Templates: various questions not worth their own thread

[ownedbycats]

program:
	publisher = $publisher;
	ids = $identifiers;
	u = select(ids, 'url');
	s = select(ids, 'standard');

	if publisher == 'Standard Ebooks' && u != s then
		n = re(u, '^^https://standardebooks.org/ebooks/()', '\1');
		ids = list_union(ids, strcat('standard:', n), ',')
	fi;

	if publisher == 'Standard Ebooks' && u && s then
		ids = list_difference(ids, strcat('url:', u), ',')
	fi;
	ids

In an Action Chain, if I run this once, it produces the standardidentifier and does not delete the url: identifier. If I run it on a book that has both standard and url: identifiers it does delete the url:. Do I need to make it two separate actions/templates?

[chaley]

Originally Posted by ownedbycats

[...]
In an Action Chain, if I run this once, it produces the standardidentifier and does not delete the url: identifier. If I run it on a book that has both standard and url: identifiers it does delete the url:. Do I need to make it two separate actions/templates?

I verified that an Action Chains single field edit will replace the ids with the result of the template. Example: given identifiers of

ao3:1, goodreads:16967, uri:https://archiveofourown/works/12345, x:y

and the action chain

Spoiler:

{
  "ActionsChainMenus": {
    "Menus": [
      {
        "active": true,
        "chain_settings": {
          "chain_links": [
            {
              "action_name": "Single Field Edit",
              "action_settings": {
                "col_name": "identifiers",
                "template": "program:\n\tlist_difference($identifiers, 'ao3:1', ',')",
                "value_type": "template"
              },
              "comment": ""
            }
          ],
          "refresh_gui": true
        },
        "image": "",
        "menuText": "Test identifiers",
        "subMenu": ""
      }
    ]
  },
  "schemaVersion": 1.8
}

then the result is

goodreads:16967, uri:https://archiveofourown/works/12345, x:y

The ao3:1 identifier is removed. I can't say why it doesn't work for you. I suggest you put print() statements in it so you can verify exactly the result of the template.

FWIW: the

n = re(u, '^^https://standardebooks.org/ebooks/()', '\1');

is strange. The \1 will always result in the empty string because the grouping () contains no pattern, so why is it there? I also don't know why there are two ^ anchors in the front. I think you mean

n = re(u, '^https://standardebooks.org/ebooks/(.*)$', '\1');

[ownedbycats]

I was having trouble with the regex and basically just put in the first thing that actually halfway worked.

Oddly, template tester isn't removing the url: identifier either:



If I take out the check for 'standard', it works:



Is it looking at the state of the column at the beginning of the template, when it only has the url identifier? That would explain why it works running it a second time.

[chaley]

Originally Posted by ownedbycats

I was having trouble with the regex and basically just put in the first thing that actually halfway worked.

Oddly, template tester isn't removing the url: identifier either:

Attachment 191593

If I take out the check for 'standard', it works:

Attachment 191595

Is it looking at the state of the column at the beginning of the template, when it only has the url identifier? That would explain why it works running it a second time.

It is looking at the value of "s", which is the result of the select() on line 5. It doesn't see any changes you make to "ids" after line 5.

However, I suspect your logic won't do what you want. If s is non-empty then there is a "standard" identifier. If s != u then you add a standard identifier, which is probably different from the original so there are now 2 of them in the list. When you save the identifiers calibre will pick one of them; which one is undefined. Thus if s is non-empty then you must remove the existing "standard" identifier before adding a different "standard" identifer.

Once you have done the above the rest of your logic assumes that 's' contains the current value of the "standard" identifier. Your template must ensure that is the case, perhaps by assigning to 's'.

Depending on what you are doing it might be easier and less buggy to loop through the identifiers, building a new identifiers list.

[ownedbycats]

For now, I think I'll just remove the check for the standard: identifier. I could probably also remove the check for url: too.

Unrelated comment: Using stored templates for complex checks cuts down a lot on if-then-fi placement mistakes.

[ownedbycats]

I'm having some trouble with using an inlist with square brackets, e.g.

program:
coll = $#kobocoll;

first_non_empty
   (
	if '[Send to Device]' inlist coll then 'sync.png' fi,
	if "[Kobo Store]" inlist coll then 'drm-locked.png' fi,
   )

incorrectly marks every book with sync.png (and if I comment the line out, then drm-locked.png). It doesn't happen if I remove the square brackets. Technically it's still working as it's not exact and still matches, but what is going on?

EDIT: I think I know what's happening. I forgot about regex. Putting in slashes to.... break (I forgot the actual word) them makes it work.

EDIT: Is there a better way to merge three lists then this? I'm a bit surprised that list_union doesn't handle more than 2 lists.

new_entries = list_union(first_list, second_list, ',';
new_list = list_union(new_entries, third_list, ',')

[chaley]

As you determined, you need to use backslashes to escape the square brackets that indicate character classes in regular expressions.

The list_union() function takes two lists. If you need to do a union of more than two lists then you do

    x = list_union(list1, list2, ',');
    x = list_union(list3, x, ',')

etc, or

    x = list_union(list1, list_union(list2, list3, ','), ',')

[SLaSZT]

Maybe a dumb question, but how can I set up a custom sort of a custom column? I literally just realized today that custom columns are a thing so I don't know if there's simple code that already does this.

