araison/explaining_framework
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

Adding XGWT configs generation

Browse source
This commit is contained in:
araison 2023-02-12 14:12:12 +01:00
parent 27e8a8a4d8
commit 408bab4bc4
3 changed files with 135 additions and 24 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

133
config_gen.py
View file

@ -10,6 +10,13 @@ from torch_geometric.graphgym.utils.io import string_to_python
from explaining_framework.utils.io import (obj_config_to_str, read_yaml, from explaining_framework.utils.io import (obj_config_to_str, read_yaml,
write_yaml) write_yaml)
def chunks(lst, n):
"""Yield successive n-sized chunks from lst."""
for i in range(0, len(lst), n):
yield lst[i : i + n]
# class BaseConfigGenerator(object): # class BaseConfigGenerator(object):
# def __init__(self,dataset_name:str,explainer_name:str, explainer_config:str, model_folder:str): # def __init__(self,dataset_name:str,explainer_name:str, explainer_config:str, model_folder:str):
# self.dataset_name=dataset_name # self.dataset_name=dataset_name
@ -68,6 +75,7 @@ def explainer_conf(explainer: str, **kwargs):
explaining_cfg["depth"] = "all" explaining_cfg["depth"] = "all"
explaining_cfg["score_map_norm"] = kwargs.get("score_map_norm") explaining_cfg["score_map_norm"] = kwargs.get("score_map_norm")
explaining_cfg["interest_map_norm"] = kwargs.get("interest_map_norm") explaining_cfg["interest_map_norm"] = kwargs.get("interest_map_norm")
elif explainer == "EIXGNN": elif explainer == "EIXGNN":
explaining_cfg["L"] = kwargs.get("L") explaining_cfg["L"] = kwargs.get("L")
explaining_cfg["p"] = kwargs.get("p") explaining_cfg["p"] = kwargs.get("p")
@ -77,6 +85,23 @@ def explainer_conf(explainer: str, **kwargs):
explaining_cfg["domain_similarity"] = kwargs.get("domain_similarity") explaining_cfg["domain_similarity"] = kwargs.get("domain_similarity")
explaining_cfg["signal_similarity"] = kwargs.get("signal_similarity") explaining_cfg["signal_similarity"] = kwargs.get("signal_similarity")
explaining_cfg["shapley_value_approx"] = kwargs.get("shapley_value_approx") explaining_cfg["shapley_value_approx"] = kwargs.get("shapley_value_approx")
elif explainer == "XGWT":
explaining_cfg["wav_approx"] = kwargs.get("wav_approx")
explaining_cfg["wav_passband"] = kwargs.get("wav_passband")
explaining_cfg["wav_norm"] = kwargs.get("wav_norm")
explaining_cfg["candidates"] = kwargs.get("candidates")
explaining_cfg["samples"] = kwargs.get("samples")
explaining_cfg["c_proc"] = kwargs.get("c_proc")
explaining_cfg["pred_thres_strat"] = kwargs.get("pred_thres_strat")
explaining_cfg["CI_thres"] = kwargs.get("CI_thres")
explaining_cfg["mix"] = kwargs.get("mix")
explaining_cfg["scales"] = kwargs.get("scales")
explaining_cfg["pred_thres"] = kwargs.get("pred_thres")
explaining_cfg["incl_prob"] = kwargs.get("incl_prob")
explaining_cfg["top_k"] = kwargs.get("top_k")
explaining_cfg["get_DAG"] = kwargs.get("get_DAG")
return explaining_cfg return explaining_cfg
@ -161,6 +186,7 @@ if "__main__" == __name__:
"GNNExplainer", "GNNExplainer",
"EIXGNN", "EIXGNN",
"SCGNN", "SCGNN",
"XGWT",
] ]
for dataset_name in DATASET: for dataset_name in DATASET:
@ -193,6 +219,8 @@ if "__main__" == __name__:
+ ".yaml", + ".yaml",
) )
explainer_path.append(path_explainer) explainer_path.append(path_explainer)
if os.path.exists(path_explainer):
continue
explainer_config.append(config) explainer_config.append(config)
write_yaml(config, path_explainer) write_yaml(config, path_explainer)
if explainer_name == "SCGNN": if explainer_name == "SCGNN":
@ -215,7 +243,89 @@ if "__main__" == __name__:
) )
explainer_path.append(path_explainer) explainer_path.append(path_explainer)
explainer_config.append(config) explainer_config.append(config)
if os.path.exists(path_explainer):
continue
write_yaml(config, path_explainer) write_yaml(config, path_explainer)
if explainer_name == "XGWT":
explainer_config = []
explainer_path = []
for wav_approx in [False]:
for wav_passband in ["heat"]:
for wav_norm in [True]:
for candidates in [10, 15, 30, 50]:
for samples in [10, 25, 50]:
for c_proc in ["auto"]:
for pred_thres_strat in ["regular"]:
for CI_thres in [0.05]:
for mix in ["uniform"]:
for scales in [
[2],
[3],
[5],
[9],
[2, 3, 5],
[2, 3, 5, 9],
[5, 9],
[2, 3],
]:
for pred_thres in [
0.1,
0.25,
0.5,
]:
for incl_prob in [
0.2,
0.4,
0.6,
]:
for top_k in [
2,
5,
10,
]:
for get_DAG in [
False
]:
config = explainer_conf(
"XGWT",
wav_approx=wav_approx,
wav_passband=wav_passband,
wav_norm=wav_norm,
candidates=candidates,
samples=samples,
c_proc=c_proc,
pred_thres_strat=pred_thres_strat,
CI_thres=CI_thres,
mix=mix,
scales=scales,
pred_thres=pred_thres,
incl_prob=incl_prob,
top_k=top_k,
get_DAG=get_DAG,
)
path_explainer = os.path.join(
explainer_folder,
"XGWT_"
+ obj_config_to_str(
config
)
+ ".yaml",
)
explainer_path.append(
path_explainer
)
explainer_config.append(
config
)
if os.path.exists(
path_explainer
):
continue
write_yaml(
config,
path_explainer,
)
for explainer_p, explainer_c in zip(explainer_path, explainer_config): for explainer_p, explainer_c in zip(explainer_path, explainer_config):
explaining_cfg = explaining_conf( explaining_cfg = explaining_conf(
dataset=dataset_name, dataset=dataset_name,
@ -236,16 +346,15 @@ if "__main__" == __name__:
+ f"dataset={dataset_name}-model={model_kind}-explainer={explainer_name}_{obj_config_to_str(explainer_c)}.yaml" + f"dataset={dataset_name}-model={model_kind}-explainer={explainer_name}_{obj_config_to_str(explainer_c)}.yaml"
) )
write_yaml(explaining_cfg, PATH) write_yaml(explaining_cfg, PATH)
os.makedirs(explaining_folder + "/0", exist_ok=True) a = sorted(glob.glob(explaining_folder + "/*.yaml"))
os.makedirs(explaining_folder + "/1", exist_ok=True)
a = glob.glob(explaining_folder + "/*.yaml")
for path in a[:8050]: num_GPU = 4
basename = os.path.basename(path) for i in range(num_GPU):
dirname = os.path.dirname(path) os.makedirs(explaining_folder + f"/{i}", exist_ok=True)
os.rename(path, dirname + "/0/" + basename) split_size = int(len(a) / num_GPU)
data = chunks(a, split_size)
for path in a[8050:]: for i, d in enumerate(data):
basename = os.path.basename(path) for path in d:
dirname = os.path.dirname(path) basename = os.path.basename(path)
os.rename(path, dirname + "/1/" + basename) dirname = os.path.dirname(path)
os.rename(path, dirname + f"/{i}/" + basename)

