neo-mamba 0.9

Python SDK for the NEO 3 blockchain

neo-mamba

This SDK intends to provide building blocks for Python developers to interact with the NEO blockchain as a means to lower the entry barrier.

It is a work in progress and thus you can expect that not all parts of the blockchain are supported. What is present should be functioning correctly unless explicitely mentioned (e.g. because they depend on structures not yet available).

Please report any issues on Github or submit ideas how to improve the SDK.

Also check out our Python smart contract compiler Boa !

Quick install

pip install wheel
pip install neo-mamba

or

git clone https://github.com/CityOfZion/neo-mamba.git
cd neo-mamba
pip install wheel
pip install wheel -e .

This installs mamba with only the MemoryDB as possible backend. In order to install the requirements to use the LevelDB backend install the leveldb via extras

pip install -e .[leveldb]

Ensure you have PIP >= 19.3

pip install --upgrade "pip>=19.3"

For full documentation including more elaborate install instructions go to https://docs.coz.io/neo3/mamba/index.html.

Roadmap

Over time the following components are expected to be implemented starting with the items in “SDK Core”. As we release versions and receive feedback components may be added or restructured. Some components may end up living in separate repositories but be included here by default (e.g. virtual machine implementations).

• Core (v0.1)

• Network (v0.1)

• Storage (v0.1)

• Virtual Machine

• Smart contracts (v0.4)

• Wallet support (v0.8)

Performance

At the time of releasing v0.7 there was no RC2 test net available for performance measuring. Very few changes in RC2 should affect performance, therefore the RC1 measurements below should still be pretty accurate. RC3 (v0.8) has even less changes. Performance measurements are on hold until other priorities have been processed.

RC1 measurements

RC1 is the first release, out of 5, where neo-mamba syncs the blockchain slower than the official client neo-cli. Specifically, 30% slower. New bottlenecks as a result of the RC1 updates have been identified and will be improved on after the required RC2 updates.

Performance was measured by letting each client sync the RC1 TestNet from an offline file up to block height 146253 as available at that time. Syncing was done without verifying the block witnesses (--noverify flag in neo-cli). Each client used LevelDB as storage backend, with no other additional plugins running. Each client was synced separately with no additional resource usage on the system. The test was performed on OSX 10.15.7 on a i7-9750H with 16 GB ram.

FAQ

1. make docs fails with no theme named 'neo3' found (missing theme.conf?). -> python setup.py install. Try again.