Templates: various questions not worth their own thread - Page 31

ownedbycats · 01-04-2023, 07:47 PM

Thanks.

When using first_non_empty, is there any performance difference between checking a default and custom column? Trying to figure out the optimal order for this (e.g. the more complex ones with multiple checks at the end).

Code:

program:
## Missing metadata
first_non_empty(
	if !$formats then 'plus.png' fi,
	if !$tags then 'tags.png' fi,
	if !$languages then 'languages.png' fi,
	if !$#booktype then 'catalog.png' fi,
	if !$#pagecount then 'cover_flow.png' fi,
	if !$#chaptercount && !'paperbook' in approximate_formats() then 'cover_flow.png' fi,
	if $#booktype == 'Documentation & Manuals' && !$#manualtype then 'catalog.png' fi,
	if $#kobostatus && $#kobostatus != 'Kobo Store' && !$#kobopath then 'snippets.png' fi,
)

chaley · 01-05-2023, 10:40 AM

Quote:

Originally Posted by ownedbycats

Thanks.

When using first_non_empty, is there any performance difference between checking a default and custom column? Trying to figure out the optimal order for this (e.g. the more complex ones with multiple checks at the end).

Fetching a custom column is nanoseconds slower than fetching a standard column. Certainly not enough to worry about.

The first_non_empty() conditions should be ordered by likelihood. Put the ones first that are most likely to match, regardless of complexity.

ownedbycats · 01-05-2023, 09:30 PM

Thanks. Another question: what is the difference, if any, between using a first_non_empty and an if-then-elif in a situation like this?

first_non_empty is a bit easier for me (less misplaced fi's, for one!) so I'm inclined to keep it unless there's some large advantage to the other.

Code:

program:

	if $#booktype == 'Fanfiction' then
		first_non_empty(
			if 
				$#fanficstatus == 'In-Progress' 
				&& days_between(format_date(today(), 'yyyy-MM-dd'), $#fanficupdated) >= 365 
			then 
				'clock.png' 
			fi,
			
			if 
				!'^cover:' inlist $#admintags 
			then 
				'test' 
			fi
		)
	fi

Code:

program:

	if $#booktype == 'Fanfiction' then
		if 
			$#fanficstatus == 'In-Progress' 
			&& days_between(format_date(today(), 'yyyy-MM-dd'), $#fanficupdated) >= 365 
		then
 			'clock.png'

		elif 
			!'^cover:' inlist $#admintags 
		then
			'test'
		fi
	fi

chaley · 01-06-2023, 05:26 AM

Quote:

Originally Posted by ownedbycats

Thanks. Another question: what is the difference, if any, between using a first_non_empty and an if-then-elif in a situation like this?

first_non_empty is a bit easier for me (less misplaced fi's, for one!) so I'm inclined to keep it unless there's some large advantage to the other.

Code:

program:

	if $#booktype == 'Fanfiction' then
		first_non_empty(
			if 
				$#fanficstatus == 'In-Progress' 
				&& days_between(format_date(today(), 'yyyy-MM-dd'), $#fanficupdated) >= 365 
			then 
				'clock.png' 
			fi,
			
			if 
				!'^cover:' inlist $#admintags 
			then 
				'test' 
			fi
		)
	fi

Code:

program:

	if $#booktype == 'Fanfiction' then
		if 
			$#fanficstatus == 'In-Progress' 
			&& days_between(format_date(today(), 'yyyy-MM-dd'), $#fanficupdated) >= 365 
		then
 			'clock.png'

		elif 
			!'^cover:' inlist $#admintags 
		then
			'test'
		fi
	fi

The list of ifs is slightly faster (less than microseconds), not enough to worry about.

I have wondered whether I should make a variant that combines first_non_empty() and switch(), perhaps named first_true(). It would take a list of argument pairs where each pair is a condition expression and a value expression. For each pair in order it would evaluate the condition. If True then it would evaluate the value expression and return it, otherwise try the next pair. There would be a 'none are true' value at the end, like switch().

This would be faster than either first_non_empty() or a series of ifs, and would probably be easier to read. But it would be yet another maintenance and documentation chore. It is also quite easy to do in a python template. I am still thinking about it.

chaley · 01-06-2023, 11:05 AM

Quote:

Originally Posted by chaley

I have wondered whether I should make a variant that combines first_non_empty() and switch(), perhaps named first_true(). It would take a list of argument pairs where each pair is a condition expression and a value expression. For each pair in order it would evaluate the condition. If True then it would evaluate the value expression and return it, otherwise try the next pair. There would be a 'none are true' value at the end, like switch().