13
explaining_framework/config/explainer_config/xgwt_config.py
View file

@ -38,17 +38,18 @@ def set_xgwt_cfg(xgwt_cfg):
xgwt_cfg.wav_approx = False xgwt_cfg.wav_approx = False
xgwt_cfg.wav_passband = "heat" xgwt_cfg.wav_passband = "heat"
xgwt_cfg.wav_normalization = True xgwt_cfg.wav_norm = True
xgwt_cfg.num_candidates = 30 xgwt_cfg.candidates = 30
xgwt_cfg.num_samples = 10 xgwt_cfg.samples = 10
xgwt_cfg.c_procedure = "auto" xgwt_cfg.c_proc = "auto"
xgwt_cfg.pred_thres_strat = "regular" xgwt_cfg.pred_thres_strat = "regular"
xgwt_cfg.CI_threshold = 0.05 xgwt_cfg.CI_thres = 0.05
xgwt_cfg.mixing = "uniform" xgwt_cfg.mix = "uniform"
xgwt_cfg.scales = [3] xgwt_cfg.scales = [3]
xgwt_cfg.pred_thres = 0.1 xgwt_cfg.pred_thres = 0.1
xgwt_cfg.incl_prob = 0.4 xgwt_cfg.incl_prob = 0.4
xgwt_cfg.top_k = 5 xgwt_cfg.top_k = 5
xgwt_cfg.get_DAG = False
def assert_cfg(xgwt_cfg): def assert_cfg(xgwt_cfg):

13
explaining_framework/utils/explaining/outline.py
View file

@ -300,16 +300,17 @@ class ExplainingOutline(object):
explaining_algorithm = XGWT( explaining_algorithm = XGWT(
wav_approx=self.explainer_cfg.wav_approx, wav_approx=self.explainer_cfg.wav_approx,
wav_passband=self.explainer_cfg.wav_passband, wav_passband=self.explainer_cfg.wav_passband,
wav_normalization=self.explainer_cfg.wav_normalization, wav_norm=self.explainer_cfg.wav_norm,
num_candidates=self.explainer_cfg.num_candidates, candidates=self.explainer_cfg.candidates,
num_samples=self.explainer_cfg.num_samples, samples=self.explainer_cfg.samples,
c_procedure=self.explainer_cfg.c_procedure, c_proc=self.explainer_cfg.c_proc,
pred_thres_strat=self.explainer_cfg.pred_thres_strat, pred_thres_strat=self.explainer_cfg.pred_thres_strat,
CI_threshold=self.explainer_cfg.CI_threshold, CI_thres=self.explainer_cfg.CI_thres,
mixing=self.explainer_cfg.mixing, mix=self.explainer_cfg.mix,
pred_thres=self.explainer_cfg.pred_thres, pred_thres=self.explainer_cfg.pred_thres,
incl_prob=self.explainer_cfg.incl_prob, incl_prob=self.explainer_cfg.incl_prob,
top_k=self.explainer_cfg.top_k, top_k=self.explainer_cfg.top_k,
get_DAG=self.explainer_cfg.get_DAG,
scales=self.explainer_cfg.scales, scales=self.explainer_cfg.scales,
) )
elif name == "SCGNN": elif name == "SCGNN":