rkvsraman / autotags Goto Github PK

Combine all AutoTags components including blacklist and white list in a single 
JS file. Provide utility 
methods for adding to the lists rather than maintaining a separate JS file.

Support multiple sets of text with different weightings (field weights) e.g. 
title would be treated 
differently to body.

Add support for "associated" tags (e.g when suggesting 'lucene' also add 
'search', 'java').

As the summary says.

What steps will reproduce the problem?
1. type watch at the end of the demo text
2.
3.

What is the expected output? 
It took 9 ms. to generate tags from 119 words!
bayesian
bayesian inference
artificial intelligence
complex models
bayes theorem
monte carlo
metropolis hastings
beast
mrbayes
p4
techniques
applications

What do you see instead?
No words are returned, it looks like it crashed

What version of the product are you using? On what operating system?
1.2 demo page, firefox, windows

Please provide any additional information below.
failing for the same reason in my app
had to explicitly add

                if(token.indexOf("watch") >= 0 ){
                    continue;
                }

makes me wonder if other words will make it crash

cheers
[email protected]

At the moment the last variation of a term is shown.

Potentially apply further weighting based on position, applying more weighting 
for terms that 
appear towards the front of the text.

This is a minimal problem but there's some optimisations to be made.

What steps will reproduce the problem?
1. A text like HTML Js would not be picked as a tag while Html Js might
2. Also all abbreviations are skipped. "GOOG and YAHO" - a text like this 
both will be skipped 

What is the expected output? What do you see instead?
All abbreviations should come up as tags and there should be no 
restriction that the first letter should be cap and subsequent letters 
should be in small case.

What version of the product are you using? On what operating system?
v1.0 on win2k

Please provide any additional information below.

Separate out language specific elements, stopwords, regular expressions etc.

Remove cache functions.

What steps will reproduce the problem?
1. Include prototype.js in you example file
2. Click Analyze

What is the expected output? What do you see instead?
I'd like to see tags generated, but the process failed!


What version of the product are you using? On what operating system?
autotags 1.2; Firefox; Mint OS

Support term normalisation ( 'youtube', 'iFilm' -> 'video sharing' )