Better support for imprints in v1.3.6

Version 1.3.6 of PageKicker provides better support for “imprints”, which give a publisher the ability to provide distinctive style and branding for a set of similar books. The –imprint option allows the operator to select an imprint from the command line.  Imprint files stored in the conf/ directory allow the operator to specify global variable values for an imprint. By default, every imprint should have its own name, logo, copyright notice, and mission statement, but it is also possible to specify stylistic variables such as $covercolor, $coverfont, $epigraphfile, and so on, as well as substantive variables such as default APIs, summarization parameters, and the like.

Version 1.3.7 makes chapter sorting alphabetical (as designed) with the –expand_seeds_to_pages flag set to “yes”, and provides an unambiguous warning message when the master configuration file is missing.



Want to install PageKicker on a VM?

I’m working on making the install process for PageKicker seamless and bulletproof with an install script.  I’m going to start with getting the script running on a Ubuntu virtual machine, so that I can easily throw away failed attempts and start fresh.  After that’s working flawlessly, on to AWS and other platforms.

I am following these instructions to build the VM:

It has taken about 30 min so far and I have encountered no obstacles.

If you want to work on this along with me, shoot me a DM or email at fred at pagekicker doht com.

Five crucial things to know about algorithmic publishing, right now

1. Version 1.0 of  PageKicker’s open source algorithmic publishing suite has been released and the command-line application is now available on github.

In fact, this occurred two weeks ago, and we are already on version 1.3.5, “Samuel Johnson”, which added several important post-1.0 features including web & news search and multiple seed submission from the command line.
2.  The command-line application enables individuals and enterprises to create complete, always updating, machine-learning-enabled Chicago-style ebooks in a matter of seconds at a cost of pennies.
The system can search any permissioned content, assemble reports in the preferred structure, deliver in any digital book or document format, and distribute via  email, ftp, e-commerce, and ONIX.
Custom-labeled web, Magento, iOS, and Android UIs are available.
3. Supported and hosted deployments with training and customization packages are now available on Linux computers, virtual machines, and AWS instances.
This is a great time to be an early adopter. The software has reached an inflection point of maturity, PageKicker is now post-revenue, and we can save you tens of thousands of dollars in labor costs within weeks of deployment.
4. PageKicker’s WordPress site provides frequent updates on the growth of the product and the state of algorithmic publishing in general.
Increasingly, I will be relying on the wordpress site and my twitter account @fredzannarbor to get out the word. These emails will continue to be quarterly.
5.  Eight pictures are worth eight thousand words.
The operator launches the book build with a simple command line invocation. This can be incorporated in batches, alerts, or smart scripts.

Dozens of command line options are available that can configure every aspect of search, fetch, analyze, build, distribute.

A proprietary summarizer and numerous off-the-shelf analytic tools are included.

Complete books are created.

Each robot author is endowed with personality.

By default covers are created with word clouds, but this is customizable.
Docs are delivered in multiple formats to the file system or to email, web, or mobile clients.

This is all done in a matter of seconds, and can then be repeated whenever the user is ready to read.  Because the software, available content, and machine learning are always improving, the latest book build will always be significantly better than the previous ones.  Books can be built on any topic or any combination of topics.

Thank you for all your interest in PageKicker over the years, and please, let me know when you would like to begin your test deployment.

Deploying PageKicker on AWS

I am deploying PageKicker on an external AWS image with Ubuntu 16.04 and in the process automating the installation process. Hitherto it has been manual — I have successfully set the system  up on several different machines, including Ubuntu 12, 14, and 16; Macintosh; and Ubuntu Bash in Windows 10.  The installation script will make this a push button process.  If you would like to be an early tester on AWS let me know at fred at pagekicker doht com.

In the meantime, I have made two dot releases that make the system more portable and resolve bugs.

Version 1.3.4:

  • makes the search engine snippet function portable
  • fixes a regression that caused errors in creating .mobi documents

Version 1.3.5:

  • changes certain default font choices to be consistent with Ubuntu defaults
  • fixes a security issue caused by an ImageMagick vulnerability



PageKicker in action: Mary Tyler Moore instant book #amwriting

From the command line on your PageKicker installation:

bin/ --seedsviacli "Mary Tyler Moore" --expand_seeds_to_pages "yes" \
--booktitle "She Could Light the World Up ..."


PageKicker searches permissioned content & identifies relevant resources:








And delivers to your workspace for further editing.

Screenshot from 2017-01-25 17-05-51.png

All in 30 seconds, and infinitely scalable.


version 1.3.3 adds search engine and news snippets

PageKicker enhances its ability to provide comprehensive content for its books with version 1.3.3, which adds two new parts of the book as defaults at the beginning of the back matter: Search Engine Snippets and News Snippets.  These sections contain up to 10 snippets for each seed term in the book drawn, as you might expect, from web search and news indexes.  The design is agnostic as to which indexes. Future releases will add various ways of exploiting this additional content, such as spidering & analyzing the additional documents referenced by the snippets.

From one of our standard test books on Paella:


From a test book I built today on #AlternativeFacts:

bin/ --seedsviacli "Alternative Facts; Sean Spicer"\
 --booktitle "#AlternativeFacts"


version 1.2.3 fixes bug in display of Acronyms section

Version 1.2.3 of PageKicker fixes a bug whereby in the Acronyms section the ebook builder interpreted acronyms beginning with the # character s markdown, thus causing hashtag acronyms to create new section headings.

Thus #NotMyPresident. produced:


now becomes:


The trailing period in #NotMyPresident is an error in the acronym filter, although it is not immediately apparent to me how the filter would know to remove the period from acronyms appearing at the end of a sentence.

The current acronym filter is a regex that could be dramatically improved.

sed 's/[[:space:]]\+/\n/g' $txtinfile  | sort -u |\
 egrep '[[:upper:]].*[[:upper:]]' | sed 's/[\(\),]//g' | uniq

Contributions very welcome!

version 1.2.2-Johnson adds multiple seeds via cli

Version 1.2.2 of PageKicker adds the ability to submit multiple semi-colon separated seeds via the command line using the –seedsviacli option.  This may seem like a minor feature but it is pretty darned useful for administrators who can now fluently create multitopic books without the necessity of creating a separate seed file.   E.g.:

bin/ --seedsviacli "Women's March 2017; Alice Paul; \
Protests against Donald Trump; Tiny Hands (song)" \
--booktitle "All About the Women's March"

Result (15 sec later) in epub, mobi, and docx is a unique book bringing together information about the Women’s March; the Swarthmore-graduate feminist Alice Paul; the global #resistance movement; and the Women’s March theme song, “Tiny Hands.”

Remember, you can run thousands of these script commands in a day, on any topics, against any permissioned content.