Skip to content

Latest commit

 

History

History
52 lines (35 loc) · 3.42 KB

README.md

File metadata and controls

52 lines (35 loc) · 3.42 KB
Raw

Running

To run this Python code, you will need to open it in Jupyter Notebooks.

You will also need to install various python libraries, depending what is already installed. I have not tracked what you will need, and would recommend running the notebook - it will give errors when a library cannot be found. To install these, you may want to use conda or pip. Instructions for pip/conda

Once you can open the notebook, I would recommend changing paths so the notebook points at the sample data log file in data/. Really, you will want to make your own, if you want to analyze your own data. I pull data using the chia command line tool. This is done in scripts/log_netspace_size.sh.

Notes

  • This is not a polished notebook, depending on what you consider polished. I know it is imperfect, but has suited my needs. I just want to make it available for anyone who wants to dick around with it.

  • I have made use of primarily a sigmoid-like function for netsize prediction. There is no real reason to think the netspace will follow a sigmoid curve - so make your own functions, it is easy!

How I got started, and use this notebook and scripts

I pulled netspace size data from https://www.chiaexplorer.com/charts/netspace for past data, then started logging on my own. I combined the two with a python script. I don't have the python code i used to pull data from the json and throw it into my csv log (i deleted the code b/c i assumed i was done with it. oops) But it isn't hard to do that. There is more data that my logger grabs, than chia explorer has. So i just filled the "unknown" columns with 0, when converting the json data to csv.

To pull data from chiaexplorer, open a dev console in the browser, refresh the page, and there should be document loaded called netSpace, from the URL: https://api2.chiaexplorer.com/chart/netSpace?period=2w. just copy-paste the data into a file. You could also just skip pulling data from chiaexplorer, but it gives a good history for prediction purposes.

For logging data on the farmer, I just have a cronjob running every 5 minutes:

*/5 * * * * /home/ethanblagg/Documents/scripts/log_netspace_size.sh

log_netspace_size.sh

#!/bin/bash

. /home/ethanblagg/Documents/chia-blockchain/activate
file="/home/ethanblagg/Documents/scripts/netspace-size.log"
epoch_time=$(date +%s)
chia_farmed=$(chia farm summary | grep "Total chia farmed" | grep -Eo "[0-9]*\.[0-9]*")
num_plots=$(chia farm summary | grep "Plot count" | grep -Eo "[0-9]+")
plots_size_TiB=$(chia farm summary | grep "Total size of plots" | grep -Eo "[0-9]*\.[0-9]*")
net_size_PiB=$(chia farm summary | grep "Estimated network space" | grep -Eo "[0-9]*\.[0-9]*")
plots_size_PiB=`echo "scale=15; $plots_size_TiB / 1024" | bc`
win_prob=`echo "scale=15; $plots_size_PiB / $net_size_PiB" | bc`


echo "Chia farmed: $chia_farmed XCH"
echo "Num plots:   $num_plots"
echo "Plots size:  $plots_size_TiB TiB"
echo "Plots size:  $plots_size_PiB PiB"
echo "Netspace:    $net_size_PiB PiB"
echo "Owned fraction of net: $win_prob"

echo "$epoch_time,$chia_farmed,$num_plots,$plots_size_PiB,$net_size_PiB,$win_prob" >> $file

So once you have some data from chiaexplorer and/or logging, just run the jupiter notebook. I am running it on my iMac, and just having it pull the log from my farmer via ssh. Then it whizzes and whirs and makes pretty colors.