog posts on the blockchain.

The purpose of this tutorial is to guide you through the implementation of a complete feedback loop: submitting data and reading this data back from the blockchain.

By completing this tutorial, you will learn about:

• Scaffolding a Cosmos SDK message
• Defining new types in protocol buffer files
• Implementing keeper methods to write data to the store
• Reading data from the store and return it as a result of a query
• Using the blockchain's CLI to broadcast transactions and query the blockchain

Note: All the functions in this chapter can be scaffolded with a single command but instead you will learn how to add each functionality individually.

Prerequisites

This series of blog tutorials is based on a specific version of Ignite CLI, so to install Ignite CLI v0.22.2 use the following command:

curl https://get.ignite.com/cli@v0.22.2! | bash

Create your blog chain

First, create a new blockchain.

Open a terminal and navigate to a directory where you have permissions to create files. To create your Cosmos SDK blockchain, run this command:

ignite scaffold chain blog --address-prefix blog

The blog directory is created with the default directory structure.

The new blockchain is scaffolded with the --address-prefix blog flag to use "blog" instead of the default "cosmos" address prefix.

High-level transaction review

So far, you have learned how to modify proto files to define a new API endpoint and modify a keeper query function to return static data back to the user. Of course, a keeper can do more than return a string of data. Its purpose is to manage access to the state of the blockchain.

You can think of the state as being a collection of key-value stores. Each module is responsible for its own store. Changes to the store are triggered by transactions that are signed and broadcasted by users. Each transaction contains Cosmos SDK messages (not to be confused with proto message). When a transaction is processed, each message gets routed to its module. A module has message handlers that process messages. Processing a message can trigger changes in the state.

Create message types

A Cosmos SDK message contains information that can trigger changes in the state of a blockchain.

First, change into the blog directory:

cd blog

To create a message type and its handler, use the message command:

ignite scaffold message createPost title body

The message command accepts message name (createPost) and a list of fields (title and body) as arguments.

The message command has created and modified several files:

modify proto/blog/tx.proto
modify x/blog/client/cli/tx.go
create x/blog/client/cli/tx_create_post.go
create x/blog/keeper/msg_server_create_post.go
modify x/blog/module_simulation.go
create x/blog/simulation/create_post.go
modify x/blog/types/codec.go
create x/blog/types/message_create_post.go
create x/blog/types/message_create_post_test.go

🎉 Created a message `createPost`.

As always, start with a proto file. Inside the proto/blog/tx.proto file, the MsgCreatePost message has been created. Edit the file to add the line that defines the id for message MsgCreatePostResponse:

message MsgCreatePost {
  string creator = 1;
  string title = 2;
  string body = 3;
}

message MsgCreatePostResponse {
  uint64 id = 1;
}

Review the message code

Review the Cosmos SDK message type with proto message. The MsgCreatePost has three fields: creator, title, and body. Since the purpose of the MsgCreatePost message is to create new posts in the store, the only thing the message needs to return is an ID of a created post. The CreatePost rpc was already added to the Msg service:

service Msg {
  rpc CreatePost(MsgCreatePost) returns (MsgCreatePostResponse);
}

Define messages logic

In the newly scaffolded x/blog/keeper/msg_server_create_post.go file, you can see a placeholder implementation of the CreatePost function. Right now it does nothing and returns an empty response. For your blog chain, you want the contents of the message (title and body) to be written to the state as a new post.

You need to do two things:

• Create a variable of type Post with title and body from the message
• Append this Post to the store

func (k msgServer) CreatePost(goCtx context.Context, msg *types.MsgCreatePost) (*types.MsgCreatePostResponse, error) {
    // Get the context
    ctx := sdk.UnwrapSDKContext(goCtx)

    // Create variable of type Post
    var post = types.Post{
        Creator: msg.Creator,
        Title:   msg.Title,
        Body:    msg.Body,
    }

    // Add a post to the store and get back the ID
    id := k.AppendPost(ctx, post)

    // Return the ID of the post
    return &types.MsgCreatePostResponse{Id: id}, nil
}

Define Post type and AppendPost keeper method

Define the Post type and the AppendPost keeper method.

When you define the Post type in a proto file, Ignite CLI (with the help of protoc) takes care of generating the required Go files.

Create the proto/blog/post.proto file and define the Post message:

syntax = "proto3";

package blog.blog;

option go_package = "blog/x/blog/types";

message Post {
  string creator = 1;
  uint64 id = 2;
  string title = 3; 
  string body = 4; 
}

The contents of the post.proto file are standard. The file defines:

• A package name blog.blog that is used to identify messages
• The Go package go_package = "blog/x/blog/types" where new files are generated
• The message message Post

Continue developing your blog chain.