This would be faster than either first_non_empty() or a series of ifs, and would probably be easier to read. But it would be yet another maintenance and documentation chore. It is also quite easy to do in a python template. I am still thinking about it.

I added the function switch_if(). It is in source now.

ownedbycats · 01-06-2023, 04:57 PM

I'll give it a test in a few days

Two questions here. #purchasecost is a floating-point with format of $x.xx.

Code:

program:

		if $$#purchasecost ==# '0.00' then '0.00' 
		elif $$#purchasecost <=# '0.99' then '$0.01 - $0.99' 
		elif $$#purchasecost <=# '4.99' then '$1.00 - $4.99' 
		elif $$#purchasecost <=# '10.00' then '$5.00 - $9.99' 		
		elif $$#purchasecost <=# '15.00' then '$10.00 - $14.99' 	
		elif $$#purchasecost <=# '15.00' then '$10.00 - $14.99' 
		elif $$#purchasecost <=# '20.00' then '$15.00 - $19.99' 
		elif $$#purchasecost <=# '30.00' then '$20.00 - $29.99' 
		elif $$#purchasecost >=# '30.00' then '$30.00 and up' 
		fi

a) This'll always try to show the first entry for undefineds. if $$#purchasecost ==# '' then '' best option?

b) More out of curiosity: The first time I made this template, I forgot to add the '#' to all the cmps. Line 11, >= '30.00', showed up on undefineds. This didn't happen for == or <=. How did it get 'undefined is more than 30.00'?

chaley · 01-06-2023, 05:12 PM

Quote:

Originally Posted by ownedbycats

I'll give it a test in a few days

Two questions here. #purchasecost is a floating-point with format of $x.xx.

Code:

program:

		if $$#purchasecost ==# '0.00' then '0.00' 
		elif $$#purchasecost <=# '0.99' then '$0.01 - $0.99' 
		elif $$#purchasecost <=# '4.99' then '$1.00 - $4.99' 
		elif $$#purchasecost <=# '10.00' then '$5.00 - $9.99' 		
		elif $$#purchasecost <=# '15.00' then '$10.00 - $14.99' 	
		elif $$#purchasecost <=# '15.00' then '$10.00 - $14.99' 
		elif $$#purchasecost <=# '20.00' then '$15.00 - $19.99' 
		elif $$#purchasecost <=# '30.00' then '$20.00 - $29.99' 
		elif $$#purchasecost >=# '30.00' then '$30.00 and up' 
		fi

a) This'll always try to show the first entry for undefineds. if $$#purchasecost ==# '' then '' best option?

b) More out of curiosity: The first time I made this template, I forgot to add the '#' to all the cmps. Line 11, >= '30.00', showed up on undefineds. This didn't happen for == or <=. How did it get 'undefined is more than 30.00'?

You are using raw_field ($$), in which case the format is ignored and an undefined number returns the string "none" instead of the empty string. The string "none" is larger than the string "30.00". If you want to check for undefined values then you should explicitly check for "none".

Alternatively, to control results you could use raw_field() explicitly to set what value you want for undefined. For example, you could say

Code:

    v = raw_field('#purchasecost', -100000.00)

which sets v to -100000.00 if #purchasecost isn't defined. Then use the variable v in the if chain so that undefined numbers are very small and less than zero.

Finally, the last branch of the if shouldn't have a condition but should instead always return '$30.00 and up'. The test serves no purpose.

ownedbycats · 01-07-2023, 01:42 AM

That worked.

Tweaking my #kobopath template:

Code:

program:

	if $#kobostatus then

# Returns 'Fanfiction/Fandom' tag for fanfics
		if $$#booktype == 'Fanfiction' then
			list_re($tags, ',', '^Fanfiction.(.*)$', 'Fanfiction/\1')	
# Returns 'PDF' for PDFs
		elif $formats == 'PDF' then
			'PDF'
# Returns 'Loans' for Loans
		elif 'OVERDRIVE' in $formats then
			'Loans'
# Returns booktype for others
		else
			$#booktype
		fi

	fi

As seen in this screenshot, something that has multiple Fanfiction.Whatever tags lists all three. Oops!

