blogwall at kent ridge mrt station janaka prasad 09/06/2008

20
BlogWall at Kent Ridge MRT Station Janaka Prasad 09/06/2008

Upload: teresa-fletcher

Post on 30-Dec-2015

217 views

Category:

Documents


0 download

TRANSCRIPT

BlogWall at Kent Ridge MRT Station

Janaka Prasad09/06/2008

Basic overview of the new application

Main parts of the application

SMS Reader Reads SMS from the modem and write to the

database SMS Processor

Process received SMS;• check banned words• locate synonyms• Tag the SMS using English POS (part of speech )

Tagger• create poetry• write back to database

Display Processing based application that displays or

Interpret data

Main parts of the application

Backend Application Provide management capabilities

• PHP website to add poetry to the system• Tool to extract web based poetry data to the

system• Tool to calculate IDF of the database• Tool to add emotional weights/ prefix to the

system• Tool to recalculate emotional weights of database• Manage poetry in the system

How it is going to work

In previous system everything was put together and you need to know how everything works before working with BlogWall

In new application all major parts are separated into different applications that run separately

This make it easy to work with the system E.g. – the artist doesn’t have to know how SMS work

or how to generate poetry. He is presented with all the data. All he needs to know is how to display it.

Data is passed from one application to the other via database

Each application has different level of complexity separating them make each of the cleaner

How it is going to work

Database

Project Plan

PHP website to add poetry to the system Task 2 – Completed Data entry operators can add poetry to the system using

web based application. PHP file called “blogwall.php” used to add the poetry to

the system Already setup in one of the computers in the lab and

Rashika has been adding poetry to the system Link: http://137.132.165.176/blogwall.php requires PHP (5), MySQL (5.0), and Apache (2.2) Data gather here goes to “poem_table”

Extract web based poetry data to the system Task 4 - completed The main task of PosTagger application is to transfer the

data collected by the PHP website into actual tables. PosTagger only process entries in “poem_table” field

“poem_processed” set to value 0. After processing it set it to 1 indicating it has been processed

Get Lines from PHP table“poem_table”

where poem_processed=0

Poetalready in

“poet”?Add poetNO

Last entry?End YES

NO

Poem already in“poem”?

YES

YES

For each line inPoem

NO

Add to “poemlines”

For each keyword inpoem line

Keyword in“keywords”?

Add Keyword NO

LastKeyword?

NO YES

Last Line?

NO

YES

Update with Emotionalweight

Add entry to “poemlines_keywords”Calculate TF,

calculate positional tag using English POSTagger

Set “poem_table”,poem_processed=1

Calculate IDF of the database Task 5 – Completed PosTagger only transfer the data to the system it can not calculate IDF. This tool will calculate IDF of the system. Every time new poems added to the system IDF has to be recalculated. The administrator only has to press start button to start the process.

Add/Edit emotional weights/ prefix Task 6 – Completed Administrators will use these tools to manage emotional weights and prefix

Recalculate emotional weights of database Task 7 – Completed Every time new poetry added to the system or new emotional words or prefix

added to the system Emotional weights need to be recalculated. The administrator only has to press start button to start this process.

Swear words

System needs to detect banned words in the SMS This tool will be used to manage them

Poetry manager

Manage poetry in the system

Reading SMS from modem

Task 9 – Completed Task 10 - Working on now I have rewritten this entire module I have tested the application with some lab members

simultaneously sending SMS

Init Modem

Read “settings.cfg” file to get portnumber

Open port

ErrorEnd YES

NO

ErrorEnd YES

NO

Set Modem to textmode

Delete all SMS

Read Modem output

OK or ERROR inmessage

NO

OK & +CMGL: & no ERROR inmessage

YES

Locate SMS messages

YES

Write to “sms” tablestatus = 0

Write to “sms_log” table

Delete all SMS

Get ALL SMSNO

Data entry

120 poems added to the database

Important issues …

Do we support multilingual poetry in Kent Ridge MRT?Processing can not display Unicode

textWe are currently building English

poetry database. We would need a large database of other poetry and English translation for it to work.

Getting the artist