Subclassing dataframe to store modified strings - pandas

Question

I want to subclass a dataframe such that when I do say
df.loc[index,column] = 'green'
the corresponding value specified by [index,column] takes on the value
'background-color: green'
as an example.
Basically I have a lot of custom formatting to do to my dataframes before I send them to excel and since the styler requires storing the styles in dataframes with these strings, I want to create a subclassed dataframe for the styles I am applying to make storing them easier.
I hope my explanation makes sense.
The conditional formatting is no good for me, as a lot of the colors I am getting are being pulled from existing html tables that I have to do a lot of web-scraping to obtain.
Edit: The way the pandas styler object works is that you take a function, which takes the df you wish to style, as an input, and it returns with another dataframe which same index and columns with the style options.
For example I have dataframe
df
A B C D
first second
A1 2 1 3 2 4
A2 2 5 2 2 3
A3 1 3 4 7 6
A4 4 4 3 3 12
A5 6 2 4 4 7
for the values in df that equal two, I want to make their background color blue when exporting this df to excel.
So I'll make a function
def make_2_blue(df):
background_color = pd.DataFrame('',columns=df.columns,index=df.index)
background_color[df == 2] = 'background-color: blue'
return
Then when I want to export this to excel I'd do
df.styler.apply(make_2_blue,axis=None).to_excel('blah.xlsx',engine='openpyxl')
Now in my example I am webscraping html to get the individual cells background color.
What I want to do is subclass a dataframe to make applying these formats easier.
Essentially I would like create some styler-dataframe where I specify the style and then I when I access the individual values in the df it will take in my string and apply the desired formatting.
So lets say I make a background-color dataframe class. I'll make an instance of it called background_color_df.
background_color_df = pd.DataFrame('',index=df.index,columns=df.columns)
>>> background_color_df
A B C \
first second
A1 2 background-color: background-color: background-color:
A2 2 background-color: background-color: background-color:
A3 1 background-color: background-color: background-color:
A4 4 background-color: background-color: background-color:
A5 6 background-color: background-color: background-color:
D
first second
A1 2 background-color:
A2 2 background-color:
A3 1 background-color:
A4 4 background-color:
A5 6 background-color:
Now I go to assign the color blue at a specific index and column
background_color_df.loc['A1','A'] = 'blue'
background_color_df
A B C \
first second
A1 2 background-color: blue background-color: background-color:
A2 2 background-color: background-color: background-color:
A3 1 background-color: background-color: background-color:
A4 4 background-color: background-color: background-color:
A5 6 background-color: background-color: background-color:
D
first second
A1 2 background-color:
A2 2 background-color:
A3 1 background-color:
A4 4 background-color:
A5 6 background-color:

Answer 1

If I got your question right, you have data like this:
color foo
background green bar
foreground red baz
That is, a dataframe with attributes for some kind of category like foreground and background and you need them combined like:
color foo
background background-color: green background-foo: bar
foreground foreground-color: red foreground-foo: baz
Please ignore the column names and index names here, as you could change them easily.
If that is, what you need, you don't need to subclass DataFrame, which will surely be a lot of work. You could just postprocess your dataframe so it looks like what you want. E.g. by doing:
# example data
df= pd.DataFrame({'color': ['green', 'red'], 'foo': ['bar', 'baz']}, index=['background', 'foreground'])
columns_to_transform= ['color', 'foo'] # maybe even: df.columns if you need to transform all columns
for col in columns_to_transform:
df[col]= df.index + '-' + col + ': ' + df[col]