Click image for larger version

Name: 2023-01-07 02_13_29-Edit template.png
Views: 40
Size: 37.8 KB
ID: 198858

Due to my tagging practices:
a) Fanfiction.Crossover is the always the first in the list;
a) I only have multiple tags when one is 'Fanfiction.Crossover'
b) I always prefer 'Crossover' for the first pick.

So based on this, I thought of two separate solutions:

Option #1: splitting the list:

Code:

	if $$#booktype == 'Fanfiction' then
		list_split(list_re($tags, ',', '^Fanfiction.(.*)$', 'Fanfiction/\1'), ',', 'ff');	
		ff_0

Disadvantage: This relies on Fanfiction.Crossover being first on the list. However, I rarely make changes to those subtags.

Option #2 A second check for 'Crossover' specifically. Here, I'm checking the taglike #fanficcat but I could also check Fanfiction.Crossover directly.

Code:

	if $$#booktype == 'Fanfiction' then
			if 'Crossover' in $#fanficcat then 'Fanfiction/Crossover'
			else list_re($tags, ',', '^Fanfiction.(.*)$', 'Fanfiction/\1')
		fi

Disavantage: This relies on 'Crossover' always being in #fanficcat (or Fanfiction.Crossover in tags). This is pretty much always the case.

Another disavantage: This also relies on 'Crossover' not being in #fanficcat (or Fanfiction.Crossover in tags) when unneeded. Thanks to the way FFF works with multi-fandom stories, this has happened.

Based on that second disadvantage, I'm inclined to go with option #1. Does this make sense?

chaley · 01-07-2023, 05:39 AM

Quote:

Originally Posted by ownedbycats

That worked.

Tweaking my #kobopath template:

[...]

Based on that second disadvantage, I'm inclined to go with option #1. Does this make sense?

I would avoid complexity and make the checks explicit. The new switch_if() makes it easy and compact to do the checks in order. Something like this:

Code:

program:

# Returns 'Fanfiction/Fandom' tag for fanfics
	booktype = 'Fanfiction';
	tags = 'Fanfiction.Bar, Fanfiction.Crossover, Fanfiction.Foo';
	if booktype == 'Fanfiction' then
		switch_if(
			'^Fanfiction\.Crossover$' inlist tags, 'Fanfiction/Crossover',
# More tests in the order you want
			'^Fanfiction\.Foo$' inlist tags, 'Fanfiction/Foo',
#			...
# If none of the tests find a value then end up picking the first one. 
# If the list is empty then this will return the empty string.
# You only need the list_re() if the list can contain values other than
# Fanfiction.XXX.
			list_item(list_re(tags, ',', '^Fanfiction.(.*)$', 'Fanfiction/\1'), 0, ',')
		)
	fi

ownedbycats · 01-07-2023, 07:48 PM

Thanks.

I converted one of my templates to get a better idea of how the switch_if works. Is this a good place to use one?

Code:

program:
	switch_if(
		'Cleanup' in virtual_libraries(), 'polish.png',
		'Fanfiction' in $#booktype, list_split(icons_fanfic(), ':', 'ff'); 
			ff_0,
		'Documentations & Manuals' in $#booktype, 'helpbook.png',
		'Loans' in virtual_libraries(), 'overdrive.png',
		'Physical Books' in virtual_libraries(), 'paperbook.png',
		approximate_formats()=='PDF', 'pdfbook.png',
		'omnibus' inlist $#admintags, 'bookshelf.png',
		'book_open.png',
	)

original first_non_empty template:

Spoiler:

(at times like this, sometimes I wish there was some sort of split template editor so I could compare the results of two templates. That's probably not too useful generally though.)

ownedbycats · 01-08-2023, 03:27 AM

Another example:

Code:

program:	
	if
		is_marked()
	then
		m = is_marked();
		switch_if(
			'reading_list_send_to_device' in m, 'sync.png',
			'reading_list_to_be_read' in m, 'list.png',
			'fff' in m, 'download-metadata.png',
			'marked.png'
		)
	fi

Original. (Trying to figure out why I used a 'contains' instead of 'in'.)

Spoiler:

chaley · 01-08-2023, 06:41 AM

Quote:

Originally Posted by ownedbycats

Thanks.

I converted one of my templates to get a better idea of how the switch_if works. Is this a good place to use one?

