How can I remove files that have an unknown number in them?

Sometimes it's easier to move the "good" files out of the way, and then delete the bad files, and then move the good files back.

If that approach is suitable then this might work

#!/bin/sh

# Temporary directory to hold the files we want to keep
mkdir .keep || exit

for a in $(cat keeplist)
do
  # These are the files we want to keep
  mv body*$a.txt .keep

  # Except this might match negative versions, so remove them
  rm -f .keep/*-$a.txt
done

# Remove the files we don't want
rm body*

# Move the good files back
mv .keep/* .

# Tidy up
rmdir .keep

So, for example, if we start with:

% ls
body-1-2126.txt  body-2-3-123.txt  body-3-3131.txt  body00125.txt  s
body-1-2127.txt  body-3-3129.txt   body00123.txt    fix
body-1-2128.txt  body-3-3130.txt   body00124.txt    keeplist

And then run that script, we end up with

% ls
body-1-2127.txt  body-3-3129.txt  body00123.txt  fix  keeplist  s

In zsh:

$ set -o extendedglob
$ list=(123 127 129)
$ echo rm body(^*(${(~j[|])list})).txt
rm body00124.txt body00125.txt body-1-2126.txt body-1-2128.txt body-3-3130.txt body-3-3131.txt

(remove the echo to actually do it).

The j[|] parameter expansion flag joins the elements of $list with |. With the ~ flag, those are interpreted as a glob operator (the alternation operator as opposed to just a literal |).

So the glob ends up being body(^*(123|127|129)).txt, ^ being the negation extendedglob operator, so matching on filenames that start with body, followed by any string not ending in 123, 127, 129, followed by .txt.

Replace * with (^*-) if you need the extra condition that the part before those numbers must not end with - if they're to be preserved so a file called body-1-1-123.txt for instance would also be removed.

For an even stricter matching, you could even do:

n='((-|)[0-9])' # digit with an optional - sign
echo rm body$~n(#c2)($~n(#c3)~(${(~j[|])list})).txt

Where (#c2) is the repetition operator, and ~ is a except (and-not) operator. $~n is like $n except the contents of $n is interpreted as a pattern as opposed to a literal string (like for the ~ parameter expansion flag above).

So we're matching on body followed by two digits each optionally preceded by a - followed by 3 of those except those that are either one of the members of $list, followed by .txt.

find has a name-matching primitive which can be negated to allow taking actions on files that do not match a name, or which don't match any of a list of names.

Since find's default is to and together multiple operations given on one line, we can write a bash script as follows:

#!/usr/bin/env bash

list=( 123 127 129 )

findcmd="find . -type f $(printf -- ' -not -name \*%s.txt' "${list[@]}")"

bash -v <<< "$findcmd"

(Note: the bash line could also be done as:

printf '%s\n' "$findcmd"
eval $findcmd

)

The output from that script is:

find . -type f  -not -name \*123.txt -not -name \*127.txt -not -name \*129.txt
./body-3-3130.txt
./body00125.txt
./body-1-2126.txt
./body00124.txt
./body-1-2128.txt
./body-3-3131.txt

Here we see two pieces of information: the find command syntax that was built from the array of numbers to keep; and the resulting list of files that do not match any of those numbers.

Inspect the list of filenames closely. Once you have confirmed that you want to delete all of those files, copy the find command syntax and paste it and append the find action directive -exec rm -v {} \; as follows (shown with a backslash-escaped line break for readability):

$ find . -type f  -not -name \*123.txt -not -name \*127.txt -not -name \*129.txt \
    -exec rm -v {} \;
./body-3-3130.txt
./body00125.txt
./body-1-2126.txt
./body00124.txt
./body-1-2128.txt
./body-3-3131.txt

Python. Straightforward way

import os
import glob

num_lst = [123, 127, 129]
num_as_str_set = set(map(str, num_lst))

# If not other files except .txt in directory, listdir() will be enough
#for filename in os.listdir():
for filename in glob.glob("*.txt"):
    #7654321
    #123.txt
    #[-7:-4] -> 123
    if filename[-7:-4] not in num_as_str_set:
        print("remove", filename)
# Uncomment to remove files
#       os.remove(filename)

The same logic on Bash

declare -A hash_map
hash_map=( [123]= [127]= [129]= )

for fn in *.txt; do
    key="${fn: -7:-4}"
    if ! [[ -v hash_map["$key"] ]]; then
        echo "$fn"
#Uncomment to actual remove
#       rm -v "$fn"
    fi  
done

Python. Tricky, may be unoptimal way (benchmark is needed)

import os
from glob import glob
from itertools import chain

num_lst = [123, 127, 129]
s = set(glob("*.txt")) - set(chain(*(glob(f"*{num}.txt") for num in num_lst)))
#Uncomment to remove files
#list(map(os.remove, s))