A library for working with IPFS in Python.
It includes a programmer-friendly API called ipfs_api for easy usage of some of the most useful IPFS features, a module called ipfs_datatransmission that allows for easy direct P2P data transmission between two IPFS nodes and a module called ipfs_lns for remembering the multiaddressess of known IPFS nodes ("contacts") to speed up connection times.

Under the hood these modules use a slightly updated version ipfshttpclient, a no-longer-maintained Python package for interacting with IPFS via its HTTP-API. The modified package can be accesses as ipfs_api.ipfshttpclient2 to gain full access to all supported ipfs interactivity in a structured manner similar to the IPFS CLI. In environments where interaction with the IPFS via HTTP is not possible, this package automatically uses ipfs_cli as a fallback, a module that provides all the functionality of ipfs_api by communicating to ipfs via its command-line interface.

Package Contents:

Python Modules:

• ipfs_api: a wrapper for the module ipfshttpclient2 that provides a simple API for some of the most useful features of IPFS
• ipfs_datatransmission: a module that enables P2P data transmission between IPFS nodes
• ipfs_lns: a module that allows IPFS nodes and their multiaddresses to be stored in app data to make them easier to find in the IP layer of the internet (easier to connect to)
• ipfshttpclient2: a modified version of the official ipfshttpclient module that has been expanded and updated where needed

Other:

• Examples: a folder of scripts demonstrating how to work with ipfs_datatransmission

Setup

Installation

pip install IPFS-Toolkit

Prerequisites

IPFS-Toolkit is made for interacting with IPFS in the Python programming language. One configuration must be applied to IPFS in order to use IPFS-Toolkit's main features.

• Install IPFS (Desktop version or CLI, doesn't matter). https://docs.ipfs.io/install/

• Enable "Libp2pStreamMounting" in IPFS:
  Desktop Version/WebUI: on the Settings tab, scroll down to "IPFS CONFIG" and change the line that reads:
  change from:
  "Libp2pStreamMounting": false,
  to:
  "Libp2pStreamMounting": true,
  Click the "Save" button and restart IPFS.

  CLI Version: run:
  ipfs config --json Experimental.Libp2pStreamMounting true

• Python3, version 3.6 or above https://www.python.org/downloads/

• Pip for Python3: https://pip.pypa.io/en/stable/installing/

• If you want to use the source code, install the following prerequisite Python modules:
  (depending on your Python installation you may need to use pip instead of pip3 in the following commands)

pip3 install appdirs httpcore httpx idna multiaddr setuptools wheel zmq

Modules Contained in IPFS-Toolkit

IPFS-API

import ipfs_api

This has a simplified and more user-friendly yet limited API for interacting with IPFS compared to the ipfshttpclient.

Usage examples:

print(ipfs_api.my_id()) # print your IPFS peer ID  
cid = ipfs_api.publish('./SomeFileOrDir') # upload file or directory to IPFS and store it's CID in a variable  

# Managing IPNS keys  
ipfs_api.create_ipns_record('MyWebsite') # generate a pair of IPNS name keys  
ipfs_api.update_ipns_record('MyWebsite', './SomeFileOrDir') # upload file to IPFS & publish   
ipfs_api.update_ipns_record_from_cid('MyWebsite', cid) # publish IPFS content to IPNS key  

# IPFS PubSub  
ipfs_api.pubsub_subscribe("test", print) # print is the eventhandler  
ipfs_api.pubsub_publish("test", "Hello there!")

HTTP API access:

This allows you to almost full access the IPFS API through a naming structure similar to its CLI through ipfshttpclient's client object. At this stage of development it is incomplete.

Examples:

import ipfs_api
ipfs_api.http_client.add("helloworld.txt")
ipfs_api.http_client.routing.findpeer("QMHash")
ipfs_api.http_client.swarm.peers()

To check and somewhat manage the status of the IPFS daemon, see Examples/Demo-Check-IPFS-Status.py

IPFS-Datatransmission

import ipfs_datatransmission

A Python module for limitless, easy, private peer-to-peer data transmission other over the IPFS Network.

