df_ads = get_df_ads()
df_ads = df_ads.set_index(["compenv", "slab_id", "active_site", ], drop=False)
df_features = get_df_features()
df_features.index = df_features.index.droplevel(level=5)
df_jobs = get_df_jobs()
df_slab = get_df_slab()
df_jobs_data = get_df_jobs_data()
df_jobs_data["rerun_from_oh"] = df_jobs_data["rerun_from_oh"].fillna(value=False)
df_dft = get_df_dft()
df_job_ids = get_df_job_ids()
df_job_ids = df_job_ids.set_index("job_id")
df_job_ids = df_job_ids[~df_job_ids.index.duplicated(keep='first')]
# -
feature_cols = df_features["features"].columns.tolist()
# + active=""
#
#
# -
# ### Collecting other relevent data columns from various data objects
# +
# #########################################################
from local_methods import get_num_revs_for_group
# -
from methods import isnotebook
isnotebook_i = isnotebook()
if isnotebook_i:
from tqdm.notebook import tqdm
verbose = True
else:
from tqdm import tqdm
verbose = False
# # Read Data
# +
df_job_ids = get_df_job_ids()
compenv_local = os.environ["COMPENV"]
# -
# # Parse `df_jobs_base` files
# +
root_dir = os.path.join(os.environ["PROJ_irox_oer"],
"dft_workflow/job_processing", "out_data")
compenvs = [
"nersc",
"sherlock",
"slac",
"wsl",
def process_group_magmom_comp(
name=None,
group=None,
write_atoms_objects=False,
verbose=False,
):
"""
"""
#| - process_group_magmom_comp
# #####################################################
group_w_o = group
# #####################################################
out_dict = dict()
out_dict["df_magmoms_comp"] = None
out_dict["good_triplet_comb"] = None
out_dict["job_ids"] = None
# out_dict[""] =
job_ids_list = list(set(group.job_id_max.tolist()))
#| - Reading data
# #########################################################
df_jobs = get_df_jobs()
# #########################################################
df_atoms_sorted_ind = get_df_atoms_sorted_ind()
df_atoms_sorted_ind = df_atoms_sorted_ind.set_index("job_id")
# #########################################################
df_job_ids = get_df_job_ids()
df_job_ids = df_job_ids.set_index("job_id")
from methods import read_magmom_comp_data
assert name != None, "Must pass name to read previous data"
magmom_comp_data_prev = read_magmom_comp_data(name=name)
if magmom_comp_data_prev is not None:
pair_wise_magmom_comp_data_prev = \
magmom_comp_data_prev["pair_wise_magmom_comp_data"]
#__|
if write_atoms_objects:
#| - Write atoms objects
df_i = pd.concat([
df_job_ids, df_atoms_sorted_ind.loc[group_w_o.job_id_max.tolist()]
],
axis=1,
join="inner")
# #########################################################
df_index_i = group_w_o.index.to_frame()
compenv_i = df_index_i.compenv.unique()[0]
slab_id_i = df_index_i.slab_id.unique()[0]
active_sites = [
i for i in df_index_i.active_site.unique() if i != "NaN"
]
active_site_i = active_sites[0]
folder_name = compenv_i + "__" + slab_id_i + "__" + str(
int(active_site_i))
# #########################################################
for job_id_i, row_i in df_i.iterrows():
tmp = 42
job_id = row_i.name
atoms = row_i.atoms_sorted_good
ads = row_i.ads
file_name = ads + "_" + job_id + ".traj"
print("Is this saving to the right place d9sf")
root_file_path = os.path.join("__temp__", folder_name)
print(os.getcwd(), root_file_path)
if not os.path.exists(root_file_path):
os.makedirs(root_file_path)
file_path = os.path.join(root_file_path, file_name)
atoms.write(file_path)
#__|
# #####################################################
#| - Getting good triplet combinations
all_triplet_comb = list(
itertools.combinations(group_w_o.job_id_max.tolist(), 3))
good_triplet_comb = []
for tri_i in all_triplet_comb:
df_jobs_i = df_jobs.loc[list(tri_i)]
# Triplet must not contain duplicate ads
# Must strictly be a *O, *OH, and *bare triplet
ads_freq_dict = CountFrequency(df_jobs_i.ads.tolist())
tmp_list = list(ads_freq_dict.values())
any_repeat_ads = [True if i > 1 else False for i in tmp_list]
if not any(any_repeat_ads):
good_triplet_comb.append(tri_i)
#__|
# #####################################################
#| - MAIN LOOP
if verbose:
print("Number of viable triplet combinations:", len(good_triplet_comb))
data_dict_list = []
pair_wise_magmom_comp_data = dict()
for tri_i in good_triplet_comb:
#| - Process triplets
data_dict_i = dict()
if verbose:
print("tri_i:", tri_i)
all_pairs = list(itertools.combinations(tri_i, 2))
df_jobs_i = df_jobs.loc[list(tri_i)]
sum_norm_abs_magmom_diff = 0.
