nullobsi
/
cs420
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Add Assignment 4, Prolog stuff

This commit is contained in:
Citlali del Rey 2024-04-30 13:58:22 -07:00
parent 1bf67d7ee3
commit de50816932
Signed by: nullobsi
GPG Key ID: 933A1F44222C2634
5 changed files with 1464 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

103
Assignment4/Assign4.pl Normal file
View File

@ -0,0 +1,103 @@
:- style_check(-singleton).
% DO NOT MODIFY
:- consult('ReadCSV.pl').
:- use_module(library(date)).
:- read_csv_and_store('EmployeeData.csv').
% DO NOT MODIFY
is_seattle_employee(Name) :- person(_, Name, _, _, _, _, _, _, _, _, _, _, 'Seattle',_).
is_senior_manager_in_IT(Name) :- person(_, Name, 'Sr. Manger', 'IT', _, _, _, _, _, _, _, _, _, _).
is_director_finance_miami(Name) :- person(_, Name, 'Director', 'Finance', _, _, _, _, _, _, _, _, 'Miami', _).
is_asian_US_manufacturing_40M(Name, Unit, Gender, Ethnicity, Age) :- person(_, Name, _, _, Unit, Gender, Ethnicity, Age, _, _, _, _, _, _), Age > 40, Unit = 'Manufacturing', Gender = 'Male', Ethnicity = 'Asian'.
% format calls write internally
greet(EEID) :- person(EEID, Name, Job, Dept, Unit, _, _, _, _, _, _, _, _, _), format('Hello, ~w, ~w, ~w, ~w!~n', [Name, Job, Dept, Unit]).
years_until_retirement(Name, Age, Until) :- person(_, Name, _, _, _, _, _, Age, _, _, _, _, _, _), Until is 65-Age.
is_rd_black_midAge(Name, Unit, Ethnicity, Age) :- person(_, Name, _, _, Unit, _, Ethnicity, Age, _, _, _, _, _, _), Ethnicity = 'Black', Unit = 'Research & Development', Age >= 25, Age =< 50.
is_ITorFin_PHXorMIAorAUS(Name, Dept, City) :- person(_, Name, _, Dept, _, _, _, _, _, _, _, _, City, _),
(Dept = 'IT'; Dept = 'Finance'),
(City = 'Phoenix'; City = 'Miami'; City = 'Austin').
is_female_senior_role(Name, Title) :- person(_, Name, Title,_,_,'Female',_,_,_,_,_,_,_,_), atom_concat('Sr.', _, Title).
% Helper predicate that converts a salary string to a number
% Remove $.
convert_salary(In, Out) :- atom_concat('$', NoDollar, In),
% Split by commas, and additionally remove trailing whitespace
split_string(NoDollar, ',', '\t\s', NoCommaList),
% Concatenate the list back into an atom.
atomic_list_concat(NoCommaList, SalaryAtom),
% Convert the atom into a number.
atom_number(SalaryAtom, Out).
is_highly_paid_senior_manager(Name, Salary) :- person(_, Name, 'Sr. Manger', _, _, _, _, _, _, Salary, _, _, _, _),
convert_salary(Salary, SalaryN),
SalaryN > 120000.
% Divisibility test
divides(N,X) :- 0 is X mod N.
% prime/2 is helper predicate.
% it is a recursive algorithm that tests all the numbers from N up to x,
% ensuring they all do not divide N.
% Base case.
prime(N, N) :- true, !.
% X should not divide N, and the next number also should not.
prime(X, N) :- not(divides(X, N)), Next is X+1, prime(Next, N).
% prime/1 calls prime/2.
prime(N) :- prime(2,N).
is_prime_age(Name, Age) :- person(_, Name, _, _, _, _, _, Age, _, _, _, _, _, _), prime(Age).
% Uses convert_salary.
total_salary(Name, Salary) :- person(_, Name, _, _, _, _, _, _, _, AnnSal, BonusP, _, _, _),
convert_salary(AnnSal, AnnSalN), atom_concat(Bonus, '%', BonusP), atom_number(Bonus, BonusN),
Salary is AnnSalN + (AnnSalN * (BonusN / 100)).
takehome_salary(Name, Title, Salary) :- person(_, Name, Title,_,_,_,_,_,_,_,_,_,_,_),
% Get the total salary.
total_salary(Name, AS),
% Semicolon causes backtracking.
(
% "One of these clauses should be true"
AS =< 50000, Tax = 20/100;
AS > 50000, AS =< 100000, Tax = 25/100;
AS > 100000, AS =< 200000, Tax = 30/100;
AS > 200000, Tax = 35/100
% Unified value for Tax is then used here.
), Salary is AS - (AS * Tax).
total_years(Name, Years) :- person(_,Name,_,_,_,_,_,_,Hire,_,_,_,_,Exit),
% parse timestamp to get the year number
split_string(Hire, '/', '', [_,_,HireYrS]),
atom_number(HireYrS, HireYrShort),
% since it's a short number, add 1900 if > 50
% but otherwise add 2000
(HireYrShort < 50, HireYr is HireYrShort + 2000;
HireYr is HireYrShort + 1900),
% find "exit" year
(Exit = '', get_time(T), stamp_date_time(T, DT, local), date_time_value(year, DT, CurYear), Years is CurYear - HireYr, !;
% parse exit year
split_string(Exit, '/', '', [_,_,ExitYrS]),
atom_number(ExitYrS, ExitYrShort),
(ExitYrShort < 50, ExitYr is ExitYrShort + 2000;
ExitYr is ExitYrShort + 1900),
Years is ExitYr - HireYr, !).
title_salary(Title, Salary) :- person(_,_,Title,_,_,_,_,_,_,SalaryS,_,_,_,_), convert_salary(SalaryS, Salary).
% bagof() gets al the possible salaries for a title. Then, i use foldl
% to sum up all of the salaries, and divide by the number of salaries,
% which is the formula for the average.
average_salary(Title, AvgSalary) :- bagof(Salary, title_salary(Title, Salary), SalList), foldl(plus, SalList, 0, Total), length(SalList, N), AvgSalary is Total / N.

