Skip to content

opentensor/polkascan-explorer

 
 

Repository files navigation

Polkascan Explorer

Screenshot 2022-10-10 at 10 03 34 Screenshot 2022-10-10 at 10 30 11

Polkascan Explorer provides a generalized block explorer for Substrate-based blockchains.

It combines data retrieved directly from the Substrate node RPC with data retrieved from a third-party indexer API.

At the moment the following third-party data are supported:

  • Polkascan API
  • Subsquid

Installation using Subsquid endpoints

See the installation instructions of the Polkascan Explorer UI repos if the explorer is solely going to run with Subsquid endpoints.

Installation using Polkascan API endpoints

Run init.sh to initialize repository; this will basically run:

  • git submodule update --init --recursive
  • cp explorer-ui-config.json explorer-ui/src/assets/config.json
  • cp explorer-ui-privacy-policy.html explorer-ui/src/assets/privacy-policy.html

The explorer-ui-config.json file contains the URLs of the exposed Substrate and Explorer API endpoints

Running the application

Docker

  • docker-compose up --build

Local

Harvester

  • cd harvester
  • pip install -r requirements.txt
  • ./harvester-cli.sh run

Explorer API

  • cd explorer-api
  • pip install -r requirements_api.txt
  • ./start-api.sh

Explorer UI

For a dev server, open a terminal and run:

cd polkadapt
npm i
npm run build

When making changes in polkadapt source files you have to build again.

Now open a second terminal and run:

npm i
npm run start

Navigate to http://localhost:4200/. The app will automatically reload if you change any of the source files.

Services

For more information or modification see the Docker compose file

Components

The explorer application consist of several components:

Harvester component

The harvester retrieves data from the connected Substrate node and stores it into a MySQL (by default) database.

Storage cron

With the processing of each block, it can also be desirable to retrieve and decode certain storage records. By default, for every block the System.Events are stored as this is a fundamental element to determine which calls are executed.

When other storage records are needed, for example a balance snapshot of all accounts every 10000 blocks, additional cron items can be added:

cd harvester
./harvester-cli.sh storage-cron add 

> Block interval (e.g. 10 = every 10th block): 10000
> Pallet: System
> Storage function: Account
> Added cron System.Account every 10000 blocks

Check the current storage cron items:

  Id    Block interval  Pallet    Storage name
----  ----------------  --------  --------------
   1                 1  System    Events
   2             10000  System    Account

Then run the harvester:

./harvester-cli.sh run 

Storage tasks

When storage cron items retrieve records as part of the block harvest process, storage tasks can be added to retrieve records for any given blocks that are already processed.

Also, this feature can be used standalone, so it basically acts as a storage harvester.

Example: Store total issuance for blocks 1-1000:

cd harvester
./harvester-cli.sh storage-tasks add
> Pallet: Balances
> Storage function: TotalIssuance
> Blocks (e.g. '100,104' or '100-200'): 1-1000
> Added task Balances.TotalIssuance for blocks 1-1000 

Then run only the 'cron' job of the harvester:

./harvester-cli.sh run --job cron 

List progress of tasks:

./harvester-cli.sh storage-tasks list
>   Id  Pallet    Storage name    Blocks                                 Complete
> ----  --------  --------------  -------------------------------------  ----------
>    1  System    Account         {'block_ids': [1, 2]}                  True
>    2  Balances  TotalIssuance   {'block_end': 1000, 'block_start': 1}  True

Explorer API component

The explorer API transforms the data via an ETL process into an explorer-specific format. It exposes a GraphQL endpoint and enables subscription-based communication to the UI.

Explorer UI component

Explorer UI is a client-sided Angular based application that utilizes PolkADAPT and its Adapters to obtain data from multiple data sources, like the Explorer API and the Substrate node. Its design is based on flat Material component design, styled in Polkascan branding.

Modifications

Substrate Node

By default, a build of Substrate Node Template is used. If a local Substrate node is already running on the host machine, you can change environment variable in the docker-compose.yml: SUBSTRATE_RPC_URL=ws://host.docker.internal:9944

Or to any public or private websocket endpoint

To use PolkadotJS Apps with the local Substrate node, go to https://polkadot.js.org/apps/?rpc=ws%3A%2F%2F127.0.0.1%3A9944#/explorer

Explorer UI

Datasources

The UI is utilizing PolkAdapt to combine on-chain data retrieved directly from the Substrate node, with the GraphQL endpoint with indexed data served by the Explorer API. In this way data like events and extrinsics can be retrieved fast, with the verification of on-chain data.

By default, the UI are using the local endpoints, new networks can be added by extending the dict:

{
  "local": {
    "substrateRpcUrlArray": ["ws://127.0.0.1:9944"],
    "explorerWsUrlArray": ["ws://127.0.0.1:8000/graphql-ws"]
  },
  "polkadot": {
    "substrateRpcUrlArray": ["wss://<PUBLIC_ENDPOINT>"],
    "explorerWsUrlArray": ["wss://<HOSTED_EXPLORER_API_ENDPOINT>/graphql-ws"]
  }
}

Known Issues and Limitations

  • Substrate runtimes that implements metadata prior to MetadataV14 decode in a different format as expected by the explorer, so not all functionality is available. This can have effect on early blocks as well, a workaround is to set BLOCK_START env setting for the harvester, to a block number with MetadataV14.
  • Errors during building explorer-ui Docker container on M1 architecture. This could happen when the chromium binary is not available for arm64, there is a workaround available
  • Currently, the explorer-api application supports Python version >3.6 and <3.10

License

https://github.com/polkascan/explorer/blob/main/LICENSE

Releases

No releases published

Packages

No packages published

Languages

  • Shell 97.7%
  • HTML 2.3%