Slow txt to mobi convertion, performance at o(n^2) as lines of txt grow?

forceps · 11-18-2012, 03:00 PM

When convert plain text to mobi, the time used grow more rapidly as text grows. It is roughly like:

T = k * n ^ 2

where T is the time used, n is the total lines of text file, and k is a constant between 2 to 3 on my system.

In my case, each line of text file is converted to a … paragraph, If for each … , Calibre try find its parent during convertion, presumingly by search everyline before that ..., then n * (n + 1) /2 search need to be done, that might be an explaination.

May I suggest add performance tuning to the future development?

theducks · 11-18-2012, 03:25 PM

You did not offer to write better code.

Moderator Notice
Please read the sticky before posting in Development. https://www.mobileread.com/forums/sho...d.php?t=122042

mrmikel · 11-19-2012, 06:24 AM

You might consider using some other program, such as Open Office or even Sigil, to establish the paragraphs yourself. Then Calibre would be much faster.

After all, how is Calibre supposed to know where the paragraphs are supposed to be? HTML input is one format Calibre is happy with.

forceps · 11-19-2012, 03:39 PM

How about an option of convert plain txt to "plain" mobi, no sections, no fonts, just like read a plain txt file, but in mobi format.

I tried to create the html from a 10M txt file with a simple python script. Inside the <body> tag, there are only .., or only , then feed that html to ebook-convert, it still runs not fast enough.

I wish I could contribute, but the task looks daunting for my skill level...

cybmole · 11-20-2012, 08:36 AM

try import text to MS word, save as filtered HTML, then have calibre convert html to mobi - that may go faster even though it is 2 step. if you don't like word, use some other program that can save as html / epub e.g. sigil.

a 10M text file. sounds awfully big though!, unless I am misunderstanding your unit of measurement.

a long novel, formatted as a text file, would be less than 1 Mb

user_none · 11-22-2012, 10:40 PM

Set the paragraph and formatting style manually. Do not use heuristic for the formatting. This will reduce the amount of processing calibre does to A) determine the paragraph style and B) format the text.

Heuristic formatting uses, you guessed it, heuristics and the larger the text the more it needs to process. Most heuristics are a series of regular expressions and increasing the amount of text will drastically slow down the process. Every regex needs to run over the entire document. So every line you add every regex needs to run over that much more text.

Quote:

Originally Posted by forceps

How about an option of convert plain txt to "plain" mobi, no sections, no fonts, just like read a plain txt file, but in mobi format.

Again, actually look at the options that are available. Formatting style: plain. This will do what you're asking for.

forceps · 11-26-2012, 11:47 AM

after set:
--formatting-type plain --input-encoding utf8 --markdown-disable-toc --paragraph-type off

conversion time of a 10M txt to mobi is reduced to less than 2 minutes, a 20M txt takes 14 minutes. It is much better.

Thanks.

11-18-2012, 03:00 PM	#1
forceps Enthusiast Posts: 26 Karma: 168 Join Date: May 2005 Location: Wuhan, China Device: Kindle DXG	Slow txt to mobi convertion, performance at o(n^2) as lines of txt grow? When convert plain text to mobi, the time used grow more rapidly as text grows. It is roughly like: T = k * n ^ 2 where T is the time used, n is the total lines of text file, and k is a constant between 2 to 3 on my system. In my case, each line of text file is converted to a <p> … </p> paragraph, If for each <p> … </p>, Calibre try find its parent during convertion, presumingly by search everyline before that <p>..., then n * (n + 1) /2 search need to be done, that might be an explaination. May I suggest add performance tuning to the future development?

11-19-2012, 03:39 PM	#4
forceps Enthusiast Posts: 26 Karma: 168 Join Date: May 2005 Location: Wuhan, China Device: Kindle DXG	How about an option of convert plain txt to "plain" mobi, no sections, no fonts, just like read a plain txt file, but in mobi format. I tried to create the html from a 10M txt file with a simple python script. Inside the <body> tag, there are only <p>..<p>, or only <br>, then feed that html to ebook-convert, it still runs not fast enough. I wish I could contribute, but the task looks daunting for my skill level...

Thread Tools	Search this Thread
Show Printable Version Email this Page	Search this Thread: Advanced Search

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
txt to mobi how to	codrutoctavian	Conversion	7	01-24-2012 10:42 PM
Convertion error txt to epub "IndexError: list index out of range"	economix	Conversion	6	12-25-2011 06:14 AM
.txt to .mobi	BroCraig	Conversion	9	03-10-2011 02:40 PM
txt to mobi - dashes becoming ?	cybmole	Calibre	5	10-14-2010 11:02 AM
inserting blank lines into rtf/txt/html	errata	Ectaco jetBook	7	07-10-2010 09:16 PM

11-18-2012, 03:25 PM	#2
theducks Well trained by Cats Posts: 30,873 Karma: 59840450 Join Date: Aug 2009 Location: The Central Coast of California Device: Kobo Libra2,Kobo Aura2v1, K4NT(Fixed: New Bat.), Galaxy Tab A	You did not offer to write better code. Moderator Notice Please read the sticky before posting in Development. https://www.mobileread.com/forums/sho...d.php?t=122042

11-19-2012, 06:24 AM	#3
mrmikel Color me gone Posts: 2,089 Karma: 1445295 Join Date: Apr 2008 Location: Central Oregon Coast Device: PRS-300	You might consider using some other program, such as Open Office or even Sigil, to establish the paragraphs yourself. Then Calibre would be much faster. After all, how is Calibre supposed to know where the paragraphs are supposed to be? HTML input is one format Calibre is happy with.

11-20-2012, 08:36 AM	#5
cybmole Wizard Posts: 3,720 Karma: 1759970 Join Date: Sep 2010 Device: none	try import text to MS word, save as filtered HTML, then have calibre convert html to mobi - that may go faster even though it is 2 step. if you don't like word, use some other program that can save as html / epub e.g. sigil. a 10M text file. sounds awfully big though!, unless I am misunderstanding your unit of measurement. a long novel, formatted as a text file, would be less than 1 Mb

11-26-2012, 11:47 AM	#7
forceps Enthusiast Posts: 26 Karma: 168 Join Date: May 2005 Location: Wuhan, China Device: Kindle DXG	after set: --formatting-type plain --input-encoding utf8 --markdown-disable-toc --paragraph-type off conversion time of a 10M txt to mobi is reduced to less than 2 minutes, a 20M txt takes 14 minutes. It is much better. Thanks.

Advert

Advert