There's one hell of a lot of redundancy in the data, which is part of the program. Compress it. Even trivial encoding like RLE reduces it from 2451 characters to 525 characters. The encoding here is simply a character followed by digits for the number of times it appears.

 75'6
1 71'11
1 67'15
1 64'17
1 61'19
1 59'20
1 57'21
1 37~13 5'7 6'9
1 34~19 3'4 8'8
1 31~24 12'7
1 28~29 9'7
1 26~32 7'6
1 25~34 5'6
1 24~36 3'5
1 23~38 1'5
1 22~37 2'4
1 21~36 2'4
1 21~35 2'3
1 20~34 2'3 2~2
1 20~32 2'3 2~4
1 20~30 2'3 2~6
1 20~28 2'3 2~8
1 20~26 2'3 2~10
1 20~24 2'3 2~11
1 20~22 2'3 2~13
1 18'2~19 2'3 2~15
1 16'5~16 2'3 2~16
1 14'6 2~13 2'3 2~17
1 13'5 5~10 2'3 2~18
1 11'6 7~6 2'3 2~20
1 9'7 10~2 2'3 2~21
1 8'7 12'4 2~21
1 6'9 8'5 2~22
1 5'10 4'9 4~17
1 4'22 9~10
1 3'20
1 2'18
1 1'16
1'14
1'11
1 2'6
[download]

This can be trivially improved by:

• Optimising single characters to have no trailing number;
• encoding numbers into single characters.

That latter is moderately interesting. After getting rid of 1s, we have to encode numbers in the range 2 to 71, and you have four unique characters in your data. So we need 74 unique characters. Represent your four data characters by the first four normally printable ASCII characters - \t, \n, space and pling. Then represent the numbers 2 to 71 with ASCII characters 34 to 103.