[GUI Plugin] Import List

[capink]

The developer of this plugin is inactive. The xpath settings for amazon has been changed and the plugin has not been updated with these changes. I've done some changes to the plugin recently and I plan on cleaning up the predefined list. In the mean time try putting these values in the webpage tab:

Rows: //li[@class="zg-item-immersion"]
Title: .//div[@class="p13n-sc-truncate p13n-sc-line-clamp-1"]/text()
Authors: .//a[@class="a-size-small a-link-child"]//text()[normalize-space()]

[floyd68]

Originally Posted by capink

The developer of this plugin is inactive. The xpath settings for amazon has been changed and the plugin has not been updated with these changes. I've done some changes to the plugin recently and I plan on cleaning up the predefined list. In the mean time try putting these values in the webpage tab:

Rows: //li[@class="zg-item-immersion"]
Title: .//div[@class="p13n-sc-truncate p13n-sc-line-clamp-1"]/text()
Authors: .//a[@class="a-size-small a-link-child"]//text()[normalize-space()]

That works!
Now I have to figure how to get rid of the "predefined" links that don't work... like all the "Book Chart" links.

[floyd68]

Is it possible to remove "predefined" lists which either:
1. Don't work
2. I don't want

For example:
BookChart domain is no longer working so I would like to remove all references in "Predefined".

THANK YOU IN ADVANCE!

[floyd68]

Disregard... I found it in config.py

I do appreciate the help earlier!

[floyd68]

Ok... I have tried to figure this out.
I opened pages that WERE pulling data (Amazon) and used "inspect Element" to look for the "Row" / "Title" information uses so I could fix my own problems.

I discovered that I am not smart enough.

Does anyone have working codes to pull information from the NY Times "Predefined" group?

[capink]

I already have these fixed in a version that I will post on this thread later. Dead entries will be removed as well.

NY Bestsellers
Rows: //ol/li
Title: .//h3/text()
Authors: .//article/div/a/p[2]/text()

[floyd68]

That would be AWESOME!!!!
I appreciate the work you have done on this!

[Rellwood]

Originally Posted by capink

The title/author match is the original way the plugin used to do automatic matching. The updates I've done has not changed anything related to title/match nor does it affect anything in the manual match. If you test with an earlier version I think you will get the same outcome you are getting with the title/match option in the updated version. If you give me steps to reproduce the problems you are having, I'll have a look at it.

Let me look into whether it was just in one library or in others. I haven't been working on my libary for a bit, so I wasn't able to "recreate" it in other libraries.

It might have to do with other elements ie, Calibre version.

[Rellwood]

I can't offer a way of indicating what would work and what won't work.

I have tried it on multiple Calibre Versions, and I even tried it with the older plugin. It's like none will work on a couple of books.

It's not just that it won't select the book, but it won't even list the book if I type in the exact name, or it won't offer the book in the list of books by the author.

I cleared out the plugin preferences in using Job Spy, I moved the books to another library, I even tried adding the book to the library with the plugin, then deleting that book.

For some books nothing I can do will match up.

I never had this problem with not finding the book after typing in the title/author properly.

What can I do? Is there some place where the plugin keeps information that prevents a book from showing even if it is manually typed in?

[Rellwood]

It might have to do with being title sort, dropping the leading "The, A, Of" etc. the plugin does. My library is sorted strictly alphabetic. I noticed that books that have the "The", or a ":" in the titles aren't matching.


Literally, the only available thing to do is to clear out the search bar and have it list every book in the library. I can find it that way.

[capink]

To be able to look at this I need the exact titles and authors of books causing troubles. You can use calibre create a catalog to export a csv of selected books. I will give it a look, but don't hold your breath.

[capink]

Version 1.4.9
Update: Add a progress dialog.
Update: Add datatype validation.
Update: Add context menu for exporting to csv.
Fix: title/author match not matching books with multiple authors even if identical.
Fix: Update predefined sources: Amazon, NY Bestsellers.
Fix: Removed dead predefined sources: BookChart, Amazon Listmania, Play.com, SFFJazz, WHSmith
Fix: Guess form headers now ignores duplicate headers.
Fix: Update universal line mode to use open newline arg.
Fix: Remove unused custom series indices headers.
Fix: Error when removing and re-adding encodings.
Fix: Buttons in encoding dialog now change their state based on selected item.

Datatype validation: This feature is added to avoid the problem of the plugin failing midway with an exception because a value was not suitable for a column it was supposed to go (e.g text value for column that only accepts int or float). If you are importing a large dataset this can happen several minutes — or hours — into the import process. Now the datatype validation checks before proceeding, and if errors are found; a validation error message appears offering details. It also highlights errors in the preview table with a red color, and offers a checkbox option to filter the rows with errors, so that you inspect them, and if you want delete them and proceed with the rest of data. Also the new export to csv feature allows you to export the rows containing the errors to a csv file to deal with them later.

The only type of validation the plugin used to do before is checking title field, and preventing the user from proceeding if an empty title was found, but without any details of what and where the failure is. The title field check is included in the datatype validation but with details and highlighting.

The datatype validation only catches errors that used to stop the import process. This includes int, float, rating, datetime and title fields. Errors that are ignored by the plugin and discarded or replaced by default values are not included in the datatype validation. Examples of these are: illegal language values which are replaced with UNDEFINED by the plugin, and illegal boolean values which are replaced by False, series_index for which invalid values are replaced with 0, ivalid enum types which are discarded by the plugin.

There is also a fix for a problem with the original title/author matching of the plugin not being able to match books with multiple authors. It is surprising this has not been reported until now given how long ago this plugin was developed. @Rellwood reported problems with author/title matching without specifying multiple authors being the problem. I don't know if this solves his issues.

As with the previous version you are advised to use this only on a test library until you are satisfied it can be used on a daily basis.

[Rellwood]

I noticed that the length/height of the "match columns" is very short, instead of a scrollbar, maybe allow for manual stretching? Also, maybe allow the tags and series columns to NOT be manditory in the output?

[capink]

There was a minor regression with the version I posted above. It is fixed now. You can download it from here. (link on the previous post also points to the fixed version).

[kcar181]

Hi guys Im trying to figure out a custom webpage

http://www.worldswithoutend.com

I figured some of it out but ive never used xpath before. Titles are fine but all the authors are adding combined. "Bob & Sue & Lee" instead of:
Bob
Sue
Lee



//*[@id="reportlist"]
.//p[@class="title"]/a/text()
.//p[@class="author"]/a/text()

This is what im using for page:
https://www.worldswithoutend.com/lists_banned_sff.asp