OpenRank SDK

EigenTrust by OpenRank

EigenTrust SDk is a python wrapper for https://github.com/Karma3Labs/go-eigentrust

Demo

Notebook: https://app.hex.tech/f7590f16-5e23-4925-9e84-6656844ee649/hex/2a1f9e4b-76e2-49fe-979a-750043589cbb/draft/logic

Video: https://www.loom.com/share/483a3a1fab244207924595d98c056c70?sid=3112f668-3970-474d-9a3f-ed925231eb6e

Input

You will need local trust and optionally pre-trust. If pre-trust is not specified, each peer will have an equal weight. Both can be specified using a CSV or an array of a dict with i, j, and v.

Sample local trust dict variable

localtrust = [{
  "i": "ek",
  "j": "sd",
  "v": 100
}, {
  "i": "vm",
  "j": "sd",
  "v": 100
}, {
  "i": "ek",
  "j": "sd",
  "v": 75
}]

Sample local trust (lt.csv):

from,to,value
ek,sd,100
vm,sd,100
ek,vm,75

Here we have 3 peers: EK, VM, and SD. Both EK and VM trust SD by 100. EK also trusts VM, by 3/4 of how much he trusts SD.

Sample pre trust dict variable

pre = [{
  "i": "ek",
  "v": 50
}, {
  "i": "vm",
  "v": 100
}]

Sample pre-trust (pt.csv):

peer_id,value
ek,50
vm,100

Here, both EK and VM are pre-trusted by the network (a priori trust). VM is trusted twice as much as EK.

Running

To run EigenTrust using the above input:

from openrank_sdk import EigenTrust

api_key = 'your_api_key'
a = EigenTrust(api_key=api_key)

# Option A - Use local variable
a.run_eigentrust(localtrust)
## run with pretrust you've defined rather than the one distributed equally
a.run_eigentrust(localtrust, pretrust)

# Option B - Use CSV
a.run_eigentrust_from_csv("./lt.csv")
## run with pretrust you've defined rather than the one distributed equally
a.run_eigentrust_from_csv("./lt.csv", "./pt.csv")

Outputs:

[
  {'i': 'vm', 'v': 0.485969387755102},
  {'i': 'sd', 'v': 0.2933673469387755},
  {'i': 'ek', 'v': 0.22066326530612243}
]

Here, the EigenTrust algorithm distributed the network's trust onto the 3 peers:

EK gets 22.0%
SD gets 29.3%
VM gets 48.5%

Appendix

Tweaking Alpha

The pre-trust input defines the relative ratio by which the network distributes its a priori trust onto trustworthy peers, in this case EK and VM.

You can also tweak the overall absolute strength of the pre-trust. This parameter, named alpha, represents the portion of the EigenTrust output taken from the pre-trust. For example, with alpha of 0.2, the EigenTrust output is a blend of 20% pre-trust and 80% peer-to-peer trust.

The CLI default for alpha is 0.5 (50%). If you re-run EigenTrust using a lower alpha of only 0.01 (1%):

from openrank_sdk import EigenTrust

api_key = 'your_api_key'
a = EigenTrust(api_key=api_key, alpha=0.01)

a.run_eigentrust(localtrust, pretrust)

We get a different result:

[
  {'address': 'vm', 'score': 0.39451931175281096},
  {'address': 'sd', 'score': 0.4401132971693594},
  {'address': 'ek', 'score': 0.16536739107782936}
]

EK and VM's trust shares got lower (EK 21.7% ⇒ 16.5%, VM 48.1% ⇒ 39.5%), whereas SD's trust share soared (30.2% ⇒ 44%) despite not being pre-trusted. This is because, with only 1% pre-trust level, the peer-to-peer trust opinions (where SD is trusted by both EK and VM) make up for a much larger portion of trust.