Skip to content

Latest commit

 

History

History
122 lines (96 loc) · 5.61 KB

README.md

File metadata and controls

122 lines (96 loc) · 5.61 KB
Raw

Filecoin Mechanistic Twin

Mechanistic model for the Filecoin Economy. You can use this model to forecast all the components underlying Filecoin's circulating supply (i.e., minting, vesting, locking and burning), based on a set of parameters that encode storage provider behavior. The model uses the following assumptions:

  • Forecasting is done daily. This means that each forecasting step corresponds to a day and the forecasted metrics correspond to the value we expect to see at the end of that day.
  • The model uses the current sector states (i.e. known schedule expirations) and it estimates future onboardings and future renewals.
  • The daily power onboarded is provided as a tunable parameter.
  • The sector renewal rates are provided as a tunable parameter.
  • Sector duration is a constant provided as a tunable parameter.
  • Filecoin Plus sectors have the same renewal rates and sector durations as other sectors.
  • The model uses the current pledge metrics (i.e. known scheduled expiration in pledge) to measure known active sectors, and it estimates pledge metrics coming from future onboardings and renewals using the same assumptions as the ones used to model storage power.
  • The model ignores storage deal collateral when computing the total locked FIL.

To learn more about how the model is designed, check the specifications linked at the end of this readme.

Prerequisites

numpy>=1.22
pandas>=1.4
requests>=2.28
matplotlib>=3.5.2
seaborn>=0.11.2

Installing

This package is available on PyPI and thus can be directly installed with pip:

pip install mechaFIL

Alternatively, this package can be installed from source by cloning this repository and installing it manually with the command:

python setup.py install

Setup for Data Access

mechaFIL uses historical data from Spacescope to provide Filecoin network econometric forecasts. Spacescope requires users to register for a unique token in order to request the historical data. Thus, each user of mechaFIL needs to register for a unique token to enable data access - this is a prerequisite to use mechaFIL.

Steps:

  1. Follow the instructions here to get your own unique bearer token for data access.
  2. Store your token in a json configuration file with the key auth_key. An example is:
{
    "auth_key": "Bearer ghp_xJtTSVcNRJINLWMmfDangcIFCjqPUNZenoVe"
}

Usage

There is a high-level function that can be used to run a forecast/simulation of circulating supply and its components. First you need to import the relevant packages:

import mechafil
import datetime

Now, you need to define some parameters:

# Starting date for the simulation
start_date = datetime.date(2021, 3, 15)
# Current date
current_date = datetime.date(2022, 11, 1) 
# Number of days to run the simulation (after current_rate)
forecast_length = 365
# Renewal rate of all sectors (percentage of raw-byte that will renew)
renewal_rate= 0.6
# Raw-byte power (in PiB) that is onboarded every day
rb_onboard_power = 12.0
# Percentage of raw-byte power onboarded that contains FIL+ deals
fil_plus_rate = 0.098
# Sector duration of newly onboarding sectors
duration = 360
# Pointer to Authentication Token
auth_config = "<UPDATE ME>"
# Method of computing QAP
qap_method = 'basic'

A few important notes regarding the inputs:

  • Due to data availability, the start date cannot be earlier than 2021-03-15.
  • The current date needs to be at least 2 days after the start date.
  • The current date needs to be at least 1 day before the actual current date. i.e. if you are running the simulation on 2023-01-15, then the maximum current date that is supported is 2023-01-14.
  • The parameters renewal_rate, rb_onboard_power and fil_plus_rate can be a single number or a vector of numbers. If they are a number, the model assumes that number as a constant throughout the simulation. If a vector is provided, then the vector needs to have the same size as the simulation length. The vector option gives the user the most flexibility since they can apply different constants throughout the simulation.
  • The optional parameter qap_method determines how network QAP will be computed. Two approaches are provided in the library, which we term basic and tunable. Setting this value to tunable will enable QAP to be computed with tunable sector duration multipliers, but note that this is an approximation. The other method is basic which does not support sector duration multipliers. See here for more details.

Now, you can call the simulation function and collect the data in a DataFrame:

cil_df = mechafil.run_simple_sim(start_date,
    current_date,
    forecast_length,
    renewal_rate,
    rb_onboard_power,
    fil_plus_rate,
    duration,
    auth_config,
    qap_method='tunable')

cil_df.head()

Please check this Google Colab notebook for a complete example.

You can also run part of the simulation separately. To see more examples, check the available notebooks.

Note for Developers

If you use mechaFIL in expert mode (i.e. don't use run_simple_sim) directly, then you will need to ensure that you setup your data access properly by running the following code

import mechafil.data as mecha_data
path_to_auth_cfg="<UPDATE ME>"
mecha_data.setup_spacecope(path_to_auth_cfg)

References