I am looking for a solution to read previous two rows where C_1=104 and next two rows where C_1=105 includes these 2 rows as well. Below are the different dataframe inputs.
The challange I am facing is-

Cond 1 – sometimes C_1=104 is first row(refer df2) in dataframe so there wouldn’t be any previous rows.

Cond 2 – sometime C_1=104 is not there but C_1=105 is first row(refer df3).

Cond 3- sometimes Cond-1 & 2 can be other way which means C_1=105 can be last row(refer df4) OR C_1=104 can we last(C_1=105 will not be there).

df1 = pd.read_csv(StringIO("""
      datetime        C_1 C_2  C_3  C_4
"18/06/2023 3:51:50"  101 128  89    4  
"18/06/2023 3:51:52"  102 128  61    4  
"18/06/2023 3:51:53"  103 130  65    0  
"18/06/2023 3:51:53"  104 132  89    0  
"18/06/2023 3:51:54"  105 134  66    2  
"18/06/2023 3:51:55"  106 134  89    0  
"18/06/2023 3:51:56"  107 156  77    0  
"18/06/2023 3:51:56"  108 112  89    0  
"18/06/2023 3:51:57"  109 109  90    0  
"""), sep="s+")

df-1 Output-

      datetime        C_1 C_2  C_3  C_4
"18/06/2023 3:51:52"  102 128  61    4 
"18/06/2023 3:51:53"  103 130  65    0  
"18/06/2023 3:51:53"  104 132  89    0  
"18/06/2023 3:51:54"  105 134  66    2  
"18/06/2023 3:51:55"  106 134  89    0 
"18/06/2023 3:51:56"  107 156  77    0

df2 = pd.read_csv(StringIO("""
      datetime        C_1 C_2  C_3  C_4  
"18/06/2023 3:51:53"  104 132  89    0  
"18/06/2023 3:51:54"  105 134  66    2  
"18/06/2023 3:51:55"  106 134  89    0  
"18/06/2023 3:51:56"  107 156  77    0  
"18/06/2023 3:51:56"  108 112  89    0  
"18/06/2023 3:51:57"  109 109  90    0  
"""), sep="s+")

df-2 Output-

      datetime        C_1 C_2  C_3  C_4
"18/06/2023 3:51:53"  104 132  89    0  
"18/06/2023 3:51:54"  105 134  66    2  
"18/06/2023 3:51:55"  106 134  89    0  
"18/06/2023 3:51:56"  107 156  77    0

df3 = pd.read_csv(StringIO("""
      datetime        C_1 C_2  C_3  C_4   
"18/06/2023 3:51:54"  105 134  66    2  
"18/06/2023 3:51:55"  106 134  89    0  
"18/06/2023 3:51:56"  107 156  77    0  
"18/06/2023 3:51:56"  108 112  89    0  
"18/06/2023 3:51:57"  109 109  90    0  
"""), sep="s+")

df-3 Output-

      datetime        C_1 C_2  C_3  C_4  
"18/06/2023 3:51:54"  105 134  66    2  
"18/06/2023 3:51:55"  106 134  89    0  
"18/06/2023 3:51:56"  107 156  77    0

df4 = pd.read_csv(StringIO("""
      datetime        C_1 C_2  C_3  C_4
"18/06/2023 3:51:50"  101 128  89    4  
"18/06/2023 3:51:52"  102 128  61    4  
"18/06/2023 3:51:53"  103 130  65    0  
"18/06/2023 3:51:53"  104 132  89    0  
"18/06/2023 3:51:54"  105 134  66    2  
"""), sep="s+")

df-4 Output-

      datetime        C_1 C_2  C_3  C_4
"18/06/2023 3:51:52"  102 128  61    4 
"18/06/2023 3:51:53"  103 130  65    0  
"18/06/2023 3:51:53"  104 132  89    0  
"18/06/2023 3:51:54"  105 134  66    2

I was trying using df[g.cumcount().lt(N) | g.cumcount(ascending=False).lt(N)] where N=2 and also trying using windows function but unable to find the solution.