Project deployer

Let's find out how to handle the requests made by writing values on a specific path set by the project user.

(1) Install ain-py

First of all, let’s install ain-py. ain-py is a Python SDK used to interact with AI Network blockchain. It can be installed through PyPI.(We used version 0.1.4 in this tutorial.)

# if your server is node js server, use ain-js
# npm install @ain-blockchain/ain-js
pip install ain-py==0.1.4

(2) Import package

Then, import the packages that we will use in the tutorial.

import os
import json
import logging
import asyncio
import torch
from ain.ain import Ain
from flask import Flask, request
from ain.types import ValueOnlyTransactionInput
from transformers import GPT2LMHeadModel, GPT2Tokenizer

ain : Package used to interact with AI network.

asyncio : Package for asynchronous programming. Used with ain.

flask : Package for web framework. Used to create API servers.

transformers : Package for dealing with natural language processing.

(3) Connect to a blockchain node

In this step, let’s connect with the blockchain node. PROVIDER_URL can contain the URL of AI Network's Test Net (https://testnet-api.ainetwork.ai/), Main Net (https://mainnet-api.ainetwork.ai/)).

To use ain-py on Main Net, chainId must be set to 1 (Test Net: 0, Main Net: 1)

PROVIDER_URL = os.environ['PROVIDER_URL']
# if use testnet, set chainId to 0
ain = Ain(PROVIDER_URL, chainId=1)

(4) Set an account

If you are connected to the blockchain node, use the account's private key to set up your account. When deploying the project to Ainize, the environmental variable AINIZE_INTERNAL_PRIVATE_KEY will be automatically set. addAndSetDefaultAccount is a function that adds an account from a private key that enters the parameter and sets the account as the default account.

AINIZE_INTERNAL_PRIVATE_KEY = os.environ['AINIZE_INTERNAL_PRIVATE_KEY']
ain.wallet.addAndSetDefaultAccount(AINIZE_INTERNAL_PRIVATE_KEY)

(5) Set values

At this point, everything should be ready! When a request comes in to the server, it will get the result value and write the result value on the AI Network Blockchain.

When a value is written in the specific path set by the project user, the trigger function will be triggered and a request will be sent to the registered endpoint.

res = json.loads(request.data.decode('utf-8'))
print(res)
# {
#     "function": {
#         "function_id": "gpt2-pride-and-prejudice",
#         "function_type": "REST",
#         "function_url": "<https://main-ainize-trigger-tutorial-scy6500.endpoint.ainize.ai/trigger>"
#     },
#     "transaction": {
#         "address": "Oxbb5ecD24E8Bf0937d8c22D4440Fa95D50a1d108E",
#         "extra": {
#             "created_at": 1640223741870,
#             "executed_at": 1640223741870
#         },
#         "hash": "Oxeeb009260026a0fb030aa4218104ec2acecbed8c8cf4dba31650c1509cc2b0a1",
#         "signature": "Oxeeb009260026a0fb030aa4218104ec2acecbed8c8cf4dba31650c1509cc2b0a1a1dfa59513aac069f4a389e925cc885ae7b3ae7f8e1e2483603c555eb34288152c7b2b4ab774553cfe5f2761d16507fe500ea862ad0b34e52b436f5a91afa5cbib",
#         "tx_body": {
#             "nonce": -1,
#             "operation": {
#                 "is_global": false,
#                 "ref": "/apps/ainize_tutorial/Oxbb5ecD24E8Bf0937d8c22D4440Fa95D50a1d108E/1640223741870/user",
#                 "type": "SET_VALUE",
#                 "value": "{
#                     "baseText": "I love him. He's not proud. I was wrong. I was entirely wrong about him.",
#                     "len": 50
#                 }"
#             },
#             "timestamp": 1640223741430
#         }
#     }
# }

First, let's check if the incoming request is valid. We need to check if transaction , tx_body, operation are among the parameters of the request.

if (not res.get('transaction') or
        not res['transaction'].get('tx_body') or
        not res['transaction']['tx_body'].get('operation')):
    return f'Invalid transaction : {res}', 400
transaction = res['transaction']['tx_body']['operation']
tx_type = transaction['type']
if tx_type != 'SET_VALUE':
    return f'Not supported transaction type : {tx_type}', 400

If the incoming request is valid, put the requested value into the model to get the result value and write the result value to the blockchain. To write a value to the blockchain, you need to use setValue of ain-py (but async was used to process it asynchronously). path contains the path to write the value, and value contains the value to write in the path.

# path = /apps/appName/path/to/your/want
async def set_value(path, value):
    result = await ain.db.ref(path).setValue(
        ValueOnlyTransactionInput(
            value=value,
            nonce=-1
        )
    )

value = eval(transaction['value'])
# Get the result by putting baseText and len as input to the model
result = make_story(value['baseText'], value['len'])
try:
    # Delete the 'user' of the ref to make the result_ref 
    # like /apps/appName/.../result
    result_ref = transaction['ref'].split('/')[:-1]
    result_ref.append('result')
    result_ref = '/'.join(result_ref)
    print(result_ref)
    # /apps/ainize_tutorial/Oxbb5ecD24E8Bf0937d8c22D4440Fa95D50a1d108E/1640223741870/result
    asyncio.run(set_value(result_ref, result))
except Exception as e:
    logging.error(f'setValue failure : {e}')
    return f'setValue failure : {e}', 500
return '', 204

You can see that the results are written correctly on the AI Network Blockchain.