for pair_i in all_pairs:
# if pair_i in list(pair_wise_magmom_comp_data_prev.keys()):
if (magmom_comp_data_prev is not None) and \
(pair_i in list(pair_wise_magmom_comp_data_prev.keys())):
magmom_data_out = pair_wise_magmom_comp_data_prev[pair_i]
else:
# print("Need to run manually")
# print("pair_i:", pair_i)
#| - Process pairs
row_jobs_0 = df_jobs.loc[pair_i[0]]
row_jobs_1 = df_jobs.loc[pair_i[1]]
ads_0 = row_jobs_0.ads
ads_1 = row_jobs_1.ads
# #############################################
if set([ads_0, ads_1]) == set(["o", "oh"]):
job_id_0 = df_jobs_i[df_jobs_i.ads == "o"].iloc[0].job_id
job_id_1 = df_jobs_i[df_jobs_i.ads == "oh"].iloc[0].job_id
elif set([ads_0, ads_1]) == set(["o", "bare"]):
job_id_0 = df_jobs_i[df_jobs_i.ads ==
"bare"].iloc[0].job_id
job_id_1 = df_jobs_i[df_jobs_i.ads == "o"].iloc[0].job_id
elif set([ads_0, ads_1]) == set(["oh", "bare"]):
job_id_0 = df_jobs_i[df_jobs_i.ads ==
"bare"].iloc[0].job_id
job_id_1 = df_jobs_i[df_jobs_i.ads == "oh"].iloc[0].job_id
else:
print("Woops something went wrong here")
# #############################################
row_atoms_i = df_atoms_sorted_ind.loc[job_id_0]
# #############################################
atoms_0 = row_atoms_i.atoms_sorted_good
magmoms_sorted_good_0 = row_atoms_i.magmoms_sorted_good
was_sorted_0 = row_atoms_i.was_sorted
# #############################################
# #############################################
row_atoms_i = df_atoms_sorted_ind.loc[job_id_1]
# #############################################
atoms_1 = row_atoms_i.atoms_sorted_good
magmoms_sorted_good_1 = row_atoms_i.magmoms_sorted_good
was_sorted_1 = row_atoms_i.was_sorted
# #############################################
# #############################################
magmom_data_out = get_magmom_diff_data(
ads_atoms=atoms_1,
slab_atoms=atoms_0,
ads_magmoms=magmoms_sorted_good_1,
slab_magmoms=magmoms_sorted_good_0,
)
#__|
pair_wise_magmom_comp_data[pair_i] = magmom_data_out
tot_abs_magmom_diff = magmom_data_out["tot_abs_magmom_diff"]
norm_abs_magmom_diff = magmom_data_out["norm_abs_magmom_diff"]
if verbose:
print(" ",
"pair_i: ",
pair_i,
": ",
np.round(norm_abs_magmom_diff, 3),
sep="")
sum_norm_abs_magmom_diff += norm_abs_magmom_diff
# #################################################
data_dict_i["job_ids_tri"] = set(tri_i)
data_dict_i["sum_norm_abs_magmom_diff"] = sum_norm_abs_magmom_diff
# #################################################
data_dict_list.append(data_dict_i)
# #################################################
#__|
#__|
# #####################################################
df_magmoms_i = pd.DataFrame(data_dict_list)
# #####################################################
out_dict["df_magmoms_comp"] = df_magmoms_i
out_dict["good_triplet_comb"] = good_triplet_comb
