Automatically exported from code.google.com/p/text-munger
Text Munger chews up sources via a variety of configurable methods.
It is a work-in-progress. See Roadmap for more details.
See http://www.xradiograph.com/WordSalad.TextMunger for my dis-organized, ranting notes....
This will probably be converted to JavaScript/node.js
There is no defined goal for TextMunger, beyond being useful.
For certain limited definitions of the word "useful." Details RECENTLY COMPLETED TASKS
improved display in selection editors 2012.04.16
Library, rule-selectors and editors.
FreeVerse transformation rule implementation 2012.04.16
which combines standalone ShortLines and InitialSpaces transformations
HeijinianAidToMemory transformation implementation 2012.04.17
indents lines according to alpha offset
save and reload text-sources (if in library or local files) 2012.03.28
basic editing control for SOURCE and OUTPUT 2012.03.27
This was pretty-much built-on, only now it's accessible via context-menu
added SNIPPETS window 2012.03.27
and copy-to-snippets from OUTPUT via context-menu
put density inside of XRML format 2012.03.18
load SOURCE from file 2012.03.19
save OUTPUT to file 2012.03.19 TASKS (in no particular order)
save and reload configured rule-sets
partially working as of 2012.04.12, but needs improvement, including
named rule-sets
auto-loading of saved-rulesets (in predefined location)
some sort of short-cut key to edit a given ruleset(s)
there's a lot of clicking in this interface.
build unit-tests
added percentage bias to text sources
although this can be crudely-done via adding the text more than once
swap OUTPUT to SOURCE, for re-processing
allow controls to resize when form is resized
separation of operation from GUI, so can be scripted
percentage-bias for word-level Transforms
i.e., don't apply to EVERY word
installation project, so installer can exist in project alongside source
better movement-controls inside of edit-areas
i.e., word-jump, etc
(better) error-handling. Right now, TM expects loaded data to be valid, etc. This task probably goes hand-in-hand with
add log4net
apply rule/ruleset to selected text in OUTPUT
re-think of rule selection/layout -- too many clicks, currently.
common tokenizer -- word, char, sentence, other. punctuation-sensitive or not.
Attempt to delete Gutenberg boilerplate
portamnteaux list, and other analysis of output (a la charNG) Transformation Rules Additions and Enhancements
vowell-to-punct rule: randomize the punctuation -- a la expletive deleted
BOWDLERIZER -- configurable regex rules, transforms words or phrases into something else
more Transformations
acrostic/mesostic generator
e-cummings-ification
more clustering random-walk for Density transform
file-based generic replacement "translator" to read from directory
ie, multiple user-controlled translators are possible
that will be an issue for serialization....
wait, I think I did this....
formatting category for rules (ie, XRML, shortlines, Heijinian, etc.)
more formatters (Howlish paragraphs, right-justify position `n`?)
custom rules have scripting of some kind
Markov updates
drop xray-references in Markov rules
add explanation of rules to rule-editor (and whatever background code)
change multiple-Markov rules to one rule, with options
make Markov-analysis case-insensitive for analysis
not for output (optional, I suppose)
optional WAAAY Down the road, but a major goal:
planar-output and processing, not just linear There are three basic types of transformation rules available:
Formatters
Translators
Generators Generators
Markov (n-gram) Chains Translators
Leet
Pig Latin
Shuffle
Disemconsonant
Disemvowell
Random Caps
Reverse
Shouty Caps
Vowell to Punctuation
Homphonerizer
Punctuize Whitespace Formatters
Short Lines
Initial Spaces
Free Verse
Heijinian Aid to Memory
XraysMonaLisa?-style