Templates: various questions not worth their own thread

[chaley]

Originally Posted by ownedbycats

Also, is there a difference between using approximate_formats and $formats for the =='PDF'?

No. The both refer to the formats in the database without looking at what is on the disk. The formats_*() functions look at the disk, which makes them much slower.

[ownedbycats]

How is switch() meant to be used? Here, I got overdrive.png when I was expecting drm-locked.png, so I think I either
a) misunderstood the syntax
b) misunderstood what instances it's meant to be used in:

[ownedbycats]

Is there a function that sorts a list into a user-defined order?

Context: I have a text column that may return results in a list-string like "foo, bar, foo, foo, foobar."

Using Action Chains, I'd like to:
a) Sort the entries in a user-defined order (e.g. "foobar, bar, foo, foo, foo)
b) Select the first item of this sorted list

And then replace the string with that.

(Additional context: When FanFicFare downloads an anthology of multiple fics, it concenates all the ratings into an unsorted string like "General Audiences, Mature Audiences, Explicit, Teen and Up Audiences, Not Rated." In order, I want to pick the first of Explicit > Mature > Teen > General > Not Rated.)

edit: can I check just one column with a first_non_empty?

[chaley]

Originally Posted by ownedbycats

Is there a function that sorts a list into a user-defined order?

Context: I have a text column that may return results in a list-string like "foo, bar, foo, foo, foobar."

Using Action Chains, I'd like to:
a) Sort the entries in a user-defined order (e.g. "foobar, bar, foo, foo, foo)
b) Select the first item of this sorted list

And then replace the string with that.

(Additional context: When FanFicFare downloads an anthology of multiple fics, it concenates all the ratings into an unsorted string like "General Audiences, Mature Audiences, Explicit, Teen and Up Audiences, Not Rated." In order, I want to pick the first of Explicit > Mature > Teen > General > Not Rated.)

No. The only way to do it that I can think of is to make a list containing all possible items, then loop through that list using for. If the actual list contains the current item then add it to the result list, otherwise skip it. Something like this:

program:
	possibles = 'Explicit, Mature, Teen, General, Not Rated';
	actuals = 'foo, Mature, Teen, Explicit';

	answer = '';
	for item in possibles:
		if '^' & item & '$' inlist actuals then
			answer = item;
			break
		fi
	rof;
	answer

edit: can I check just one column with a first_non_empty?

I don't understand. Are you asking if you can have only one rule? Yes.

[chaley]

Originally Posted by ownedbycats

How is switch() meant to be used? Here, I got overdrive.png when I was expecting drm-locked.png, so I think I either
a) misunderstood the syntax
b) misunderstood what instances it's meant to be used in:

You misunderstand the syntax. The first argument is a value, perhaps a $column selector. The rest of the arguments are (pattern, value) pairs. The pattern is checked against the first argument. If it matches then the value is returned.

The syntax you are using is switch_if().

[ownedbycats]

Here's what I came up with:

program:
## list_remove_duplicates probably isn't necessary
	list = list_remove_duplicates($#fanficrating, ',');

	first_non_empty(
		if 'Explicit' inlist list then 'Explicit' fi,
		if 'Mature Audiences' inlist list then "Mature Audiences" fi,
		if "Teen and Up Audiences" inlist list then "Teen and Up Audiences" fi,
		if "General Audiences" inlist list then "General Audiences" fi,
		if "Not Rated" inlist list then "Not Rated" fi
	)

The first_non_empty question I was thinking of something along the lines of:

	first_non_empty inlist list(
		"Explicit",
		"Mature Audiences"
		"Teen and Up Audiences"
		"General Audiences"
		"Not Rated"
	)

[chaley]

Originally Posted by ownedbycats

Here's what I came up with:

program:
## list_remove_duplicates probably isn't necessary
	list = list_remove_duplicates($#fanficrating, ',');

	first_non_empty(
		if 'Explicit' inlist list then 'Explicit' fi,
		if 'Mature Audiences' inlist list then "Mature Audiences" fi,
		if "Teen and Up Audiences" inlist list then "Teen and Up Audiences" fi,
		if "General Audiences" inlist list then "General Audiences" fi,
		if "Not Rated" inlist list then "Not Rated" fi
	)

That works well. If you haven't already done it, put the checks in probability order.

Note that what you wrote is exactly equivalent to this switch_if. I used the assignment to avoid writing the string twice. It isn't needed as the last test shows.

program:
## list_remove_duplicates probably isn't necessary
	list = list_remove_duplicates($#fanficrating, ',');

	switch_if(
		(a = 'Explicit') inlist list, a,
		(a = 'Mature Audiences') inlist list, a,
		(a = "Teen and Up Audiences") inlist list, a,
		(a = "General Audiences") inlist list, a,
		"Not Rated" inlist list, "Not Rated",
		''
	)

[ownedbycats]

Originally Posted by chaley

That works well. If you haven't already done it, put the checks in probability order.