out_dict["pair_wise_magmom_comp_data"] = pair_wise_magmom_comp_data
out_dict["job_ids"] = job_ids_list
# #####################################################
return (out_dict)
def old_get_ORR_PLT():
"""
"""
#| - get_ORR_PLT
# #########################################################
# df_ads = get_df_ads()
# df_ads = df_ads[~df_ads.g_oh.isna()]
# df_m = df_ads
# #########################################################
df_dft = get_df_dft()
# #########################################################
df_job_ids = get_df_job_ids()
# #########################################################
df_features_targets = get_df_features_targets()
smart_format_dict = [
[{"stoich": "AB2"}, {"color2": "black"}],
[{"stoich": "AB3"}, {"color2": "grey"}],
]
ORR_PLT = ORR_Free_E_Plot(
free_energy_df=None,
state_title="ads",
free_e_title="ads_g",
smart_format=smart_format_dict,
color_list=None,
rxn_type="OER")
# # df_m.g_ooh = 1.16 * df_m.g_oh + 2.8
# df_m["g_ooh"] = df_m.g_oh + 2.8
# df_m = df_m.set_index(["compenv", "slab_id", ], drop=False)
new_col = (df_features_targets["targets"]["g_oh"] + 2.8)
new_col.name = ("targets", "g_ooh", "", )
df_features_targets = pd.concat([
new_col,
df_features_targets,
], axis=1)
paths_dict = dict()
# for name_i, row_i in df_m.iterrows():
for name_i, row_i in df_features_targets.iterrows():
#| - Loop through data and add to ORR_PLT
# #####################################################
g_o_i = row_i[("targets", "g_o", "", )]
g_oh_i = row_i[("targets", "g_oh", "", )]
g_ooh_i = row_i[("targets", "g_ooh", "", )]
slab_id_i = row_i[("data", "slab_id", "")]
active_site_i = row_i[("data", "active_site", "")]
job_id_o_i = row_i[("data", "job_id_o", "")]
job_id_oh_i = row_i[("data", "job_id_oh", "")]
# #####################################################
# #####################################################
df_job_ids_i = df_job_ids[df_job_ids.slab_id == slab_id_i]
bulk_ids = df_job_ids_i.bulk_id.unique()
mess_i = "SIJFIDSIFJIDSJIf"
assert len(bulk_ids) == 1, mess_i
bulk_id_i = bulk_ids[0]
# #########################################################
row_dft_i = df_dft.loc[bulk_id_i]
# #########################################################
stoich_i = row_dft_i.stoich
# #########################################################
data_dict_list = [
{"ads_g": g_o_i, "ads": "o", },
{"ads_g": g_oh_i, "ads": "oh", },
{"ads_g": g_ooh_i, "ads": "ooh", },
{"ads_g": 0., "ads": "bulk", },
]
df_i = pd.DataFrame(data_dict_list)
df_i["stoich"] = stoich_i
prop_name_list = [
"stoich",
]
# #########################################################
# name_i = "IDSJFISDf"
name_i = slab_id_i + "__" + str(int(active_site_i))
ORR_PLT.add_series(
df_i,
plot_mode="all",
overpotential_type="OER",
property_key_list=prop_name_list,
add_overpot=False,
name_i=name_i,
)
#__|
return(ORR_PLT)