Skip to content

Commit

Permalink
Remove internal options tests
Browse files Browse the repository at this point in the history 
Mostly broken by the formulation updates, so they don't serve
much purpose.
  • Loading branch information
@simonbowly
simonbowly committed Oct 25, 2024
1 parent c02be79 commit f5e6b89
Showing 1 changed file with 1 addition and 386 deletions.
387 changes: 1 addition & 386 deletions tests/opf/test_internal.py
View file
Original file line number Diff line number Diff line change
@@ -1,398 +1,13 @@
# Tests that solve models using the internal API (more available parameters)

import os
import unittest
from dataclasses import dataclass
from typing import List

import gurobipy as gp
from gurobipy import GRB

from gurobi_optimods.datasets import load_opf_example
from gurobi_optimods.opf import (
converters,
grbformulator_ac,
grbformulator_common,
grbformulator_dc,
grbformulator_iv,
)
from gurobi_optimods.opf.api import _solve_opf_model_internal


@unittest.skipIf(
os.environ.get("CI", "false") == "true",
"Tests expensive internal options; skipped for CI",
)
class TestInternal(unittest.TestCase):
# Test internal options we haven't exposed on the public API yet

def setUp(self):
self.env = gp.Env()
self.case = load_opf_example("case9")

def tearDown(self):
self.env.close()

def test_ac_polar(self):
# Test polar formulation. It's expensive to test all combinations,
# so for now just test the defaults.
self.env.setParam("SolutionLimit", 1)
solution = _solve_opf_model_internal(
self.env,
self.case,
opftype="AC",
polar=True,
useef=False,
usejabr=False,
ivtype="aggressive",
branchswitching=0,
usemipstart=True,
minactivebranches=0.9,
useactivelossineqs=False,
)
self.assertIsNotNone(solution)
self.assertEqual(solution["success"], 1)

def test_ac_settings(self):
# Test all AC solver options with polar=False.
settingslist = [
dict(
opftype="AC",
polar=False,
useef=useef,
usejabr=usejabr,
branchswitching=branchswitching,
usemipstart=usemipstart,
useactivelossineqs=useactivelossineqs,
minactivebranches=minactivebranches,
ivtype="aggressive",
)
for useef in [False, True]
for usejabr in [False, True]
for branchswitching in [0, 1, 2]
for usemipstart in [False, True]
for useactivelossineqs in [False, True]
for minactivebranches in [0, 0.5, 0.95]
]

# Solve opf model for the same case and return a solution.
# The model has to be feasible for every setting combination.
self.env.setParam("SolutionLimit", 1)
for acopf_settings in settingslist:
with self.subTest(acopf_settings):
solution = _solve_opf_model_internal(
self.env,
self.case,
**acopf_settings,
)
self.assertIsNotNone(solution)
self.assertEqual(solution["success"], 1)

def test_dc_settings(self):
# DC has more limited configurations: polar, ef, jabr,
# branchswitching=2, and ivtype are not available
settingslist = [
dict(
opftype="DC",
branchswitching=branchswitching,
usemipstart=usemipstart,
useactivelossineqs=useactivelossineqs,
minactivebranches=minactivebranches,
polar=False,
useef=True,
usejabr=True,
ivtype="aggressive",
)
for branchswitching in [0, 1]
for usemipstart in [False, True]
for useactivelossineqs in [False, True]
for minactivebranches in [0, 0.5, 0.95]
]

# Solve opf model for the same case and return a solution.
# The model has to be feasible for every setting combination.
self.env.setParam("SolutionLimit", 1)
for acopf_settings in settingslist:
with self.subTest(acopf_settings):
solution = _solve_opf_model_internal(
self.env,
self.case,
**acopf_settings,
)
self.assertIsNotNone(solution)
self.assertEqual(solution["success"], 1)

def test_iv_settings(self):
# IV has more limited configurations: polar, ef, jabr,
# branchswitching, and mipstart are not available
settingslist = [
dict(
opftype="IV",
ivtype=ivtype,
useactivelossineqs=useactivelossineqs,
minactivebranches=minactivebranches,
polar=False,
useef=True,
usejabr=True,
branchswitching=0,
usemipstart=True,
)
for ivtype in ["plain", "aggressive"]
for useactivelossineqs in [False, True]
for minactivebranches in [0, 0.5, 0.95]
]

# Solve opf model for the same case and return a solution.
# The model has to be feasible for every setting combination.
self.env.setParam("SolutionLimit", 1)
for acopf_settings in settingslist:
with self.subTest(acopf_settings):
solution = _solve_opf_model_internal(
self.env,
self.case,
**acopf_settings,
)
self.assertIsNotNone(solution)
self.assertEqual(solution["success"], 1)

# test IV formulation
def test_ivopf(self):
# currently all other cases take very long in IV formulation
# solve opf model and return a solution
solution = _solve_opf_model_internal(
self.env,
self.case,
opftype="IV",
ivtype="aggressive",
polar=False,
useef=True,
usejabr=True,
branchswitching=0,
usemipstart=True,
minactivebranches=0.9,
useactivelossineqs=False,
)
# check whether the solution points looks correct
self.assertIsNotNone(solution)
self.assertEqual(solution["success"], 1)
self.assertIsNotNone(solution["f"])
# differences can be quite big because we solve only to 0.1% optimality
self.assertLess(abs(solution["f"] - 5297.0142014), 1e1)
self.assertLess(abs(solution["bus"][0]["Vm"] - 1.0986917), 1e-1)
self.assertLess(abs(solution["gen"][1]["Pg"] - 132.87813), 1e1)
self.assertLess(abs(solution["gen"][2]["Qg"] + 22.802347), 1e1)
self.assertLess(abs(solution["branch"][3]["Pf"] - 95.113306), 1e1)
self.assertLess(abs(solution["branch"][4]["Qt"] + 18.431373), 1e1)