Define keeper methods

The next step is to define the AppendPost keeper method.

Create the x/blog/keeper/post.go file and start thinking about the logic of the function and what you want to call the prefixes. The file will be empty for now.

• To implement AppendPost you must first understand how the key store works. You can think of a store as a key-value database where keys are lexicographically ordered. You can loop through keys and use Get and Set to retrieve and set values based on keys. To distinguish between different types of data that a module can keep in its store, you can use prefixes like product/ or post/.

• To keep a list of posts in what is essentially a key-value store, you need to keep track of the index of the posts you insert. Since both post values and post count (index) values are kept in the store, you can use different prefixes: Post/value/ and Post/count/.

Then, add these prefixes to the x/blog/types/keys.go file in the const and add a comment that describes the keys:

const (
    // ...

    // Keep track of the index of posts
    PostKey      = "Post/value/"
    PostCountKey = "Post/count/"
)

Your blog is now updated to take these actions when a Post message is sent to the AppendPost function:

• Get the number of posts in the store (count)
• Add a post by using the count as an ID
• Increment the count
• Return the count

Write data to the store

In the x/blog/keeper/post.go file, draft the AppendPost function. You can add these comments to help you visualize what you do next:

package keeper

import (
    "encoding/binary"

    "github.com/cosmos/cosmos-sdk/store/prefix"
    sdk "github.com/cosmos/cosmos-sdk/types"

    "blog/x/blog/types"
)

// func (k Keeper) AppendPost() uint64 {
//   count := k.GetPostCount()
//   store.Set()
//   k.SetPostCount()
//   return count
// }

First, implement GetPostCount:

func (k Keeper) GetPostCount(ctx sdk.Context) uint64 {
    // Get the store using storeKey (which is "blog") and PostCountKey (which is "Post/count/")
    store := prefix.NewStore(ctx.KVStore(k.storeKey), []byte(types.PostCountKey))

    // Convert the PostCountKey to bytes
    byteKey := []byte(types.PostCountKey)

    // Get the value of the count
    bz := store.Get(byteKey)

    // Return zero if the count value is not found (for example, it's the first post)
    if bz == nil {
        return 0
    }

    // Convert the count into a uint64
    return binary.BigEndian.Uint64(bz)
}

Now that GetPostCount returns the correct number of posts in the store, implement SetPostCount:

func (k Keeper) SetPostCount(ctx sdk.Context, count uint64) {
    // Get the store using storeKey (which is "blog") and PostCountKey (which is "Post/count/")
    store := prefix.NewStore(ctx.KVStore(k.storeKey), []byte(types.PostCountKey))

    // Convert the PostCountKey to bytes
    byteKey := []byte(types.PostCountKey)

    // Convert count from uint64 to string and get bytes
    bz := make([]byte, 8)
    binary.BigEndian.PutUint64(bz, count)

    // Set the value of Post/count/ to count
    store.Set(byteKey, bz)
}

Now that you have implemented functions for getting the number of posts and setting the post count, at the top of the same x/blog/keeper/post.go file, implement the logic behind the AppendPost function:

func (k Keeper) AppendPost(ctx sdk.Context, post types.Post) uint64 {
    // Get the current number of posts in the store
    count := k.GetPostCount(ctx)

    // Assign an ID to the post based on the number of posts in the store
    post.Id = count

    // Get the store
    store := prefix.NewStore(ctx.KVStore(k.storeKey), []byte(types.PostKey))

    // Convert the post ID into bytes
    byteKey := make([]byte, 8)
    binary.BigEndian.PutUint64(byteKey, post.Id)

    // Marshal the post into bytes
    appendedValue := k.cdc.MustMarshal(&post)

    // Insert the post bytes using post ID as a key
    store.Set(byteKey, appendedValue)

    // Update the post count
    k.SetPostCount(ctx, count+1)
    return count
}

By following these steps, you have implemented all of the code required to create new posts and store them on-chain. Now, when a transaction that contains a message of type MsgCreatePost is broadcast, the message is routed to your blog module.

• k.CreatePost calls AppendPost
• AppendPost gets the number of posts from the store, adds a post using the count as an ID, increments the count, and returns the ID

Now that you have added the functionality to create posts and broadcast them to our chain, you can add querying.

Display posts

To display posts, scaffold a query:

ignite scaffold query posts --response title,body

Two components are responsible for querying data:

• An rpc inside service Query in a proto file that defines data types and specifies the HTTP API endpoint
• A keeper method that performs the querying from the key-value store

First, review the services and messages in proto/blog/query.proto. The Posts rpc accepts an empty request and returns an object with two fields: title and body. Now you can make changes so it can return a list of posts. The list of posts can be long, so add pagination. When pagination is added, the request and response include a page number so you can request a particular page when you know what page has been returned.

