netmem
Basic memory synchronization across the network in Python.
This package has a network-synchronized dictionary that runs on asyncio event loops. It supports binding to the dictionary similar to tk.Variable() and is also compatible with tkinter and its event loops.
Installation
The easiest way is to just open your favorite terminal and type
pip install netmem
Alternatively you can clone this repo and install it with
python setup.py install
Requirements
The amazing aiohttp library
Python v3.5+
The basis for netmem is asynchronous IO and event loops, so I apologize to Python v2.x users and for that matter, Python v3.4. Although Python v3.4 supports asyncio, I really like the async for and async with constructs introduced in v3.5, and I use them in a number of places. Since Python is already on v3.6 at the time of this writing, I do not feel too terribly bad leaving v3.4 behind.
Usage
Here is the smallest meaningful example I can come up with. Run it on two different computers on the same network.
import tkinter as tk
import netmem
def main():
print("Run this on two different computers.")
mem = netmem.NetworkMemory()
mem.connect_on_new_thread(netmem.UdpConnector(local_addr=("225.0.0.1", 9991)))
tk1 = tk.Tk()
lbl = tk.Label(tk1, text="Favorite operating system:")
lbl.pack()
txt = tk.Entry(tk1, textvariable=mem.tk_var("fav_os"))
txt.pack()
tk1.mainloop()
if __name__ == "__main__":
main()
You can bind a listener to the NetworkMemory object to be notified when a value changes, such as when an update arrives over the network. The listener works like the following code snippet.
def memory_changed(netmem_dict, key, old_val, new_val)
print("Update {}:{}".format(key, new_val))
def main():
mem = netmem.NetworkMemory()
mem.add_listener(memory_changed)
mem["foo"] = "bar"
The output from this would be the following
Update foo:bar
Incidentally the underlying BindableDict class is pretty handy on its own, without even the network synchronizing capabilities.
Data Structure
NetworkMemory subclasses a Python dictionary, so you can access the data within it as you do any dictionary object. Additionally you can bind listeners to NetworkMemory (because in fact it subclasses a bindable dictionary, which is something I borrowed from other code I wrote).
Metadata
Release files for netmem 0.2.4
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | Actions |
|---|---|---|---|---|
| netmem-0.2.4.tar.gz | 13.8 kB | Without provenance | Details for netmem-0.2.4.tar.gz |
Release files / netmem-0.2.4.tar.gz
| Download URL | netmem-0.2.4.tar.gz |
|---|---|
| Size | 13.8 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
84eb2bd1f58f3e1fe52516d568f5847c358d06bd30029802ac082d170da01044
|
|
BLAKE2b-256 checksum How to use checksums |
c9ae8fdaec238ecc275beca0b658611efda8100fb29100ad28ba11299a6841a6
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
Security for netmem 0.2.4
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.