Code:

program:
	switch_if(
		'Cleanup' in virtual_libraries(), 'polish.png',
		'Fanfiction' in $#booktype, list_split(icons_fanfic(), ':', 'ff'); 
			ff_0,
		'Documentations & Manuals' in $#booktype, 'helpbook.png',
		'Loans' in virtual_libraries(), 'overdrive.png',
		'Physical Books' in virtual_libraries(), 'paperbook.png',
		approximate_formats()=='PDF', 'pdfbook.png',
		'omnibus' inlist $#admintags, 'bookshelf.png',
		'book_open.png',
	)

Yes, this is exactly the situation I envisioned for using switch_if().

Comment: you often use 'in' instead of 'inlist'. Using 'in' can lead to errors because it can find substrings. For example, if you have virtual libraries named "Loans" and "Temp Loans", the "in" will find them both, while inlist will find only the one as long as you use regular expression anchors on the search value.

Quote:

Originally Posted by ownedbycats

Another example:

Code:

program:	
	if
		is_marked()
	then
		m = is_marked();
		switch_if(
			'reading_list_send_to_device' in m, 'sync.png',
			'reading_list_to_be_read' in m, 'list.png',
			'fff' in m, 'download-metadata.png',
			'marked.png'
		)
	fi

IIRC a book can have at most one mark. In this case you should use == instead of 'in', because == is much faster. You could also rewrite the if/then to get rid of it.

Example:

Code:

program:	
	m = is_marked();
	switch_if(
		'reading_list_send_to_device' == m, 'sync.png',
		'reading_list_to_be_read' == m, 'list.png',
		'fff' == m, 'download-metadata.png',
		m, 'marked.png',
		''
	)

ownedbycats · 01-08-2023, 06:45 AM

Quote:

Originally Posted by chaley

IIRC a book can have at most one mark. In this case you should use == instead of 'in', because == is much faster. You could also rewrite the if/then to get rid of it.

Example:

Code:

program:	
	m = is_marked();
	switch_if(
		'reading_list_send_to_device' == m, 'sync.png',
		'reading_list_to_be_read' == m, 'list.png',
		'fff' == m, 'download-metadata.png',
		m, 'marked.png',
		''
	)

Here's the slight problem with that (applies to all three listed):

Click image for larger version

Name: 2023-01-08 07_44_51-Template tester.png
Views: 63
Size: 26.3 KB
ID: 198891

chaley · 01-08-2023, 06:46 AM

Quote:

Originally Posted by ownedbycats

Here's the slight problem with that (applies to all three listed):

Attachment 198891

Ooops. I didn't know that the mark had more text. Using 'in' is the right way to go.

ownedbycats · 01-08-2023, 06:55 AM

I do keep forgetting that virtual_libraries is a list though. I can see that especially messing up since I have 'read' and 'unread' vls.

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

01-05-2023, 09:30 PM	#453
ownedbycats Custom User Title Posts: 8,657 Karma: 61234567 Join Date: Oct 2018 Location: Canada Device: Kobo Libra H2O, formerly Aura HD	Thanks. Another question: what is the difference, if any, between using a first_non_empty and an if-then-elif in a situation like this? first_non_empty is a bit easier for me (less misplaced fi's, for one!) so I'm inclined to keep it unless there's some large advantage to the other. Code: program: if $#booktype == 'Fanfiction' then first_non_empty( if $#fanficstatus == 'In-Progress' && days_between(format_date(today(), 'yyyy-MM-dd'), $#fanficupdated) >= 365 then 'clock.png' fi, if !'^cover:' inlist $#admintags then 'test' fi ) fi Code: program: if $#booktype == 'Fanfiction' then if $#fanficstatus == 'In-Progress' && days_between(format_date(today(), 'yyyy-MM-dd'), $#fanficupdated) >= 365 then 'clock.png' elif !'^cover:' inlist $#admintags then 'test' fi fi Last edited by ownedbycats; 01-05-2023 at 11:49 PM.