To define the types in proto files, make the following updates in proto/blog/query.proto:

1. Add the import:

import "blog/post.proto";

2. Add pagination to the post request:

message QueryPostsRequest {
  // Adding pagination to request
  cosmos.base.query.v1beta1.PageRequest pagination = 1;
}

3. Add pagination to the post response:

message QueryPostsResponse {
  // Returning a list of posts
  repeated Post Post = 1;

  // Adding pagination to response
  cosmos.base.query.v1beta1.PageResponse pagination = 2;
}

To implement post querying logic in the x/blog/keeper/grpc_query_posts.go file, delete the contents of that file and replace it with:

package keeper

import (
    "context"

    "github.com/cosmos/cosmos-sdk/store/prefix"
    sdk "github.com/cosmos/cosmos-sdk/types"
    "github.com/cosmos/cosmos-sdk/types/query"
    "google.golang.org/grpc/codes"
    "google.golang.org/grpc/status"

    "blog/x/blog/types"
)

func (k Keeper) Posts(c context.Context, req *types.QueryPostsRequest) (*types.QueryPostsResponse, error) {
    // Throw an error if request is nil
    if req == nil {
        return nil, status.Error(codes.InvalidArgument, "invalid request")
    }

    // Define a variable that will store a list of posts
    var posts []*types.Post

    // Get context with the information about the environment
    ctx := sdk.UnwrapSDKContext(c)

    // Get the key-value module store using the store key (in our case store key is "chain")
    store := ctx.KVStore(k.storeKey)

    // Get the part of the store that keeps posts (using post key, which is "Post-value-")
    postStore := prefix.NewStore(store, []byte(types.PostKey))

    // Paginate the posts store based on PageRequest
    pageRes, err := query.Paginate(postStore, req.Pagination, func(key []byte, value []byte) error {
        var post types.Post
        if err := k.cdc.Unmarshal(value, &post); err != nil {
            return err
        }

        posts = append(posts, &post)

        return nil
    })

    // Throw an error if pagination failed
    if err != nil {
        return nil, status.Error(codes.Internal, err.Error())
    }

    // Return a struct containing a list of posts and pagination info
    return &types.QueryPostsResponse{Post: posts, Pagination: pageRes}, nil
}

Add gRPC to the module handler

In the x/blog/module.go file:

1. Add "context" to the imports, don't save the file yet.

import (
    "context"

    // ...
)

2. Update the RegisterGRPCGatewayRoutes function to register the query handler client:

// RegisterGRPCGatewayRoutes registers the gRPC Gateway routes for the module.
func (AppModuleBasic) RegisterGRPCGatewayRoutes(clientCtx client.Context, mux *runtime.ServeMux) {
    types.RegisterQueryHandlerClient(context.Background(), mux, types.NewQueryClient(clientCtx))
}

3. Now that you've modified the file with the two updates, now it's safe to save the file.

Use the CLI to create a post

Now that you have implemented logic for creating and querying posts, you can interact with your blog chain using the command line. The blog chain binary is blogd.

First, start the chain on your development machine by running the following command in the blog directory:

ignite chain serve

The binary is built by the ignite chain serve command bit it can also be built by running:

ignite chain build

To create a post at the command line:

blogd tx blog create-post foo bar --from alice

The transaction is output to the terminal. You are prompted to confirm the transaction:

{"body":{"messages":[{"@type":"/blog.blog.MsgCreatePost","creator":"blog1ctxp3pfdtr3sw9udz2ptuh59ce9z0eaa2zvv6w","title":"foo","body":"bar"}],"memo":"","timeout_height":"0","extension_options":[],"non_critical_extension_options":[]},"auth_info":{"signer_infos":[],"fee":{"amount":[],"gas_limit":"200000","payer":"","granter":""}},"signatures":[]}

confirm transaction before signing and broadcasting [y/N]: y

Type y to sign and broadcast the transaction.

Congratulations, you built a chain binary and used the blogd binary CLI to create a blog post.

Use the CLI to query posts

To query the list of all on-chain posts:

blogd q blog posts

The result:

Post:
- body: bar
  creator: blog1ctxp3pfdtr3sw9udz2ptuh59ce9z0eaa2zvv6w
  id: "0"
  title: foo
pagination:
  next_key: null
  total: "1"

Conclusion

Congratulations. You have built a blog blockchain!

You have successfully completed these steps:

• Write blog posts to your chain
• Read from blog posts
• Scaffold a Cosmos SDK message
• Define new types in protocol buffer files
• Write keeper methods to write data to the store
• Register query handlers
• Read data from the store and return it as a result a query
• Use the CLI to broadcast transactions