Specifically, I wanted to pick the "highest" rating from the string, hence the specific order. Explicit is first, even though it's the rating I have the fewest books of.

[ownedbycats]

Would there be any general use for a first_in_list(val, separator, value [, value]*) type function?

It would check a list for each pattern in order, and then return the found list item as the value.

Template: first_in_list($tags, ',', "Fanfiction.Crossover")
List contents: Fanfiction.Crossover, Fanfiction.Half-Life & Portal, Fanfiction.Mass Effect
Template value: Fanfiction.Crossover

Template: first_in_list(list, ',', "Explicit", "Mature Audiences", "Teen and Up Audiences", "General Audiences", "Not Rated")
List contents: General Audiences, Explicit, Not Rated, Not Rated
Template value: Explicit

Regex pattern matching may be useful but as with the in/inlist, it lead to unintended results:

Template: first_in_list(list, ',', "foo", "^foobar")
List contents: foobar, foo, bar, bar, foo
Template value: foobar
whoops! the regex matched 'foo'

EDIT: Perhaps as a custom function.

[chaley]

Originally Posted by ownedbycats

Would there be any general use for a first_in_list(val, separator, value [, value]*) type function?

It would check a list for each pattern in order, and then return the found list item as the value.

Template: first_in_list($tags, ',', "Fanfiction.Crossover")
List contents: Fanfiction.Crossover, Fanfiction.Half-Life & Portal, Fanfiction.Mass Effect
Template value: Fanfiction.Crossover

Template: first_in_list(list, ',', "Explicit", "Mature Audiences", "Teen and Up Audiences", "General Audiences", "Not Rated")
List contents: General Audiences, Explicit, Not Rated, Not Rated
Template value: Explicit

Regex pattern matching may be useful but as with the in/inlist, it lead to unintended results:

Template: first_in_list(list, ',', "foo", "^foobar")
List contents: foobar, foo, bar, bar, foo
Template value: foobar
whoops! the regex matched 'foo'

EDIT: Perhaps as a custom function.

This would really be sub-variant of switch, such as switch_list(). As proposed the function eliminates the second piece of the pair, the result, which I think is over-specialized. I think switch_list() is useful and will look at implementing it.

If I add switch_list() then I will probably have it set the built-in variable '$' to the result of the match. That would let you write (untested):

switch_if(list, ',',
    '^foo$', $,
    strcat('bar', $#status), $,
    '')

[chaley]

Originally Posted by ownedbycats

EDIT: Perhaps as a custom function.

For fun, here is a Python stored template that does what you want. Stored templates like this obviate the need for a lot of seldom-used built-in functions.

The template:

python:
def evaluate(book, context):
	args = context.arguments
	if args is None or len(args) != 3:
		raise ValueError('first_in_list requires 3 arguments')
	sep = args[1]
	lst = tuple(l.strip() for l in args[0].split(',') if l.strip())
	test_lst = tuple(l.strip() for l in args[2].split(',') if l.strip())
	if not test_lst:
		return ''
	for v in test_lst:
		if v in lst:
			return v
	return ''

The list match isn't a regex and is case sensitive. EDIT: it is trivial to make it case insensitive.

Calling it:

program:
	first_in_list($tags, ',', 'a, b, Space Opera, c')

When called with a book with tags "Science Fiction, Space Opera, onDevice" it returns "Space Opera".

[ownedbycats]

That works! I'll use it until (if) switch_list() is added.

[ownedbycats]

I just noticed that empty lines at the end of a template get removed, at least in tester. That's a nice little QoL detail.

[ownedbycats]

Is there any risk of permanent damage if I accidentally open a stable version of Calibre where a new function hadn't been added?

The only thing I can think of is a error'd save path if I send a book to device while the template is (indirectly, using a composite column) using switch_if().

[chaley]

Originally Posted by ownedbycats

Is there any risk of permanent damage if I accidentally open a stable version of Calibre where a new function hadn't been added?

The only thing I can think of is a error'd save path if I send a book to device while the template is (indirectly, using a composite column) using switch_if().

The template will internally throw an exception if a function doesn't exist. In the case of save_to_disk, this exception propagates out and the save fails.

In other contexts, which I think includes most places including device drivers, the value of the template is the error message. This can affect the value of composite columns, so anything that depends on them will "see" the wrong answer. Whether this results in permanent damage I can't say.