Getting on average 10 error messages like:

[61.0, 61.514828897338404, 62.02965779467681, 62.54448669201521, 63.05931558935361, 63.57414448669201, 64.08897338403042, 64.60380228136881, 65.11863117870722, 65.63346007604562, 66.14828897338403, 66.66311787072243, 67.17794676806083, 67.69277566539924,
.....

732.3368821292777, 732.851711026616, 733.3665399239544, 733.8813688212929, 734.3961977186312, 734.9110266159696, 735.4258555133081, 735.9406844106464, 736.4555133079848, 736.9703422053233, 737.4851711026616, 738.0]

this is probably the result of my code, yet not what was expected

How to get closer to what is expected ?
is it really necessary to analyze this output to find out about the input ?

The previous comments on precision seem not to be helpful, anyway.

My code works well with all delivered tests,
yet the attempt-tests go wrong

to debug, could s.o. give the attempt-input or some routine that creates
random inputs please ?

one of the following should be sufficient. The wronger the better

{'pos': [1, 4, 8, 10, 14, 16, 19, 21],
'peaks': [12, 15, 2, 10, 13, 13, 19, 15]}
should equal
{'pos': [1, 4, 6, 8, 10, 14, 16, 19, 21],
'peaks': [12, 15, 0, 2, 10, 13, 13, 19, 15]}

{'pos': [6, 11, 14, 18, 20, 22],
'peaks': [18, 16, 18, 13, 10, 16]}
should equal
{'pos': [6, 11, 14, 18, 20, 22, 25],
'peaks': [18, 16, 18, 13, 10, 16, 10]}

{'pos': [2], 'peaks': [15]}
should equal
{'pos': [2, 5, 9, 12, 14],
'peaks': [15, 18, 17, 19, 20]}

{'pos': [2, 6, 10, 13],
'peaks': [16, 13, 6, 4]}
should equal
{'pos': [2, 6, 10, 13, 16, 20, 22],
'peaks': [16, 13, 6, 4, 13, 18, 17]}

{'pos': [1], 'peaks': [5]}
should equal
{'pos': [1, 5, 9, 11, 13, 18],
'peaks': [5, 9, 17, -1, 12, 16]}

{'pos': [1, 5, 9], 'peaks': [20, 20, 20]}
should equal {'pos': [1, 5, 9, 14, 16, 18],
'peaks': [20, 20, 20, 16, 20, 10]}

{'pos': [1, 5, 8, 11, 13],
'peaks': [10, -1, 16, 11, 3]}
should equal {'pos': [1, 5, 8, 11, 13, 16, 19],
'peaks': [10, -1, 16, 11, 3, 20, 14]}

{'pos': [1, 4, 7, 9, 13, 15],
'peaks': [20, 9, 19, 11, 13, 18]}
should equal
{'pos': [1, 4, 7, 9, 13, 15, 18, 21, 24, 26, 28],
'peaks': [20, 9, 19, 11, 13, 18, 11, 15, 19, 5, 15]}
{'pos': [1, 3, 7],
'peaks': [12, 11, 10]}
should equal
{'pos': [1, 3, 7, 12, 15, 17],
'peaks': [12, 11, 10, 15, 10, 15]}

{'pos': [5, 7, 9, 11],
'peaks': [17, 13, 11, 11]}
should equal
{'pos': [5, 7, 9, 11, 15, 20],
'peaks': [17, 13, 11, 11, 16, 17]}

{'pos': [2, 4, 6],
'peaks': [16, 12, 3]}
should equal {'pos': [2, 4, 6, 11],
'peaks': [16, 12, 3, 20]}

Please give a hint conc. "multiply"

Should
[1, 5],
[2],
[-1, -3]

be treated as

[1, 5],

[2, 0],
[-1, -3]
?

"smallest" gives the imagination that there is some choice possible.
This would be in contrast to simple vector multiplication, right ?

I do notunderstand the instructions
Moreover:
Is this

[1, 5],
[2],
[-1, -3]

to be underrstood as:

[1, 5],
[2, 0 ],
[-1, -3]

?

multiply the contents of each nested array

appears to me like:

a1*b1 + a2 * b2 + a3 *b3...

but may as well be:

a1b1c1 + a2b2c2
....

But what menas:
the "smallest" total ?

in vector multiplication there is always a unique solution

Thanx, anyway my approach was more stupid. After changing print(changedstring) to return(changedstring) it works

Its a bit strange, my code yields the desired test results, the ouput is a string (checked by type())
but it is rejected. What went wromg ?
Am I allowed to post the listing here ?
Thanx

Test Results:
Log
20 8 5 19 21 14 19 5 20 19 5 20 19 1 20 20 23 5 12 22 5 15 3 12 15 3 11
None should equal '20 8 5 19 21 14 19 5 20 19 5 20 19 1 20 20 23 5 12 22 5 15 3 12 15 3 11'
Log
20 8 5 14 1 18 23 8 1 12 2 1 3 15 14 19 1 20 13 9 4 14 9 7 8 20
None should equal '20 8 5 14 1 18 23 8 1 12 2 1 3 15 14 19 1 20 13 9 4 14 9 7 8 20'
None should equal ''

There are many more tests than ust "Welcome" as given in the instructions. Where to find them ?