This module has three main pairs functions for use by the user:

• Simple Data Transmission:
  • transmit_data(data, peerID, listener_name)
  • listen_for_transmissions(listener_name, eventhandler)
• File Transmission:
  • transmit_file(data, peerID, listener_name)
  • listen_for_file_transmissions(listener_name, eventhandler)
• Conversations (Ping-Pong Data & File Transmission):
  • start_conversation(data, peerID, listener_name)
  • listen_for_conversations(listener_name, eventhandler)

See the Examples folder for the different ways of using these functions. They are designed in a way that makes them easy to use for simple applications but feature-rich for more demanding ones, such as encryption integration and access to low-level protocol settings.

Listener Functions:

The listener functions (listen_for_transmissions, listen_for_file_transmissions, and listen_for_conversations) produce Listener objects () that run on their own threads waiting for incoming data transmission or conversation requests. When another computer wants to send the first computer something, the other computer sends a transmission request which the Listener object receives and reacts to by creating a Transmission reception object () which runs on its own separate thread to receive the transmission, or in the case of the conversation listener, calls its user-defined eventhandler so that the conversation can be joined or ignored. This system of receiving data transmissions/conversations allows a computer to receive multiple transmissions addressed to the same listener simultaneously. Multiple Listeners of the same type (the types being DataTransmission, FileTransmission, or Conversation) can be run simultaneously and independently, for example by different programs. Those different listeners (instances of listen_for_transmissions) must be named to distinguish them from each other for addressing purposes. This name is the "listener_name" parameter in the listener functions. The name is chosen by the user when creating the listener (e.g. when calling ListenForTransmission()) and must be provided when starting a transmission (e.g. when calling transmit_data()).

IPFS Technicalities:

How does IPFS-DataTransmission use IPFS to send data to another computer? After all, IPFS is a file system made for sharing and storing files, creating a content addressed internet. Does it really provide the functionality for two computers to communicate directly with each other?
Yes, IPFS does provide that functionality, although as of November 2021 that is still an experimental feature. That's why you have to configure IPFS to enable that feature as described in Prerequisites.txt.
After all, IPFS relies on peers sending data to each other over its decentralized network using a communication technology called libp2p. IPFS-DataTransmission essentially uses the libp2p module inside of the IPFS process running on the computer to communicate to other computers via an access point to the module that IPFS provides, the experimental feature that IPFS calls Libp2pStreamMounting. Libp2pStreamMounting works by allowing the user to set up a port forwarding system on the computer between the user's program (in this case IPFS-DataTransmission) and the libp2p process inside of IPFS. After setting up this port forwarding the user (in this case the IPFS-DataTransmission script) can communicate to the chosen local port on their computer, and IPFS forwards that communication to the libp2p module running inside of IPFS on the other computer, which forwards it in turn to a preconfigured port on that computer, to be listened to by the user (in this case an IPFS-DataTransmission Listener object).
Because the Python API for interacting with IPFS (ipfshttpclient) doesn't yet support this experimental feature, I have updated the module myself, the result of which is the ipfshttpclient2 folder in this code project, which will remain part of this project until the changes are integrated into the official ipfshttpclient module.

RoadMap

ipfs_toolkit will become a multimodal IPFS library with a unified API for alternative modes of using IPFS:

• running an embedded IPFS node
• interacting with a separate IPFS node via its HTTP RPC

py_ipfs_node, the python library where the embedded IPFS node is being developed, will be incorporated into ipfs_toolkit when py_ipfs_node's API has stabilised.

You can check out a prototype version of the combination of these two libraries under the kubo_python branch of the IPFS-Toolkit repo: https://github.com/emendir/IPFS-Toolkit-Python/tree/kubo_python

Related Projects

• py_ipfs_node allows you to run IPFS nodes as a library where the IPFS software itself is go compiled into a library which python wraps around.

• IPFS-Toolkit-Python provides an interface to a separately running IPFS node (e.g. kubo running as a service in the background on your operating system) using its HTTP RPC API. It is based on ipfs-shipyard/py-ipfs-http-client, which hasn't been worked on properly for 4 years.

• ipfs-shipyard/py-ipfs is an effort to rewrite the full IPFS software in python, but it is "Not even remotely done yet", and doesn't look like it's been worked on properly for 7 years.

Links

This project's IPFS URL:
ipns://k2k4r8nismm5mmgrox2fci816xvj4l4cudnuc55gkfoealjuiaexbsup#IPFS-Toolkit