I have a custom column called Status that contains the fixed text values of Reading, Priority, Paused, and Read. I'd like them to sort in that order (descending):

1. Reading
2. Priority
3. Paused
4. Read

Right now, it seems to be sorting alphabetically. I think that I could probably force it to sort the way I want by explicitly numbering them, but I don't want to have that numbering visible in the column.

Thanks in advance.

[chaley]

Originally Posted by SLaSZT

Maybe a dumb question, but how can I set up a custom sort of a custom column? I literally just realized today that custom columns are a thing so I don't know if there's simple code that already does this.

I have a custom column called Status that contains the fixed text values of Reading, Priority, Paused, and Read. I'd like them to sort in that order (descending):

1. Reading
2. Priority
3. Paused
4. Read

Right now, it seems to be sorting alphabetically. I think that I could probably force it to sort the way I want by explicitly numbering them, but I don't want to have that numbering visible in the column.

Thanks in advance.

You can't set a custom sort. A column is sorted depending on its type. Text columns will sort alphabetically.

As far as I can see, your choices are:

• Pick words that you understand their meaning and that sort the way you want. It seems that if you replace "Read" with "Finished" then you have your sort.
• Prefix with numbers or characters that make it sort the way you want.
• Use two columns, one that sorts correctly and a "column built from other columns" that displays the text you want based on the first column's value.

[ownedbycats]

When I run the following template on a book that has "Not Set" in my #chapters (composite) column, it gives me EXCEPTION: could not convert string to float: 'Not Set'.

I thought that the check on line 7 would prevent it. Did I make a mistake and leave out an else or misunderstand the != compare?

program:
f = re($#chapters, '(.*)/.*', '\1');
s = re($#chapters, '.*/(.*)', '\1');
newpercent = round(multiply ((f / s), 100));

	if 
		$#chapters != 'Not Set'
			then 
## updates #percentread for updated fanfics comparing no. of read chapters to total. 
## fanfic_unreadchaps() check for unread chaps based on kobo bookmark.
				if and(
					fanfic_unreadchaps(),
					$$#percentread ==#100,
					) 
					then 
						if 
							newpercent <#99 
						then 
							newpercent
						else 
							98 
						fi 
				else 
					$$#percentread
				fi

		else $$#percentread
	fi

EDIT: Commenting out line four makes it return #percentread (expected result). Was it trying to do math even if there was nothing matching for f and s? I guess I need to fix the newpercent bit somehow or rearrange the template so that lines 2-4 run after the 'not set' check.

[chaley]

Originally Posted by ownedbycats

EDIT: Commenting out line four makes it return #percentread (expected result). Was it trying to do math even if there was nothing matching for f and s? I guess I need to fix the newpercent bit somehow or rearrange the template so that lines 2-4 run after the 'not set' check.

The re() function isn't doing what you think it does. If the regular expression in

s = re($#chapters, '.*/(.*)', '\1')

matches nothing then s is assigned the value of $#chapters. You need to account for what happens if the pattern doesn't match the value. One way to do it is to verify that there is a '/' in $#chapters before extracting the values and doing the computation.

Yes, it will try to do the math with the values in s and f no matter what those values are.

You can use the debugger and breakpoints to look at the values of the variables and expressions to better get a handle on what your template is doing.

[ownedbycats]

Yes I was looking at my fanfic_unreadchaps() as it was similar and for that one I had it check that '/' was present -- there might be a better way but it was simple and accurate enough. So either I can insert that check or move lines 2-4 further down, perhaps below the unreadchap check itself.

EDIT: I found it easier to move the lines. Also tweaked the unreadchaps code to return the number instead of just 'true' -- isn't much effort to change existing templates and might be useful to have the exact number.

program:

	if 
		$#chapters != 'Not Set'
			then 
				if and(
						fanfic_unreadchaps() >=# 1,
						$$#percentread ==#100,
						) 
					then 
						f = re($#chapters, '(.*)/.*', '\1');
						s = re($#chapters, '.*/(.*)', '\1');
						newpercent = round(multiply ((f / s), 100));
					
						if newpercent <#99 then newpercent
						else 98 
						fi 
				else 
					$$#percentread
				fi

		else $$#percentread
	fi

[ownedbycats]

Okay, here's something semi-related to the last one:

program:
	if $#fanficcat && '/' in $#chapters then
		f = re($#chapters, '(.*)/.*', '\1');
		s = re($#chapters, '.*/(.*)', '\1');
		if s - f >=# 1 then s - f fi
	fi

Is there a less kludgey way to handle line 5 so that '0' returns an empty value?

[chaley]

Originally Posted by ownedbycats

Okay, here's something semi-related to the last one:

program:
	if $#fanficcat && '/' in $#chapters then
		f = re($#chapters, '(.*)/.*', '\1');
		s = re($#chapters, '.*/(.*)', '\1');
		if s - f >=# 1 then s - f fi
	fi

Is there a less kludgey way to handle line 5 so that '0' returns an empty value (e.g. if fanfic_unreadchaps() then 'output' type things)

What I would do is:

program:
	res = '';
	if $#fanficcat && '/' in $#chapters then
		f = re($#chapters, '(.*)/.*', '\1');
		s = re($#chapters, '.*/(.*)', '\1');
		if s > f then
			res = s - f
		fi
	fi

You can get rid of 'res' if the result of the template is the value of the if

[ownedbycats]

That works. I didn't realize it was possible to re-use a variable name and have it return different things. That will be useful to know.