Skip to content
This repository has been archived by the owner on Jul 19, 2024. It is now read-only.

Latest commit

 

History

History
256 lines (177 loc) · 9.68 KB

modules.md

File metadata and controls

256 lines (177 loc) · 9.68 KB
title
Writing counterblock Plug-in Modules

counterblock is a modular application that allows developers to turn on or off various bits of its out-of-the-box functionality, as well as extending it with new functionality, through its plug-in architecture.

This document introduces the built-in modules, as well as discussing how you can write your own custom modules for counterblock that extend its functionality beyond what is possible out of the box.

Built-in Modules

counterblock ships with the following built-in modules:

  • assets: Implements basic asset functionality, such as extended asset info parsing, basic asset-related APIs, and more.
  • betting: Implements betting-specific API calls, tasks, and more.
  • dex: Implements API methods, order book and market info parsing, and more for Counterparty's distributed exchange. (Requires that the assets module be loaded for use.)
  • transaction_stats: Handles the compliation of transaction statistics.
  • counterwallet: Implements Counterwallet-specific API calls, tasks, and more. (Requires that the assets module be loaded for use.)

Any of these above modules may be enabled or disabled, allowing you to tune counterblock to your exact needs out of the box.

Custom Module Development

A counterblock module is simply a python module that utilizes a special plug-in API, to provide runtime integration into the counterblock base code.

For some examples of modules in use, check out the modules directory of the counterblock respository. These built-in modules are written just like any custom module would be, and provide a good launching point to see what is possible.

Processors: Hooking into runtime functionality

Most of the API functionality follows a specific Python decorator type syntax, to integrate into things like blockchain message processing, block level processing, startup/initialization processing and more. The general syntax is:

    from lib.processor import <processor_name>
    
    @<processor_name>.subscribe(enabled=<bool>, priority=<int>)
    def my_function(param1, param2):
        bla = do_foo()

Some other notes on processors:

  • If not specified, the defaults are enabled=true, priority=0.
  • When a processor is triggered methods are run in order of priority from highest to lowest.
  • Please note that any priority less than 0 or greater than 1000 is reserved for internal counterblock functionality, and custom plugins should only utilize priority settings under this number.

MessageProcessor

MessageProcessor runs once for each message as obtained from the counterpartyd message feed, for all activity that has been confirmed on the blockchain (i.e. at least 1 Bitcoin confirmation). msg will pass the message in the same format as the get_messages counterpartyd api method, msg_data corresponds to json.loads(msg['bindings']).

    @MessageProcessor.subscribe(enabled=True, priority=90) 
    def custom_received_xcp_alert(msg, msg_data):
        if msg['category'] != 'sends': return
        if message['status'] != 'valid': return
        if not msg_data['destination'] in MY_ADDRESS_LIST: return
        if not msg_data['asset'] == 'XCP': return 
        print('Received %s XCP at My Address %s from %s' %(
        	(float(msg_data['quantity'])/10**8), msg_data['destination'], msg_data['source']))
        return

Note that with MessageProcessor handlers, you can return 'ABORT_THIS_MESSAGE_PROCESSING' to prevent the running of further MessageProcessors (i.e. of lesser priority than the current one) for the message being currently processed.

MempoolMessageProcessor

MempoolMessageProcessor works similar to MessageProcessor, however, for messages out the mempool (i.e. that are not confirmed and included on the blockchain yet). The format of the data supplied to the processor is slightly different though, and looks like this:

    @MempoolMessageProcessor.subscribe(enabled=True, priority=90) 
    def custom_received_xcp_alert(msg, msg_data):
        assert msg['_message_index'] == 'mempool'
        assert msg['tx_hash']
        assert msg['command']
        assert msg['category']
        assert msg_data #the actual message data
        assert msg['timestamp']
        assert msg['viewed_in_block']
        
        #prevent running of further MempoolMessageProcessor's of lesser priority for the message being processed
        return 'ABORT_THIS_MESSAGE_PROCESSING'

BlockProcessor

BlockProcessor run once per new block, after all MessageProcessor functions have completed.

