[GUI Plugin] Action Chains

[chaley]

@capink: what do you think about a template function

is_book_var_defined('var_name')

that returns '1' or ''.

The function would get around the ambiguity between a defined book_var that contains '' and an undefined book_var. Something like this:

	if is_book_var_defined('foo') then
		result = book_vars('foo')
	else
		result = field('whatever')
	fi;
	result

[capink]

Yes, I can see the use for this. Will add to the next release.

[capink]

On a second thought, wouldn't a defined_book_vars() that returns a list of defined vars for book cover more ground? We can test for the above using:

str_in_list(defined_book_vars(), ',', 'var_name', 1, '')

[chaley]

Originally Posted by capink

On a second thought, wouldn't a defined_book_vars() that returns a list of defined vars for book cover more ground? We can test for the above using:

str_in_list(defined_book_vars(), ',', 'var_name', 1, '')

Not sure.

On the plus side: in addition to checking for individual book vars, your suggestion has the advantage of being able to ask about more than one variable at once, as in

list_contains(defined_book_vars(), ',', '^(var_name1|var_name2)$', 1, '')

or check for prefixes

list_contains(defined_book_vars(), ',', '^var_name.*$', 1, '')

However, it is a bit more complicated to use in the simple case.

I think I lean in the same direction you do: the generality is better. And anyway, if someone really wants to use the original suggestion then they can do it with a stored template.

program:
	arguments(var_name);
	if str_in_list(defined_book_vars(), ',', 'var_name', 1, '') then
		'1'
	else
		''
	fi

PS: I am thinking about adding "return" to the template language, which would allow the above template to be written

program:
	arguments(var_name);
	if str_in_list(defined_book_vars(), ',', 'var_name', 1, '') then
		return('1')
	else
		return('')
	fi

[chaley]

Another topic: while playing with actions today I noticed that some of the template dialogs, for example the Template Scope one, have extra fields related to stored templates, specifically the template name and the documentation. I am changing TemplateDialog to permit you to hide them.

With the changed TemplateDialog I tested this in Action Chains.templates.py and it works. Starting at line 404:

        TemplateDialog.__init__(self,
                                parent,
                                text,
                                mi=mi,
                                text_is_placeholder=text_is_placeholder,
                                all_functions=self.all_functions,
                                builtin_functions=builtin_functions,
                                global_vars=global_vars)
        self.setup_saved_template_editor(show_buttonbox=not dialog_is_st_editor,
                                         show_doc_and_name=False)

I will let you know when the changes are in calibre source.

[capink]

Originally Posted by chaley

PS: I am thinking about adding "return" to the template language, which would allow the above template to be written

That would be useful addition.

Originally Posted by chaley

On the plus side: in addition to checking for individual book vars, your suggestion has the advantage of being able to ask about more than one variable at once, as in

This is what I had in mind when thinking about this. The prefix thing is another good use for this.

Originally Posted by chaley

Another topic: while playing with actions today I noticed that some of the template dialogs, for example the Template Scope one, have extra fields related to stored templates, specifically the template name and the documentation. I am changing TemplateDialog to permit you to hide them.

With the changed TemplateDialog I tested this in Action Chains.templates.py and it works. Starting at line 404:

        TemplateDialog.__init__(self,
                                parent,
                                text,
                                mi=mi,
                                text_is_placeholder=text_is_placeholder,
                                all_functions=self.all_functions,
                                builtin_functions=builtin_functions,
                                global_vars=global_vars)
        self.setup_saved_template_editor(show_buttonbox=not dialog_is_st_editor,
                                         show_doc_and_name=False)

I will let you know when the changes are in calibre source.

Yes, would make it a lot better and less confusing. Will be waiting for the changes.

[chaley]

Originally Posted by capink

Yes, would make it a lot better and less confusing. Will be waiting for the changes.

The changes are in calibre source.

I also changed the dialog layout a lot. Let me know if you see anything wrong/strange.

[capink]

Originally Posted by chaley

The changes are in calibre source.

I also changed the dialog layout a lot. Let me know if you see anything wrong/strange.

It is working great.

[Terisa de morgan]

Thank you both of you, I'll try to make it work tomorrow, and I'll be back with my misadventures for sure

[ownedbycats]

I found a tyop!

