Issue with collections template in KoboTouch config

[isarl]

I have a saved template function, transform_awards(). However, the following template code gives an error when trying to set it for the Collections template in the Collections, covers & uploads tab:

program: transform_awards()

This used to work with the KoboTouchExtended driver, but now when I try to click OK to persist my changes in the KoboTouch configuration, I get:

calibre, version 8.7.0
ERROR: Invalid template for Collections template:: <p>The template "program: transform_awards()" is invalid:<br>Interpreter: Internal error evaluating an expression: 'list index out of range' - line number 1

However, as you can see from the template tester, it should work fine:



The code for my template function:

def evaluate(self, formatter, kwargs, metadata, _locals):
    awards = metadata.get('#awards')
    if not awards:
        return ''

    transformed_awards = []
    for award in awards:
        # Split the award record into its components
        components = award.split('.')

        # Extract the name of the award and the category
        award_name = components[0]
        award_name = award_name.removesuffix(' Award').removesuffix(' Awards')
        category_name = components[1]
        category_name = category_name.removeprefix('Best ')

        # Extract the year from the record and determine the decade it falls in
        year = int(components[2])
        decade = str(year // 10 * 10) + "s"

        # Extract the rank from the record and determine whether or not it is a winner
        rank = components[3]
        if rank == "winner":
            rank_message = "winners"
        else:
            rank_message = "nominees, etc."

       # Combine the components into the desired format
        transformed_awards.append(f"Awards: {award_name}, {category_name} ({rank_message})")

    return ':@:'.join(transformed_awards)

This seems like a bug, which is why I am reporting it, but if I am doing something wrong, then I appreciate any help getting me back in operation with my collections template. Thank you!

[chaley]

The template check works on fake metadata, not real books. This means that the contents of #awards is a single word, "#awards". Your template fails in this case because it expects any value to be a triplet. Change the template to handle "malformed" awards, ones that don't have the "correct" number of periods, by checking the length of components after the split().

[isarl]

Ah, I need to validate the result of the split has the expected number of components. So obvious now that you point it out – thank you very much!

[isarl]

Just in case anybody else finds it helpful, my updated, now functional, template:

def evaluate(self, formatter, kwargs, metadata, _locals):
    awards = metadata.get('#awards')
    if not awards:
        return ''

    transformed_awards = []
    for award in awards:
        # Split the award record into its components
        components = award.split('.')
        if len(components) < 4:
            return 'error with award: ' + award

        # Extract the name of the award and the category
        award_name = components[0]
        award_name = award_name.removesuffix(' Award').removesuffix(' Awards')
        category_name = components[1]
        category_name = category_name.removeprefix('Best ')

        # Extract the year from the record and determine the decade it falls in
        year = int(components[2])
        decade = str(year // 10 * 10) + "s"

        # Extract the rank from the record and determine whether or not it is a winner
        rank = components[3]
        if rank == "winner":
            rank_message = "winners"
        else:
            rank_message = "nominees, etc."

       # Combine the components into the desired format
        transformed_awards.append(f"Awards: {award_name}, {category_name} ({rank_message})")

    return ':@:'.join(transformed_awards)

[chaley]

Looking at the rest of the template, don't you want to either limit the split to the first 3 commas (maxsplit=3) or (preferred by me) check that the len(components) == 4? And to be complete, you should check that int(components[2]) doesn't throw a ValueError exception.

[isarl]

Good points, thank you! Yes, it's probably better to check that len(components) is exactly 4, as more than 4 would also be a problem. (I initially posted the fixed code checking against 3, then noticed my error and edited my above post.)

In fact I could probably be doing a lot more validation here, like adding a whitelist for known awards or to make sure that the rank field looks right (i.e., is a known word or a positive integer). Since this template is for transforming metadata on the fly while loading books onto my device, I will probably repurpose that added validation elsewhere as well, e.g. to show me a warning icon in my library view for books which fail.

Thanks again for the debugging help! <3