Templates: various questions not worth their own thread

[chaley]

Originally Posted by ownedbycats

Could it have been possible there were, via some glitch, two versions of the stored template temporarily in memory? I checked and saw there was only one icons_fanfic(), but restarting Calibre fixed the error.

Yes, it is possible. The compiled templates functions are kept in a few places such as in the db layer, in the template processor, and in action chains. It could be that one of these isn't being notified when the functions change.

[ownedbycats]

Quick question: I'm having issues with the placement of a fi:

program:
vl = current_virtual_library_name();

	if 'OVERDRIVE' inlist approximate_formats() && vl!='Loans' then 'overdrive.png'
	elif 'PAPERBOOK' inlist approximate_formats() && vl!='Physical Books' then 'book.png'

	elif '^(|Read|Unread|Currently Reading|On Device|Cleanup)$' in vl then
		if $#manualtype then 'helpbook.png'	
		elif $#fanficcat then icons_fanfic() 
		else 'view.png'	
		fi

	fi

In this one, 'view.png' only appears when the current VL is All Books|Read|Unread|Currently Reading|On Device|Cleanup.

program:
vl = current_virtual_library_name();

	if 'OVERDRIVE' inlist approximate_formats() && vl!='Loans' then 'overdrive.png'
	elif 'PAPERBOOK' inlist approximate_formats() && vl!='Physical Books' then 'book.png'

	elif '^(|Read|Unread|Currently Reading|On Device|Cleanup)$' in vl then
		if $#manualtype then 'helpbook.png'	
		elif $#fanficcat then icons_fanfic() 
		fi

	else 'view.png'	

	fi

With this, view.png doesn't appear when the current VL is All Books|Read|Unread|Currently Reading|On Device|Cleanup -- I think the fi after icons_fanfic() is making the template stop processing?

If I want view.png to appear on any book that doesn't match the other checks on any VL, where should I put it?

[chaley]

Originally Posted by ownedbycats

If I want view.png to appear on any book that doesn't match the other checks on any VL, where should I put it?

This appears to be the same problem we discussed in posts 336/337. The same solution should work

[ownedbycats]

Thanks.

When using Action Chains to populate an enumerated column, should it be treated the same as other columns? I'm not sure what would happen if you tried to add a value that didn't exist in the permissible values list.

[chaley]

Originally Posted by ownedbycats

When using Action Chains to populate an enumerated column, should it be treated the same as other columns?

What does "treated the same as other columns" mean?

I'm not sure what would happen if you tried to add a value that didn't exist in the permissible values list.

An attempt to set a non-permitted value is ignored. The column doesn't change. No error is signaled.

[ownedbycats]

Basically, just inputting values the same as you would with a regular text column.

[chaley]

Originally Posted by ownedbycats

Basically, just inputting values the same as you would with a regular text column.

Then yes. These columns are internally equivalent to "Text, column shown in the Tag browser" where a book can hold a single value. As I said above, assignment of a value that is not in the list of permitted values is ignored. It is up to you to ensure the value you assign is permitted.

[ownedbycats]

Quick question:

program: if $something then 'Yes' fi

program: if $something then 'Yes' else 'No' fi

Is there any difference in performance of returning 'No' for yes/no sorting?

[chaley]

Originally Posted by ownedbycats

Quick question:

program: if $something then 'Yes' fi

program: if $something then 'Yes' else 'No' fi

Is there any difference in performance of returning 'No' for yes/no sorting?

Returning 'No' is perhaps a microsecond per book slower when the template is processed, which is once per book when sorting. 'No' vs the empty string should make no difference at all.

[ownedbycats]

Question:

I have two timedate columns with a difference of less than an hour.* If I put the raw input into a days_between, depending on what order I either get 0.0 or -0.0.

Is the negative intended or a bug? There's something called a signed zero but I don't quite understand what it means.

I tried multiplying to see what would happen but still got -0.0.

* 2022-03-19 02:03:12+00:00 and 2022-03-19 03:00:00+00:00 exactly.

[chaley]

Originally Posted by ownedbycats

Question:

I have two timedate columns with a difference of less than an hour.* If I put the raw input into a days_between, depending on what order I either get 0.0 or -0.0.

Is the negative intended or a bug? There's something called a signed zero but I don't quite understand what it means.

I tried multiplying to see what would happen but still got -0.0.

* 2022-03-19 02:03:12+00:00 and 2022-03-19 03:00:00+00:00 exactly.

It means that there are digits you aren't seeing. It is formatted with 1 decimal place, the 10ths column. The full result will have a value in the 100ths or maybe even the 1000ths column. Thus the answer is less than zero.

If you don't like it, use round(), floor(), or ceiling(). Or convert the dates to timestamps, do the arithmetic, divide the result by 86,400 to get the fractional days, then format it with whatever precision you want.

[chaley]

Originally Posted by ownedbycats

Question:

I have two timedate columns with a difference of less than an hour.* If I put the raw input into a days_between, depending on what order I either get 0.0 or -0.0.

Is the negative intended or a bug? There's something called a signed zero but I don't quite understand what it means.

I tried multiplying to see what would happen but still got -0.0.

* 2022-03-19 02:03:12+00:00 and 2022-03-19 03:00:00+00:00 exactly.

For fun try the template

program: format_number((-0.001), '.1f')

You will see the negative 0. It is what python does.

You can see it without the formatter as the intermediary using the python command line interpreter.

>>> '%.1f'%(0.01)
'0.0'
>>> '%.1f'%(-0.01)
'-0.0'
>>> '%.1f'%(-0.01/10)
'-0.0'
>>> '%.3f'%(-0.01/10)
'-0.001'

[ownedbycats]

I found this: https://en.wikipedia.org/wiki/IEEE_754

Moreover, there are two zero values, called signed zeros: the sign bit specifies whether a zero is +0 (positive zero) or −0 (negative zero).

That makes more sense than the other thing I tried to read on signed zeros

[chaley]

Originally Posted by ownedbycats

I found this: https://en.wikipedia.org/wiki/IEEE_754



That makes more sense than the other thing I tried to read on signed zeros

This might be helpful: Negative Binary Numbers. Two of the three representations, sign-and-magnitude and ones-complement, can represent negative zero. Twos-complement, the most common, cannot represent negative zero.

The above discussion leaves out the complexity of floating point numbers that usually have three separate components: a sign, a "value", and an "exponent". Because (usually) the sign is separate from the number is it similar to sign-and-magnitude representation. See https://floating-point-gui.de/formats/fp/ for more explanation.

[ownedbycats]

Question: What would be the best way to take an integer (say, from days_between) and then turn that into "x years, x months, x days"?

I started trying to make a template using fractional_part and I think multiply but it didn't come out right.