df_j_tmp = df_tmp_2
# +
# df_i
# +
if using_SOAP:
df_to_use = df_j_tmp
else:
df_to_use = df_i
df_j = simplify_df_features_targets(
df_to_use,
# df_i,
# df_j_tmp,
target_ads="o",
# feature_ads="o",
feature_ads="o",
)
df_format = df_features_targets[(
"format",
"color",
"stoich",
)]
# -
# ### Removing columns with no variance
# +
df_j_info = df_j.describe()
new_cols.append(("features", "o", col_i))
df_SOAP_i.columns = pd.MultiIndex.from_tuples(new_cols)
df_tmp = df_i
df_tmp_2 = pd.concat([
df_tmp,
df_SOAP_i,
], axis=1)
df_j_tmp = df_tmp_2
# +
df_j = simplify_df_features_targets(
df_j_tmp,
target_ads="o",
feature_ads="o",
)
df_format = df_features_targets[(
"format",
"color",
"stoich",
)]
# -
# ### Removing columns with no variance
# +
df_j_info = df_j.describe()
# y_array = df_comb["targets"]["g_oh"].tolist()
# trace = go.Scatter(
# x=x_array,
# y=y_array,
# mode="markers",
# )
# data = [trace]
# fig = go.Figure(data=data)
# fig.show()
# +
df_j = simplify_df_features_targets(
df_comb,
target_ads=target_ads_i,
feature_ads=feature_ads_i,
)
df_format = df_features_targets[(
"format",
"color",
"stoich",
)]
# -
# ### Creating pre-DFT features
# +
# pre_dft_features = True
pre_dft_features = False