bvanderlugt / pypocket

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

pypocket

Python Client SDK for Pocket Network

Installation

git clone git@github.com:pokt-foundation/pypocket.git
cd pypocket
pip install wheel
pip install -e .
pip install duckdb pyarrow pandas protobuf

Requirements

A Pocket RPC URL Can be acquired through either:

  1. Access to a Pocket Node with RPC
  2. An app endpoint from the Portal

Usage

Pocket RPC /query

All RPC methods that begin with /query are currently supported. These are all exposed under the PoktRPCDataProvider.

from pokt import PoktRPCDataProvider

rpc_url = "https://mainnet.gateway.pokt.network/v1/lb/<PortalID>"
pokt_rpc = PoktRPCDataProvider(rpc_url)

height = pokt_rpc.get_height()
supported_chains = pokt_rpc.get_supported_chains()
supply = pokt_rpc.get_supply()

Pocket DB

CLI script available as pokt-index for pulling in transactions into a format that can be queried.

$ pokt-index --help
usage: pokt-index [-h] [-s START] [-e END] [-j N_CORES] [-u URL] [-d INDEX_DIR] [-b BATCH_SIZE]

Index the pocket network blockchain data

optional arguments:
  -h, --help            show this help message and exit
  -s START, --start START
                        The block to start indexing from, defaults to either the first block, or the
                        last indexed block.
  -e END, --end END     The block to index to. Defaults to the latest block.
  -j N_CORES, --n-cores N_CORES
                        The number of cores to use when indexing, defaults to 4 less than the total
                        core count.
  -u URL, --url URL     The rpc url, defaults to http://localhost:8081.
  -d INDEX_DIR, --index-dir INDEX_DIR
                        The directory where the indexed files should be written to. Defaults to
                        'index' of the current working directory.
  -b BATCH_SIZE, --batch-size BATCH_SIZE
                        The number of blocks to write to each parquet file. Defaults to 250.

Schema of the available tables defined in pokt/index/schema.py

Tables are initially written in batches to parquet files, available at: ${INDEX_DIR}/<table-name>/*.parquet.

These tables can either be queried directly:

import duckdb

tx_dir = "/path/to/INDEX_DIR/txs/*.parquet"
con = duckdb.connection()
# Get result at python object
res = con.execute("SELECT COUNT(*) FROM read_parquet({})".format(tx_dir)).fetchall()
# Get result in pandas dataframe (needs pandas install)
df = con.execute("SELECT * FROM read_parquet({}) LIMIT 500".format(tx_dir)).df()
con.close()

Or ingested into an engine (using duckdb as reference):

import duckdb

tx_dir = "/path/to/INDEX_DIR/txs/*.parquet"
# Create a txs table on a new duckdb file.
con_new = duckdb.connection("new.duckdb")
con_new.execute("CREATE TABLE txs AS SELECT * FROM read_parquet({})".format(tx_dir))
con_new.close()
# Insert txs into an existing table on a duckdb file.
con = duckdb.connection("existing.duckdb")
con.execute("INSERT INTO txs SELECT * FROM read_parquet({})".format(tx_dir))
con.close()

Create Account

from pokt import create_new_ppk

ppk = create_new_ppk("super-secret-password")

Import Account

from pokt import PPK

ppk = PPK.from_file("/path/to/keyfile.json")

Sign a Message

from pokt import PPK, unlock_ppk

ppk = PPK.from_file("/path/to/keyfile.json")

unlocked = unlock_ppk(ppk, "super-secret-password")
signed = unlocked.sign("This message is from me!")

Verify a Signature

from pokt import verify_signature

pub_key="public_key" # see PoktRPCDataProvider.get_signing_info("address")
signed_message="signed message"

isValid = verify_signature(pub_key, signed_message)

Current Functionality

[x] -> Complete
[~] -> Partial functionality

Pocket RPC:

  • Pocket RPC /query
  • [~] Pocket DB
  • [] Pocket RPC /client

Wallet:

  • Create Account (PPK file/Private Key)
  • Import Account (from PPK/Private Key)
  • Sign message with account
  • Verify signed message
  • [] Create and sign transactions from account
  • [] Create and sign multisig transactions from account

Portal RPC:

  • [~] pocket RPC provider
  • [] web3 providers
  • [] figure out other types

About

License:MIT License


Languages

Language:Python 99.8%Language:Shell 0.2%