@unittest.skip("varies between versions")
class TestFingerprints(unittest.TestCase):
# Check model fingerprints for some specific cases. Useful while
# refactoring the code, but

def setUp(self):
self.env = gp.Env()
self.data = [
{
"name": "case9",
"case": load_opf_example("case9"),
"dc_fingerprint": 1980532444,
"dc_switching_fingerprint": 1441352667,
"ac_fingerprint": -121125607,
"ac_switching_fingerprint": 1863109018,
"ac_relax_fingerprint": -552798165,
"ac_relax_switching_fingerprint": 1908254825,
"iv_fingerprint": 696846664,
},
{
"name": "caseNY",
"case": load_opf_example("caseNY"),
"dc_fingerprint": 1466917423,
"dc_switching_fingerprint": 1343022114,
"ac_fingerprint": 1587606503,
"ac_switching_fingerprint": 549489140,
"ac_relax_fingerprint": -557544349,
"ac_relax_switching_fingerprint": 1800895367,
"iv_fingerprint": -199906229,
},
]

def tearDown(self):
self.env.close()

def test_dc(self):
for example in self.data:
with self.subTest(name=example["name"]):
with gp.Model(env=self.env) as model:
alldata = converters.convert_case_to_internal_format(
example["case"]
)
alldata.update(
converters.build_internal_settings(
opftype="DC",
branchswitching=False,
usemipstart=False,
useactivelossineqs=False,
minactivebranches=0.0,
polar=False,
useef=True,
usejabr=True,
ivtype="aggressive",
)
)
grbformulator_dc.lpformulator_dc_body(alldata, model)
model.update()
self.assertEqual(model.Fingerprint, example["dc_fingerprint"])

def test_dc_switching(self):
for example in self.data:
with self.subTest(name=example["name"]):
with gp.Model(env=self.env) as model:
alldata = converters.convert_case_to_internal_format(
example["case"]
)
alldata.update(
converters.build_internal_settings(
opftype="DC",
branchswitching=True,
usemipstart=False,
useactivelossineqs=False,
minactivebranches=0.9,
polar=False,
useef=True,
usejabr=True,
ivtype="aggressive",
)
)
grbformulator_dc.lpformulator_dc_body(alldata, model)
model.update()
self.assertEqual(
model.Fingerprint, example["dc_switching_fingerprint"]
)

def test_ac(self):
for example in self.data:
with self.subTest(name=example["name"]):
with gp.Model(env=self.env) as model:
alldata = converters.convert_case_to_internal_format(
example["case"]
)
alldata.update(
converters.build_internal_settings(
opftype="AC",
polar=False,
useef=True,
usejabr=True,
branchswitching=False,
usemipstart=False,
useactivelossineqs=False,
minactivebranches=0.0,
ivtype="aggressive",
)
)
grbformulator_ac.lpformulator_ac_body(alldata, model)
model.update()
self.assertEqual(model.Fingerprint, example["ac_fingerprint"])

def test_ac_switching(self):
for example in self.data:
with self.subTest(name=example["name"]):
with gp.Model(env=self.env) as model:
alldata = converters.convert_case_to_internal_format(
example["case"]
)
alldata.update(
converters.build_internal_settings(
opftype="AC",
polar=False,
useef=True,
usejabr=True,
branchswitching=True,
usemipstart=False,
useactivelossineqs=False,
minactivebranches=0.9,
ivtype="aggressive",
)
)
grbformulator_ac.lpformulator_ac_body(alldata, model)
model.update()
self.assertEqual(
model.Fingerprint, example["ac_switching_fingerprint"]
)

def test_ac_relax(self):
for example in self.data:
with self.subTest(name=example["name"]):
with gp.Model(env=self.env) as model:
alldata = converters.convert_case_to_internal_format(
example["case"]
)
alldata.update(
converters.build_internal_settings(
opftype="AC",
polar=False,
useef=False,
usejabr=True,
branchswitching=False,
usemipstart=False,
useactivelossineqs=False,
minactivebranches=0.0,
ivtype="aggressive",
)
)
grbformulator_ac.lpformulator_ac_body(alldata, model)
model.update()
self.assertEqual(model.Fingerprint, example["ac_relax_fingerprint"])

def test_ac_relax_switching(self):
for example in self.data:
with self.subTest(name=example["name"]):
with gp.Model(env=self.env) as model:
alldata = converters.convert_case_to_internal_format(
example["case"]
)
alldata.update(
converters.build_internal_settings(
opftype="AC",
polar=False,
useef=False,
usejabr=True,
branchswitching=True,
usemipstart=False,
useactivelossineqs=False,
minactivebranches=0.9,
ivtype="aggressive",
)
)
grbformulator_ac.lpformulator_ac_body(alldata, model)
model.update()
self.assertEqual(
model.Fingerprint, example["ac_relax_switching_fingerprint"]
)

def test_iv(self):
for example in self.data:
with self.subTest(name=example["name"]):
with gp.Model(env=self.env) as model:
alldata = converters.convert_case_to_internal_format(
example["case"]
)
alldata.update(
converters.build_internal_settings(
opftype="IV",
ivtype="aggressive",
useactivelossineqs=False,
minactivebranches=0.0,
polar=False,
useef=True,
usejabr=True,
branchswitching=0,
usemipstart=True,
)
)
grbformulator_iv.lpformulator_iv_body(alldata, model)
model.update()
self.assertEqual(model.Fingerprint, example["iv_fingerprint"])
from gurobi_optimods.opf import grbformulator_common


@dataclass
Expand Down

0 comments on commit f5e6b89

Please sign in to comment.