Experiment in computer generated poetry.
deadbok_cover_me
deadbok
Overview
Abstract
Cover me.
=========
This is a Markov chain generator seeded with about 16000 lines of David Bowie
lyrics.
Why?
----
When i was young and my heart was an open book, I used to write something I
defined as poetry. Life and my world view has since become way more diverse, to
the degree that I have not written more than scattered lines for many years, for
fear that I might oversimplify things.
When David Bowie died I had the urge to write something to commemorate him,
but that would be entering even more dangerous ground. A friend of mine, had
just discovered Markov chain generators, and told me about them. I knew these,
but had newer implemented one. Thinking of David Bowie's use of cut up technique,
an idea formed as to how I could get around writing something that made me
feel narrow minded and stupid. Let the computer do it, it is basically stupid
anyway.
How?
----
``wstat.py`` creates a dictionary of word associations and number of
occurrences needed for the generator. Since I dare not include the actual lyrics
of David Bowie's songs, this program saves the words and statistics in JSON
format, that the Markov chain generator can digest.
usage: wstat.py [-h] [-o OUTFILE] [-v] corpusfile
positional arguments:
corpusfile Text file to process.
optional arguments:
-h, --help show this help message and exit
-o OUTFILE, --output OUTFILE
Write output to a this file.
-v, --verbose Be verbose.
``markov.py`` is the actual Markov chain generator, that outputs something
that with a little love, might actually end up being
quite readable, sometime even touching, funny, and beautiful. It needs a text
corpus JSON file like the ones generated by ``wstat.py``.
usage: markov.py [-h] [-v] [-o OUTFILE] [-t TPL_FILE] corpusfile
positional arguments:
corpusfile Text file to process.
optional arguments:
-h, --help show this help message and exit
-v, --verbose Be verbose.
-o OUTFILE, --output OUTFILE
Write output to a this file.
-t TPL_FILE, --template TPL_FILE
Use this file as template.
Templates.
----------
Templates may be used to control the output of the Markov generator. The
templates are text files. There are a few command tokens, to control the
generator. Anything that is not recognised as a command is copied to the output
file.
Because of the way I handle newlines there is a quirk when starting a template.
* If there is no newline at the beginning of the template, the first block
will not be capitalised.
* If there is a newline at the beginning of the template, the first block
will not be capitalised.
Command syntax:
* ``{n}``: Use the Markov generator to insert ``n`` words.
* ``{!n}``: Same as above, but newlines are not allowed.
* ``{.n}``: Use the Markov generator to insert ``n`` words, capitalise first letter.
* ``{-n}``: Copy a Markov generated block ``n`` from the top.
* ``{rl,w}``: Rhyme on word on line l, word w.
Output examples.
----------------
**Welcome to the world**
Welcome to the world
Welcome to the baby,
You gotta have
Let him loose, chi child
Yourself
Deafen me
Absolute
Welcome to the world
Welcome to the funhouse
Away our heads
Mouth
Saviour machine
Work's down
Is the possibilities it
**Welcome to the world**
Welcome to the world
Welcome to the dreamers
And your window
All is a suffragette city
Spring went
Shopsdown
Welcome to the world
Welcome to the wagon
It's only dancing
Destruction as the dames
Leisure
The way into gloom
**Falling in**
Dredging the pain, in vain sheets on refugees changed
Say he pinned you, might have to touch the bushland over now we
Tip get papa chains
Beggar's face show respect even i am
Daydreams made of hit him right on garbo's
Sailor who would they don't grab that
Fingerprints will come quick on high
Vain showbiz anyway beautiful baby baby got, Moon
Knowing right from day falling in you sleep!
Sweeping over stiles woman, now tomorrow calls me to the
Nightclubbing we're gonna pay, and
Loneliness in town crazies with name york's,
Reelin' like daisies, lie behind my day after today
Romantic as forever i'm glad i've
Sailor who would they don't grab that
Fingerprints will come quick on high
Vain showbiz anyway beautiful baby baby got,Moon
Knowing right from day falling in you sleep!
Devil breaking!
Yards at!
Utter destruction!
**Good again for**
Enemy dropped a god-awful small enough to her
Threading daisies lie low with their eyes
Flipping time in a look into the flame filled sunset
Bullshit from my head up there was so enticing to face, frightening, do-doh
Elvis is my god
Wanking to their whiskey voice croaked
Dot marks
=========
This is a Markov chain generator seeded with about 16000 lines of David Bowie
lyrics.
Why?
