AIKIF

Artificial Intelligence Knowledge Information Framework (Alpha)

This is an information classification framework that maps structured or freeform data to a standard knowledge store.

Manages a dataset of your applications, the source data, parameters, runs and results and then uses your business rules to convert and store the information in a machine usable format.

Your AI software can link to AIKIF by setting up logging watch-points to define success / failure along with the range of input parameters. Goals and plans are defined by breaking them down to smaller tasks until the task can be run by a tool in the Toolbox.

A tool is any python wrapped function or application and is easily extensible.

Quick Start

This github repository https://github.com/acutesoftware/AIKIF contains the latest code, but the current public release is available via

pip install aikif

To start the API server use aikif/api_main.py and run the tests/test_api.py

 * Running on http://127.0.0.1:5000/
 * Restarting with reloader
127.0.0.1 - - [28/May/2015 19:22:49] "GET /facts HTTP/1.1" 200 -
127.0.0.1 - - [28/May/2015 19:22:49] "GET /help HTTP/1.1" 200 -
127.0.0.1 - - [28/May/2015 19:22:49] "GET /users/1 HTTP/1.1" 200 -

To start the web interface use aikif/web_app/web_aikif.py or the batch file aikifgo_web_aikif

Simple Usage

In its simplest form AIKIF can be used to manage your projects and tasks, by updating information from scripts and tracking via the web application

my_biz = project.Project(name='Acute Software', type='business', desc='Custom Software')
my_biz.add_detail('website', 'http://www.acutesoftware.com.au')
my_biz.add_detail('email', 'djmurray@acutesoftware.com.au')

Logging data

You can use AIKIF as a database to manage adhoc data logging tasks

proj2 = project.Project(name='Sales Log', desc='Record list of sales')
proj2.add_detail('Note', 'List of sales taken from manual entries in test program')

tbl_exp = cls_datatable.DataTable('expenses.csv', col_names=['date', 'amount', 'details'])
proj2.record(tbl_exp, 'Expense', ['2015-02-13', 49.94, 'restaurant'])
proj2.record(tbl_exp, 'Expense', ['2015-02-15', 29.00, 'petrol'])
proj2.record(tbl_exp, 'Expense', ['2015-02-17', 89.95, 'fringe tickets'])

Data Collection Usage

p = aikif.project.Project('update Country reference', type='Auto')
p.add_task(1, 'download file', aikif.toolbox.web_download)
p.add_task(2, 'extract zip', aikif.toolbox.zip_util)  # not implemented
p.add_task(3, 'overwrite TXT to database staging', aikif.toolbox.data_load)

p.add_param(task=1, url='http://www.')
p.add_param(task=1, dest_zip = 'T:datadownloadcountry')
p.add_param(task=3, tbl='S_REF_COUNTRY')
p.execute()

This will execute the methods for each task using the specified parameters to update the table from the web

Map information

Define how columns in raw data should be mapped

m = aikif.mapper('custom mapper for countries', tbl = 'S_REF_COUNTRY')
m.add_col('code', data_type='STR', map_to_col='COUNTRY_CODE')
m.add_col('Name', data_type='STR', map_to_col='COUNTRY_NAME')
m.add_col('Continent', data_type='STR', map_to_col='CONTINENT')
m.add_col('Population', data_type='NUMBER', map_to_col='POPULATION')

Define your own Toolbox methods

Say you have a program 'my_average.py' which calculates averages that you want to include in the toolbox methods

t = aikif.toolbox.Toolbox()
t.add_tool(1, 'Calc Average', src=T:devsrcpythonmy_toolsmy_average.py')

p2 = aikif.project.Project('Aggregate Country by Continent')
p2.add_task(1, 'Fetch source data', aikif.toolbox.data_view)
p2.add_task(2, 'Aggregate Population', t['Calc Average'])

p2.add_param(task=1, tbl = 'S_REF_COUNTRY' )
p2.add_param(task=2, group_by_col = 'CONTINENT', measure_col='POPULATION' )

p2.execute()  # with no parameters, data outputs to console

More Information

Requirements Documentation
Design Notes
Overview Diagram