Can I find similar named files ignoring case, dashes, spaces or other characters? - bash

Question

EDIT 2:
lets say I have 2 directories one contains:
/dir1/Test File Name.txt
/dir1/This is anotherfile.txt
/dir1/And-Another File.txt
Directory 2 looks like:
/dir2/test-File_Name.txt
/dir2/test file_Name.txt
/dir2/This Is another file.txt
/dir2/And another_file.txt
How can I find (or match) files that are named similar, in this example file 1 from dir1 would match with file 1 and 2 on dir2 and so on
Trying to do this in bash. Say I have a file named "Test File 1.txt" I want to find any file that is named similar like:
test-file 1.txt
test file 1.txt
Test-file-1.txt
test-file_1.zip
etc etc
I can ignore case with find ./files/ -maxdepth 1 -iname $FILE but don't know how to ignore all the other characters.
Is there a way I can do this in bash?
EDIT:
Sorry, I forgot to mention that I need to iterate on all files, the file name is not always the same, I just used an example.
so it could be named "Test File 1.txt" or it could also be named something completely different "Something Else.txt"
So I want to look for all similar named files using a complete file name as base, but this file name can be different, hope I make more sense.

Answer 1

If Perl is your option, please try the following:
perl -e '
#files1 = glob "dir1/*";
#files2 = glob "dir2/*";
foreach (#files2) {
$f2 = $_;
s#.*/##; # remove directory name
# s#\..*?$##; # remove extension (wrong)
s#\.[^.]*$##; # remove extension (corrected)
s#[\W_]#[\\W_]?#g; # replace non-alphanumric chars
$pat = $_ . "\\.\\w+\$";
# print $pat, "\n"; # uncomment to see the regex pattern
foreach $f1 (#files1) {
if ($f1 =~ m#/$pat#i) {
print "$f1 <=> $f2\n";
}
}
}'
Output:
dir1/And-Another File.txt <=> dir2/And another_file.txt
dir1/Test File Name.txt <=> dir2/test file_Name.txt
dir1/Test File Name.txt <=> dir2/test-File_Name.txt
dir1/This is anotherfile.txt <=> dir2/This Is another file.txt
[Explanations]
The concept is to generate a regex pattern on the fly from a filename
in one directory and match it with the files in the other directory.
File extension is replaced with a pattern which matches it.
Non-alphanumeric character and underscore are replaced with a pattern
which matches them including the case the character is missing so that
anotherfile and another file match.
i option added to the pattern enables case-insensitive match.
You can see the generated regex by uncommenting the noted line.
The possible problem is we can not generate a pattern which matches with
another file from the filename anotherfile. In other words, the
matching is one-directional. A possible workaround is to neglect non-alphanumeric characters and underscores at all in matching. It may result in unexpected overmatching depending on the word and punctuation. We will need to specifically define the similarity to step further.
[Edit]
In order to get the result back to bash variables, please try:
while read -r -d "" line; do
# do something with the bash variable "line"
echo "$line"
done < <(
perl -e '
#files1 = glob "dir1/*";
#files2 = glob "dir2/*";
foreach (#files2) {
$f2 = $_;
s#.*/##; # remove directory name
# s#\..*?$##; # remove extension (wrong)
s#\.[^.]*$##; # remove extension (corrected)
s#[\W_]#[\\W_]?#g; # replace non-alphanumric chars
$pat = $_ . "\\.\\w+\$";
# print $pat, "\n"; # uncomment to see the regex pattern
foreach $f1 (#files1) {
if ($f1 =~ m#/$pat#i) {
push(#result, "$f1 <=> $f2");
# if you want just the list of filenames, comment out the line above
# and uncomment the line below
#push(#result, $f1, $f2);
}
}
}
print join("\0", #result) . "\0";
')
The results is stored in the bash variable line in line by line.
If you want to tweak the output format, please modify the line push(#result, ...).
[EDIT]
Modified to work with the following filename pairs:
"Sample Filename.txt" <=> "Sample Filename (100).txt"
"Sample.Filename.txt" <=> "Sample Filename.txt"
Here's the updated code:
while read -r -d "" line; do
# do something with the bash variable "line"
echo $line
done < <(
perl -e '
#files1 = glob "dir1/*";
#files2 = glob "dir2/*";
foreach (#files2) {
$f2 = $_;
s#.*/##; # remove directory name
s#\.[^.]*$##; # remove extension
s#\s*\(.*?\)##; # remove parenthesis if any
s#\s*\[.*?\]##; # remove square bracket if any
s#[\W_]#[\\W_]?#g; # replace non-alphanumric chars
$pat = $_ . "\\s?((\\(.*?\\))|(\\[.*?\\]))?" . "\\.\\w+\$";
#print $pat . "\n"; # uncomment to see the regex pattern
foreach $f1 (#files1) {
if ($f1 =~ m#/$pat#i) {
push(#result, "$f1 <=> $f2");
# if you want just the list of filenames, comment out the line above
# and uncomment the line below
#push(#result, $f1, $f2);
}
}
}
print join("\0", #result) . "\0";
')