Assignment 1 Simple Knowledge-Base

COMP 6591: Introduction to Knowledge-Base Systems
Assignment 1
Contents
1 General Information 2
2 Introduction 2
3 Ground rules 2
4 Overview 2
4.1 Dataset Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
5 Your Assignment 3
5.1 Knowledge Representation in Relational Model using SQLite . . . . . . . . . 4
5.2 Knowledge Representation using Prolog . . . . . . . . . . . . . . . . . . . . . 4
5.3 Knowledge-Base Queries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
6 What to Submit 5
7 Grading Scheme 7
1
1 General Information
Date posted: Monday, June 27th, 2022.
Date due: Friday, July 8th, 2022, by 23:59.1
.
Weight: 6% of the overall grade.
2 Introduction
This assignment targets basic Knowledge Representation and Queries using Prolog. In this
assignment you implement a small database system to represent movie information in both
relational as well as Prolog database format.
3 Ground rules
You work on a team of maximum 2 students(including yourself). Each team should designate
a leader who will submit the assignment electronically. See Submission Notes for the details.
ONLY one copy of the assignment is to be submitted by the team leader. Upon submission,
you must book an appointment with the marker team and demo the assignment. All members
of the team must be present during the demo to receive the credit. Failure to do so may
result in zero credit.
This is an assessment exercise. You may not seek any assistance from others while expecting
to receive credit (You must work strictly within your team). Failure to do so will
result in penalties or no credit.
4 Overview
This assignment you create a simple Knowledge-Base in SQLite and Prolog. The knowledgebase consists of some movie information. The following information is suggested to be
included in the assignment.
1
see submission notes
2
ˆ Movie titles: including the IMDB id, name, description / plot, etc.
ˆ Directors and Writers
ˆ Cast (names and possibly roles)
ˆ Keywords (a few common keywords per movie)
ˆ Genres
ˆ Languages
ˆ image urls
4.1 Dataset Requirements
While the format and size of the data is open, the following non-functional requirements
must be addressed. You may use IMDB datasets (1), any public services, or any other movie
databases, to gather the above data.
1. The data-set must include at least 10 titles.
2. The data-set must include at least two movies with more than one language.
3. At least 3 cast members per movie must be included.
4. There must exist a movie with no director or writer info.
5. Include at least 5 keywords per movie. Make sure some movies share one or two
keywords.
6. Make sure at least two genres are available.
7. Not all movies have images.
Do not include any TV-series.
5 Your Assignment
Your assignment consist of three parts: (i) Knowledge Representation in Relational Model
using SQLite, (ii) Knowledge Representation using Prolog, (iii) Knowledge-Base Queries
It is strongly recommended that you use command line interface (CLI) for both parts.
3
5.1 Knowledge Representation in Relational Model using SQLite
SQLite (3) is a light-weight relational database engine that enables implementing a small
local database in a single le. The rst step of the assignment is to create a small database
in SQLite that stores the above mentioned movie info.
5.1.1 Database Scheme and DDL
Provide the database schema. Include your DDL queries.
5.1.2 Database Queries
Using the database you created above, write the queries corresponding to the 7 data-set
requirements in 4.1 to verify that they are met.
5.1.3 CLI Scripts to export the data into CSV
Write two sample shell scripts to export two of the tables into CSV format 2
.
5.2 Knowledge Representation using Prolog
In this section, we want to represent the movie information in Prolog. To do so, you may
use the technique you in 5.1.3 or you may create the whole knowledge-base from scratch in
Prolog.
5.2.1 Database Facts
Represent the movie information as facts in Prolog.
While many of the facts may be presented using simple terms, make sure your database contains some instances of compound terms and complex structures, including Prolog lists. Here
are some suggestions: using lists for movie languages (i.e. title(..., [en fr], ...).,
2To achieve this, you may use SQLite CLI and IO redirection. An example of a sample script to provide
a text output in bash is given in the following:
echo -e "select * from titles;\n" | sqlite3 "$dbfile" 2>&1
4
using compound terms for casts (to include name and role), etc. You may however choose
and apply the complex structures on any pieces of information, as you feel is applicable.
