Update README.md

This commit is contained in:
Marcell Vazquez-Chanlatte 2018-12-19 11:21:27 -08:00 committed by GitHub
parent ba56886a4d
commit 74dc779b2d
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23

103
README.md
View file

@ -32,7 +32,58 @@ To begin, we import `mtl`.
import mtl
```
## Propositional logic (using parse api)
There are **two** APIs for interacting with the `mtl` module. Namely, one can specify the MTL expression using:
1. python operators.
2. string + the parse api.
We begin with the Python Operator API:
## Python Operator API
### Propositional logic (using python syntax)
```python
a, b = mtl.parse('a'), mtl.parse('b')
phi0 = ~a
phi1 = a & b
phi2 = a | b
# TODO: add
phi3 = a ^ b
phi4 = a.iff(b)
phi5 = a.implies(b)
```
### Modal Logic (using python syntax)
```python
a, b = mtl.parse('a'), mtl.parse('b')
# Eventually `a` will hold.
phi1 = a.eventually()
# `a & b` will always hold.
phi2 = (a & b).always()
# `a` until `b`
phi3 = a.until()
# `a` weak until `b`
phi4 = a.weak_until(b)
# Whenever `a` holds, then `b` holds in the next two time units.
phi5 = (a.implies(b.eventually(lo=0, hi=2))).always()
# We also support timed until.
phi6 = a.timed_until(b, lo=0, hi=2)
# `a` holds in two time steps.
phi7 = a >> 2
```
## String based API
### Propositional logic (parse api)
```python
# - Lowercase strings denote atomic predicates.
phi0 = mtl.parse('atomicpred')
@ -49,20 +100,7 @@ phi6 = mtl.parse('~a')
phi7 = mtl.parse('~(a)')
```
## Propositional logic (using python syntax)
```python
a, b = mtl.parse('a'), mtl.parse('b')
phi0 = ~a
phi1 = a & b
phi2 = a | b
# TODO: add
phi3 = a ^ b
phi4 = a.iff(b)
phi5 = a.implies(b)
```
## Modal Logic (parser api)
### Modal Logic (parser api)
```python
# Eventually `x` will hold.
@ -90,34 +128,15 @@ phi6 = mtl.parse('(a U[0, 2] b)')
phi7 = mtl.parse('XX a')
```
## Modal Logic (using python syntax)
```python
a, b = mtl.parse('a'), mtl.parse('b')
# Eventually `a` will hold.
phi1 = a.eventually()
# `a & b` will always hold.
phi2 = (a & b).always()
# `a` until `b`
phi3 = a.until()
# `a` weak until `b`
phi4 = a.weak_until(b)
# Whenever `a` holds, then `b` holds in the next two time units.
phi5 = (a.implies(b.eventually(lo=0, hi=2))).always()
# We also support timed until.
phi6 = a.timed_until(b, lo=0, hi=2)
# `a` holds in two time steps.
phi7 = a >> 2
```
## Boolean Evaluation
Given a property `phi`, one can evaluate is a timeseries satisifies `phi`. Time Series can either be
defined using a dictionary mapping atomic predicate names to lists of (`time`, `val`) pairs **or** using
the [DiscreteSignals](https://github.com/mvcisback/DiscreteSignals) API (used internally).
There are two types of evaluation. One uses the boolean semantics of MTL and the other uses Signal Temporal Logic like semantics.
```python
# Assumes piece wise constant interpolation.
data = {
@ -139,7 +158,7 @@ print(phi(data, dt=0.2, quantitative=False))
# output: True
```
## Quantitative Evaluate
## Quantitative Evaluate (Signal Temporal Logic)
```python
# Assumes piece wise constant interpolation.
data = {