01-06-2023, 04:57 PM	#456
ownedbycats Custom User Title Posts: 8,657 Karma: 61234567 Join Date: Oct 2018 Location: Canada Device: Kobo Libra H2O, formerly Aura HD	I'll give it a test in a few days Two questions here. #purchasecost is a floating-point with format of $x.xx. Code: program: if $$#purchasecost ==# '0.00' then '0.00' elif $$#purchasecost <=# '0.99' then '$0.01 - $0.99' elif $$#purchasecost <=# '4.99' then '$1.00 - $4.99' elif $$#purchasecost <=# '10.00' then '$5.00 - $9.99' elif $$#purchasecost <=# '15.00' then '$10.00 - $14.99' elif $$#purchasecost <=# '15.00' then '$10.00 - $14.99' elif $$#purchasecost <=# '20.00' then '$15.00 - $19.99' elif $$#purchasecost <=# '30.00' then '$20.00 - $29.99' elif $$#purchasecost >=# '30.00' then '$30.00 and up' fi a) This'll always try to show the first entry for undefineds. if $$#purchasecost ==# '' then '' best option? b) More out of curiosity: The first time I made this template, I forgot to add the '#' to all the cmps. Line 11, >= '30.00', showed up on undefineds. This didn't happen for == or <=. How did it get 'undefined is more than 30.00'? Last edited by ownedbycats; 01-06-2023 at 05:01 PM.

01-07-2023, 01:42 AM	#458
ownedbycats Custom User Title Posts: 8,657 Karma: 61234567 Join Date: Oct 2018 Location: Canada Device: Kobo Libra H2O, formerly Aura HD	That worked. Tweaking my #kobopath template: Code: program: if $#kobostatus then # Returns 'Fanfiction/Fandom' tag for fanfics if $$#booktype == 'Fanfiction' then list_re($tags, ',', '^Fanfiction.(.)$', 'Fanfiction/\1') # Returns 'PDF' for PDFs elif $formats == 'PDF' then 'PDF' # Returns 'Loans' for Loans elif 'OVERDRIVE' in $formats then 'Loans' # Returns booktype for others else $#booktype fi fi As seen in this screenshot, something that has multiple Fanfiction.Whatever tags lists all three. Oops! Due to my tagging practices: a) Fanfiction.Crossover is the always the first in the list; a) I only have multiple tags when one is 'Fanfiction.Crossover' b) I always prefer 'Crossover' for the first pick. So based on this, I thought of two separate solutions: Option #1: splitting the list: Code: if $$#booktype == 'Fanfiction' then list_split(list_re($tags, ',', '^Fanfiction.(.)$', 'Fanfiction/\1'), ',', 'ff'); ff_0 Disadvantage: This relies on Fanfiction.Crossover being first on the list. However, I rarely make changes to those subtags. Option #2 A second check for 'Crossover' specifically. Here, I'm checking the taglike #fanficcat but I could also check Fanfiction.Crossover directly. Code: if $$#booktype == 'Fanfiction' then if 'Crossover' in $#fanficcat then 'Fanfiction/Crossover' else list_re($tags, ',', '^Fanfiction.(.)$', 'Fanfiction/\1') fi Disavantage: This relies on 'Crossover' always being in #fanficcat (or Fanfiction.Crossover in tags). This is pretty much always the case. Another disavantage: This also relies on 'Crossover' not* being in #fanficcat (or Fanfiction.Crossover in tags) when unneeded. Thanks to the way FFF works with multi-fandom stories, this has happened. Based on that second disadvantage, I'm inclined to go with option #1. Does this make sense? Last edited by ownedbycats; 01-07-2023 at 02:17 AM.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Library Management: various questions not worth their own thread	ownedbycats	Library Management	138	04-23-2024 11:49 AM
[Metadata Source Plugin] Questions regarding parse select, docs and ref templates	Boilerplate4U	Development	13	07-07-2020 02:35 AM
Questions on Kobo [Interfered with another thread topic]	spdavies	Kobo Reader	8	10-12-2014 11:37 AM
[OLD Thread] Some questions before buying the fire.	darthreader13	Kindle Fire	7	05-10-2013 09:19 PM
Thread management questions	meme	Feedback	6	01-31-2011 05:07 PM

01-08-2023, 06:55 AM	#465
ownedbycats Custom User Title Posts: 8,657 Karma: 61234567 Join Date: Oct 2018 Location: Canada Device: Kobo Libra H2O, formerly Aura HD	I do keep forgetting that virtual_libraries is a list though. I can see that especially messing up since I have 'read' and 'unread' vls. Also, is there a difference between using approximate_formats and $formats for the =='PDF'?

Advert

Advert