    @BlockProcessor.subscribe(priority=0) 
    def alertBlock(): 
        print('Finished processing messages for this block')

StartUpProcessor

StartUpProcessor runs once on counterblock startup.

    @StartUpProcessor.subscribe()
    def my_db_config(): 
        config.my_db = pymongo.Connection()['my_db'] 

CaughtUpProcessor

CaughtUpProcessor runs once when counterblock catches up to the latest Counterpartyd block.

    @CaughtUpProcessor.subscribe()
    def caughtUpAlert(): 
        print('counterblock is now caught up to Counterpartyd!') 

RollbackProcessor

RollbackProcessor runs whenever the counterblock database is rolled back (either due to a blockchain reorg, or an explicit rollback command being specified to counterblock via the command line).

Note that if this processor runs and None is passed as max_block_index, it means that there was a reparse of all block data.

    @RollbackProcessor.subscribe()
    def rollbackAlert(max_block_index): 
        print('counterblock block database rolled back! Anything newer than block index %i removed!' % max_block_index) 

Other integration points

config.state

A number of internal state variables that a module may need to access are stored in config.state.

For example, if you want to run a process for every new block (but not when counterblock is catching up):

    @BlockProcessor.subscribe() 
    def my_custom_block_event(): 
        if not (config.state['cpd_latest_block_index'] - config.state['my_latest_block']['block_index']) == 1: 
            return
        #Do stuff here

Enhancing the API

To add an API method for counterblock to provide:

    from lib.processor import API
    
    #(note that the dispatcher add_method does not take arguments) 
    @API.add_method
    def my_foo_api_method(): 
        return 'bar'

Upon doing the above, my_foo_api_method is now a valid API method, and callable from any client that that utilizes your counterblock JSON RPC API.

start_task

To start a task that runs in a seperate lightweight thread (either immediately, or with a delay), use start_task:

    from lib.processor import start_task
    
    def run_my_task():
        print("Foo bar!!")
        #start again in 5 minutes
        start_task(run_my_task, delay=5*60)
        
    #start task the first time with no delay
    start_task(run_my_task)

Module configuration file

After creating your module, you will need to tell counterblock about it, so that it can load it on startup. To do this, you should edit (or create) a file called modules.conf (or modules.testnet.conf for testnet), which should be located in the counterblock config-dir (~xcp/.config/counterblock on a federated node).

To load your custom module, specify the module's path under [LoadModule] relative to the counterblock base-dir. i.e.:

[LoadModule]
'lib/vendor' = True

The above configuration would look for a vendor.py, or vendor folder (with required __init__.py file present), and load the plugin code from there. Note that you should not include any .py suffix on the filename.

To change the default behavior for counterblock modules/events, change the corresponding processor config. For instance:

To disable a processor:

#(must be bool)
[BlockProcessor]
generate_wallet_stats = False

To change a processor's priority:

#(must be int) 
[MessageProcessor]
parse_issuance = 5

To change priority and enable:

#(tuple, order does not matter)
[MessageProcessor]
parse_issuance = 5, True 
parse_issuance = True, 5

Here's an extensive counterblock modules.conf example config file:

[MessageProcessor]
#Tweak core messaging processing
# (don't use these unless you know what you're doing)
handle_exceptional = True
handle_invalid = True
parse_insert = True
handle_reorg = True
parse_issuance = 10, True
parse_balance_change = True
parse_trade_book = True
parse_broadcast = True

[StartUpProcessor]
#Enable/disable core functionality (all enabled by
# default, don't use these unless you know what you're doing)
start_cpd_blockfeed = True
check_blockchain_service = True
expire_stale_orders = True
start_api = True

[LoadModule]
#Load custom modules
lib/modules/reparse_timer = True

Please note that function names must be exact.

Command-line functions

Counterblock also provides the ability to load/unload/disable/etc modules via the command line. (Although, please note that we recommend manually editing the appropriate modules config file(s), at least for starters.)

To enable a custom module, run the following command:

counterblock enmod 'lib/vendor'

To disable a loaded module:

counterblock dismod 'lib/vendor' 

To list loaded modules and processors:

counterblock listmod