5.2.2 Queries and Rules
This assignment does not specify any specify Prolog rule to be implemented. Instead, you
decide what rules are necessary. See section 5.3 for more details.
5.3 Knowledge-Base Queries
This section denes the queries to be implemented in both relational and deductive models. For each of the following, write the query/ies in prolog, followed by the corresponding
sql in relational model.
1. Search for a movie title.
Prolog: see atom_concat for matching substrings in the title.
2. list all movie titles.
Prolog: return the movies individually, and all in one list.
3. list all movies played by an actor.
4. list all distinct movies played by two given actors.
5. list all distinct movies played by either of two given actors.
6. list all distinct movies played by some actor within a list of years.
7. list movie ids, and titles sorted by id.
8. list movie ids, and titles sorted by title.
6 What to Submit
The whole assignment is submitted by the due date under the corresponding assignment
box. It has to be completed by ALL members of the team in one submission le.
Submission Notes
Clearly include the names and student IDs of all members of the team in the submission.
Indicate the team leader.
5
IMPORTANT: You are allowed to work on a team of 2 students at most (including yourself).
Any teams of 3 or more students will result in 0 marks for all team members. If your work on
a team, ONLY one copy of the assignment is to be submitted. You must make sure that you
upload the assignment to the correct assignment box on Moodle. No email submissions are
accepted. Assignments uploaded to the wrong system, wrong folder, or submitted via email
will be discarded and no resubmission will be allowed. Make sure you can access Moodle
prior to the submission deadline. The deadline will not be extended.
Naming convention for the uploaded le: Create one zip le, containing all needed les for
your assignment using the following naming convention. The zip le should be called
a#_studids, where # is the number of the assignment, and studids is the list of student ids of all team members, separated by (_). For example, for the rst assignment,
student 12345678 would submit a zip le named a1_12345678.zip. If you work on a
team of two and your IDs are 12345678 and 34567890, you would submit a zip le named
a1_12345678_34567890.zip.
Submit your assignment electronically on Moodle based on the instruction given by your
instructor as indicated above: https://moodle.concordia.ca
Please see course outline for submission rules and format, as well as for the required demo
of the assignment. A working copy of the code and a sample output should be submitted
for the tasks that require them. A text le with answers to the dierent tasks should be
provided. Put it all in a le layout as explained below, archive it with any archiving and
compressing utility, such as WinZip, WinRAR, tar, gzip, bzip2, or others. You must keep
a record of your submission conrmation. This is your proof of submission, which you may
need should a submission problem arises.
6
7 Grading Scheme
Relational Database 15 pts
Verication Queries 5 pts
CSV Export 5 pts
KB is Prolog 15 pts
Prolog Queries 30 pts
SQL-Queries 10 pts
Use of Complex Structures 20 pts
Total: 100 pts.
References
1. IMDB Datasets: https://www.imdb.com/interfaces/
2. SWI-Prolog: https://www.swi-prolog.org/Download.html
3. SQLite: https://www.sqlite.org/index.html
4. SQLite DDL: https://www.sqlitetutorial.net/sqlite-create-table/
5. SQLite CLI: https://www.sqlite.org/cli.html
6. SQLite to CSV: https://www.sqlitetutorial.net/sqlite-export-csv/
7. proSQLite: Prolog File Based Databases via an SQLite Interface:
https://link.springer.com/content/pdf/10.1007/978-3-642-45284-0.pdf
8. proSQLite on SWI-Prolog:
https://www.swi-prolog.org/pack/list?p=prosqlite
9. Sort by key in Prolog:
https://www.swi-prolog.org/pldoc/man?predicate=keysort/2
7