[capink]

Version 1.7.0

• Update: Add persistent storage for templates. Can be accessed using these new template functions: persistent_vars(), set_persistent_vars(), list_persistent_vars(), delete_persistent_vars(), list_persistent_namespaces(), delete_persistent_namespaces()
• Update: Copy To Clipboard: Add option to write to a file.
• Update: Chain Variables: Add option to ask for variables at runtime.
• Update: Library View Double Clicked: tweak to allow single click edit (action_chains_single_click_edit).
• Update: Add a tweak to allow disabling submenu inheriting first action icon (action_chains_default_sub_menu_icon).
• Update: Update the way table state is saved to avoid messing it up every-time a new column is added.
• Update: Add options to customize scope icon and tooltip.
• Update: Add new template function: list_book_vars()

The new persistent storage allows template functions to save data permanently. Here is how you can save a variable to persistent storage:

set_persistent_vars('name', 'value')

It also supports namespace hierarchies as well for more flexible structure. For example if you want to save the same variable name but with a different value for each library, use the third argument as the namespace:

set_persistent_vars('name', 'value', current_library_name())

If you want to store permanent book vars (the same variable with different value per book per library), use a hierarchical namespace (in the form of library_name.book_id):

ns = strcat(current_library_name(), '.', field(id));
set_persistent_vars('name', 'value', ns)

You retrieve the value in the same way

ns = strcat(current_library_name(), '.', field(id));
persistent_vars('name', ns)

@ownedbycats: The new file option in the Copy To Clipboard action might solve your problems with command line limit if the program you are using supports reading filenames from a file. You can output you data to a file using Copy To Clipboard, and feed the file to the Open With action using the template option.

Finally, the action_chains_single_click_edit tweak makes it possible still edit by single clicking, which you lose when you set the tweak (in case you want to connect the double click event to a chain):

doubleclick_on_library_view = 'do_nothing'

Unfortunately, there is bug there where it also open the field editor when double clicking as well. I don't know the cause for this, so if anyone can help with this.

[chaley]

Originally Posted by capink

Finally, the action_chains_single_click_edit tweak makes it possible still edit by single clicking, which you lose when you set the tweak (in case you want to connect the double click event to a chain):

doubleclick_on_library_view = 'do_nothing'

Unfortunately, there is bug there where it also open the field editor when double clicking as well. I don't know the cause for this, so if anyone can help with this.

I don't see this behavior. I set the calibre tweak

doubleclick_on_library_view = 'do_nothing'

I then set the plugin tweak

action_chains_single_click_edit = True

A single click in the library view edits the cell. A double click in the library view runs the action. Isn't this what is supposed to happen?

[capink]

Originally Posted by chaley

A single click in the library view edits the cell. A double click in the library view runs the action. Isn't this what is supposed to happen?

Yes, that is what supposed to happen. Have you tried double clicking on an already selected book? It does not happen when you double click on book that is not selected.

[chaley]

Originally Posted by capink

Yes, that is what supposed to happen. Have you tried double clicking on an already selected book? It does not happen when you double click on book that is not selected.

Yes. For example:

1. Select a book by single-clicking on a cell such as title. The line turns blue. No editor opens. No action runs.
2. Single-click on a cell on that line, for example authors. The editor opens.
3. Double-click on a cell, for example authors. The action runs.

Inverting the order of the last two steps doesn't change anything.

[capink]

Originally Posted by chaley

Yes. For example:

1. Select a book by single-clicking on a cell such as title. The line turns blue. No editor opens. No action runs.
2. Single-click on a cell on that line, for example authors. The editor opens.
3. Double-click on a cell, for example authors. The action runs.

Inverting the order of the last two steps doesn't change anything.

When I reach number 3, the action runs and the cell opens for editing.

After more testing, it seems to be chain dependent. When I tested with some chains that contain only one action, it behaved normally. The chain I was initially testing with (attached below) has two conditional actions and it always triggers the abnormal behavior.

This might suggest that the problem is somewhere in my code? Not really sure. Will investigate more and see if I get to the bottom of this.

Edit: After testing some more, I think I found the culprit. The chain loop makes this call between actions:

QApplication.processEvents()

If I comment this out, the problem does not happen anymore. I don't know calling processEvents() leads to this?