sort_cols/sort_cols.py

import ast
import pdb

import pandas as pd

SOURCE_DF = "context_summary_group.csv"

df = pd.read_csv("sort_cols/context_summary.csv")
# df.columns = pd.MultiIndex.from_tuples([ast.literal_eval(col) for col in df.columns])
df.columns = [ast.literal_eval(col) for col in df.columns]

names = [("Model", "family"), ("Model", "model"), ("Model", "tag")]
model_names = []
for ind, row in df.iterrows():
    model_names.append(":".join(row[names].to_list()))

df = pd.read_csv(SOURCE_DF)
names = ["family", "model", "tag"]
row_dict = {}
for ind, row in df.iterrows():
    row_dict[":".join(row[names].to_list())] = row

# reconstruct rows
rows = []
for name in model_names:
    rows.append(pd.DataFrame(row_dict[name]).T)
cat_df = pd.concat(rows, axis=0)
cat_df.reset_index(drop=True).to_csv(SOURCE_DF, index=False)