Using regex with Exiftool

[philbond87]

I don't think this is possible with Exiftool but I want to be sure – Exiftool is so comprehensive I don't want to assume anything.

What I want to do is take a nested folder structure of image files and create string using parts of each filename (using regex) and write that created string to a tag.
I know that Exiftool will crawl recursively but I would be surprised if could create a unique string for each file based on the filename, then allow me to apply it to a tag of that file.

The reason I ask is that I have already written an application that does this – digs through the directories, creates the unique string then makes a call to Exiftool to apply that to the tag. With a great number of large files this can take quite a long time. If it could all be done within a single Exiftool command it would be significantly faster.

Thanks!

[StarGeek]

You're a bit light on details, but I don't see anything there that's exiftool can't do.

As an example, assume we have a filename that has a descriptive text at the beginning and then the date, but reversed (Day Month Year).  So, for a filename of Uncle John's third wedding-22-07-2020.jpg, you could use this (assuming Windows)
"-Keywords+<${Filename;m/(.*)-(\d\d)-(\d\d)-(\d{4})/;$_=\"$4-$3-$2 at $1\"}"

Example:

Code Select Expand

C:\>exiftool -g1 -a -s -keywords "Y:\!temp\cccc\Uncle John's third wedding-22-07-2020.jpg"
---- IPTC ----
Keywords                        : Original Keyword

C:\>exiftool -P -overwrite_original "-Keywords+<${Filename;m/(.*)-(\d\d)-(\d\d)-(\d{4})/;$_=\"$4-$3-$2 at $1\"}" "Y:\!temp\cccc\Uncle John's third wedding-22-07-2020.jpg"
    1 image files updated

C:\>exiftool -g1 -a -s -keywords "Y:\!temp\cccc\Uncle John's third wedding-22-07-2020.jpg"
---- IPTC ----
Keywords                        : Original Keyword, 2020-07-22 at Uncle John's third wedding

Add the -r (-recurse) option option and let it run.  Though I'm not sure what the results would be if there wasn't a match in this case.

[philbond87]

Thank you for that, @StarGeek.
How would that differ for MacOS?
I've tried to apply it to my specific case but I'm not understanding the use of multiple variables that correspond to the four different capture groups.

I'm trying to capture six digits, two or more upper case letters between underscores then another six digit string preceded by an underscore. What I've come up with is:

Code Select Expand

'-comment<${filename;m/(d{6})(\_[A-Z]{2,}_)(\d{6})/;$_=$1$2$3}' FILE

Each of those alone does find the strings I'm searching for in the filenames however my syntax is clearly not right when I combine them.

Thanks

[Phil Harvey]

Can you give an example file name and what you want in the comment?

- Phil

[philbond87]

Hi Phil,

A filename might be:

Code Select Expand

123456_some-arbitrary-words_ABC_987654.jpg

What I would like to end up with in the -comment, from that filename, is:

Code Select Expand

123456ABC987654

(Although the way I built the regex expressions I realize they include underscores I was thinking I could

Code Select Expand

tr/_// them out at the end.)

[Phil Harvey]

The problem is that your regex doesn't allow for the arbitrary words.  Try this:

Code Select Expand

'-comment<${filename;m/(d{6}).*_([A-Z]{2,})_.*(\d{6})/;$_=$1$2$3}'

I have also moved the underlines out of the capture for $2.

- Phil

[Phil Harvey]

Ooops.  There were a couple of other problems in your expression.  This should work:

Code Select Expand

'-comment<${filename;m/(\d{6}).*_([A-Z]{2,})_.*(\d{6})/;$_="$1$2$3"}'

[philbond87]

Thanks Phil.

I see that you've added double quotes around the $ arguments. Was there something else I was missing?

However with the file:

Code Select Expand

123456_some-arbitrary-words_JN_987654.HEIC

using the code:

Code Select Expand

'-comment<${filename;m/(\d{6}).*_([A-Z]{2,})_.*(\d{6})/;$_="$1$2$3"}'

I'm getting:
0 image files updated
1 image files unchanged

[StarGeek]

The Comment tag is a JPG only tag.  It doesn't exist in an HEIC.

I suspect this might be a case where you need to figure out the exact tag you want to write, as many programs give different names for the actual tag.  See FAQ #3.

[philbond87]

Understand, thanks.
I will chose another tag.

I just found another issue related to what I'm trying to do.
In the second sequence of 6 characters I want to include the possibility of there being one (or more) of few specific letters.

What I've come up with is (for the jpg example):

Code Select Expand

'-comment<${filename;m/(\d{6}).*_([A-Z]{2,})_.*([abcDEF]*\d{6})/;$_="$1$2$3"}'

however for the filename:

Code Select Expand

123456_some-arbitrary-words_XY_a987654.jpg

it's not pulling the 'a' out of the second sequence of six numbers, preceded by the 'a', and just returning

Code Select Expand

123456XY987654

[StarGeek]

it's not pulling the 'a' out of the second sequence of six numbers, preceded by the 'a', and just returning

That's because .* is "Greedy" and you want it to be "Lazy".  Try adding a question mark
_.*?([abcDEF]*\d{6})

Is there actually the possibility of arbitrary characters between capital letter group and the final group or is it always just a single underscore?  Maybe this instead
([A-Z]{2,})_([abcDEF]*\d{6})

[philbond87]

That worked a charm. Thank you both very much.
(and no, no arbitrary characters between the uppercase letter group and final sequence group.)

[Phil Harvey]

I see that you've added double quotes around the $ arguments. Was there something else I was missing?

There was also a missing "\" for the first "\d".

- Phil

[philbond87]

There was also a missing "\" for the first "\d"

Ah, right.

So it's working as I would like from the terminal.
When I call it in a shell from within my application the regex isn't working as it did in the terminal – I suspect it's because of the placement and type of quotes.

In my app the command string I send to the shell instance is this:

Code Select Expand

theCommand = "usr/local/bin/exiftool -m -progress: -overwrite_original_in_place -r '-transmissionreference<${filename;m/(\d{6}).*_([A-Z]{2,})_.*?([iwOSEV]*\d{6})/;$_'$1$2$3'}' " + theFolder.ShellPath

Do I somehow need to escape the internal quotes in some way?

[StarGeek]

$_'$1$2$3'}'

The quoting here is incorrect and the equal sign is missing.  There needs to be double quotes around $1$2$3 so perl will interpolate them as variables and replace them with the contents of those variables.  See Strings in Perl: quoted, interpolated and escaped.

I would guess that the double quotes would need to be escaped with backslashes, though you would have to check the docs for the language you're using.

Alternatively, you could use the concatenation operator without quotes.
$_=$1.$2.$3