----
When i was young and my heart was an open book, I used to write something I
defined as poetry. Life and my world view has since become way more diverse, to
the degree that I have not written more than scattered lines for many years, for
fear that I might oversimplify things.
When David Bowie died I had the urge to write something to commemorate him,
but that would be entering even more dangerous ground. A friend of mine, had
just discovered Markov chain generators, and told me about them. I knew these,
but had newer implemented one. Thinking of David Bowie's use of cut up technique,
an idea formed as to how I could get around writing something that made me
feel narrow minded and stupid. Let the computer do it, it is basically stupid
anyway.
How?
----
``wstat.py`` creates a dictionary of word associations and number of
occurrences needed for the generator. Since I dare not include the actual lyrics
of David Bowie's songs, this program saves the words and statistics in JSON
format, that the Markov chain generator can digest.
usage: wstat.py [-h] [-o OUTFILE] [-v] corpusfile
positional arguments:
corpusfile Text file to process.
optional arguments:
-h, --help show this help message and exit
-o OUTFILE, --output OUTFILE
Write output to a this file.
-v, --verbose Be verbose.
``markov.py`` is the actual Markov chain generator, that outputs something
that with a little love, might actually end up being
quite readable, sometime even touching, funny, and beautiful. It needs a text
corpus JSON file like the ones generated by ``wstat.py``.
usage: markov.py [-h] [-v] [-o OUTFILE] [-t TPL_FILE] corpusfile
positional arguments:
corpusfile Text file to process.
optional arguments:
-h, --help show this help message and exit
-v, --verbose Be verbose.
-o OUTFILE, --output OUTFILE
Write output to a this file.
-t TPL_FILE, --template TPL_FILE
Use this file as template.
Templates.
----------
Templates may be used to control the output of the Markov generator. The
templates are text files. There are a few command tokens, to control the
generator. Anything that is not recognised as a command is copied to the output
file.
Because of the way I handle newlines there is a quirk when starting a template.
* If there is no newline at the beginning of the template, the first block
will not be capitalised.
* If there is a newline at the beginning of the template, the first block
will not be capitalised.
Command syntax:
* ``{n}``: Use the Markov generator to insert ``n`` words.
* ``{!n}``: Same as above, but newlines are not allowed.
* ``{.n}``: Use the Markov generator to insert ``n`` words, capitalise first letter.
* ``{-n}``: Copy a Markov generated block ``n`` from the top.
* ``{rl,w}``: Rhyme on word on line l, word w.
Output examples.
----------------
**Welcome to the world**
Welcome to the world
Welcome to the baby,
You gotta have
Let him loose, chi child
Yourself
Deafen me
Absolute
Welcome to the world
Welcome to the funhouse
Away our heads
Mouth
Saviour machine
Work's down
Is the possibilities it
**Welcome to the world**
Welcome to the world
Welcome to the dreamers
And your window
All is a suffragette city
Spring went
Shopsdown
Welcome to the world
Welcome to the wagon
It's only dancing
Destruction as the dames
Leisure
The way into gloom
**Falling in**
Dredging the pain, in vain sheets on refugees changed
Say he pinned you, might have to touch the bushland over now we
Tip get papa chains
Beggar's face show respect even i am
Daydreams made of hit him right on garbo's
Sailor who would they don't grab that
Fingerprints will come quick on high
Vain showbiz anyway beautiful baby baby got, Moon
Knowing right from day falling in you sleep!
Sweeping over stiles woman, now tomorrow calls me to the
Nightclubbing we're gonna pay, and
Loneliness in town crazies with name york's,
Reelin' like daisies, lie behind my day after today
Romantic as forever i'm glad i've
Sailor who would they don't grab that
Fingerprints will come quick on high
Vain showbiz anyway beautiful baby baby got,Moon
Knowing right from day falling in you sleep!
Devil breaking!
Yards at!
Utter destruction!
**Good again for**
Enemy dropped a god-awful small enough to her
Threading daisies lie low with their eyes
Flipping time in a look into the flame filled sunset
Bullshit from my head up there was so enticing to face, frightening, do-doh
Elvis is my god
Wanking to their whiskey voice croaked
Dot marks
Authoring entity
| Agency Name | Role |
|---|---|
| deadbok | owner |
Language
English
Methods, software and scripts
Software
| Name | Libraries or packages used |
|---|---|
| Python |
json, pronouncing, random, argparse, re, collections
|
| Smarty |
License
| Name |
|---|
| GNU General Public License v3.0 |
Metadata production
Producers
| Name | Role |
|---|---|
| GitHub Bot | bot |
Date of Production
01 December 2019