1001
Assignment4/EmployeeData.csv Normal file
View File

File diff suppressed because it is too large Load Diff

17
Assignment4/ReadCSV.pl Normal file
View File

@ -0,0 +1,17 @@
:- use_module(library(csv)).
% Predicate to read data from a CSV file and store it as rules
read_csv_and_store(Filename) :-
csv_read_file(Filename, [_|Rows], []),
process_rows(Rows).
% Process each row in the CSV file and store data as rules
process_rows([]).
process_rows([Row|Rows]) :-
process_row(Row),
process_rows(Rows).
% Store data from a row as a rule
process_row(row(EEID, Name, Job, Department, Unit, Gender, Ethnicity, Age, Hired, Salary, Bonus, Country, City, Exited)) :-
assert(person(EEID, Name, Job, Department, Unit, Gender, Ethnicity, Age, Hired, Salary, Bonus, Country, City, Exited)).

BIN
Assignment4/report4.pdf Normal file
View File

Binary file not shown.

343
Assignment4/report4.tex Normal file
View File

@ -0,0 +1,343 @@
%! TeX program = lualatex
\RequirePackage[l2tabu,orthodox]{nag}
\DocumentMetadata{lang=en-US}
\documentclass[a4paper]{scrartcl}
\usepackage{geometry}
\usepackage{graphicx}
%\usepackage{tikz}
%\usepackage{tikz-uml}
\usepackage{hyperref}
\usepackage{caption}
\usepackage{subcaption}
\usepackage{newfloat}
\usepackage{fancyvrb}
\usepackage{minted}[newfloat=true]
\usepackage{bookmark}
\usepackage{fontspec}
\usepackage{microtype}
% Math packages
%\usepackage{amsmath}
%\usepackage{mathtools}
%\usepackage{amsthm}
%\usepackage{thmtools}
%\usepackage{lualatex-math}
%\usepackage[warnings-off={mathtools-colon,mathtools-overbracket},math-style=ISO,bold-style=ISO]{unicode-math}
% Fonts
\usepackage{newcomputermodern}
\setmonofont{0xProto}[Scale=MatchLowercase]
\newcommand*{\figref}[2][]{%
\hyperref[{fig:#2}]{%
Figure~\ref*{fig:#2}%
\ifx\\#1\\%
\else
\,#1%
\fi
}%
}
\newcommand*{\lstref}[2][]{%
\hyperref[{lst:#2}]{%
Listing~\ref*{lst:#2}%
\ifx\\#1\\%
\else
\,#1%
\fi
}%
}
\setminted{breaklines=true,frame=single,linenos=true}
\newenvironment{longlisting}{\captionsetup{type=listing}}{}
\newenvironment{longfigure}{\captionsetup{type=figure}}{}
\setlength{\belowcaptionskip}{10pt}
\setlength{\abovecaptionskip}{8pt}
%\DeclarePairedDelimiter\ceil{\lceil}{\rceil}
%\DeclarePairedDelimiter\floor{\lfloor}{\rfloor}
%\declaretheorem[within=chapter]{definition}
%\declaretheorem[sibling=definition]{theorem}
%\declaretheorem[sibling=definition]{corollary}
%\declaretheorem[sibling=definition]{principle}
\usepackage{polyglossia}
\usepackage[backend=biber]{biblatex}
\setdefaultlanguage[variant=american,ordinalmonthday=true]{english}
\day=3
\month=5
\year=2024
\title{Logical Programming in Prolog}
\subtitle{Assignment 4}
\author{Juan Pablo Zendejas}
\date{\today}
\begin{document}
\maketitle
%\listoftheorems[ignoreall,onlynamed={theorem,corollary,principle}]
%\listoftheorems[ignoreall,onlynamed={definition},title={List of Definitions}]
%\tableofcontents
In this assignment, I was tasked with creating Prolog predicates that
query the database of employees given in a CSV file. I created 15 rules
as specified in the assignment PDF. Overall, this assignment was very
interesting to me. Prolog is very different from any other programming
language I've used, even Haskell. Getting used to the eccentricities of
SWI-Prolog and reading its documentation gave me a lot of insight. In
addition, my background of formal logic from Discrete Mathematics taken
here at SDSU gave me a lot of parallels that helped me understand how to
write code.
I did not receive any external help except from consulting the SWI
Prolog handbook/documentation found on their website. This documentation
was very useful to figure out what predicates to use to transform some
of the data.
\section{CSV File Importing}
First I had to change the \texttt{ReadCSV.pl} file to import all the
rows from the CSV, ignoring the first title row. The source code is
displayed in \lstref{readcsv}.
\begin{listing}
\inputminted[label={ReadCSV.pl}]{prolog}{ReadCSV.pl}
\caption{CSV reader.}
\label{lst:readcsv}
\end{listing}
This part wasn't too bad. I simply had to ignore the first row, which I
did by using Pattern Matching on the \texttt{csv\_read\_file} predicate to
ignore the first entry of the list and simply take the tail as the
\texttt{Rows} variable.
Then, on \texttt{process\_row}, I simply filled in the \texttt{assert}
call with all the names of the columns.
\section{Writing Rules}
Now, I was tasked with using the imported facts from the CSV and writing
rules that would be used as queries to the dataset.
\subsection{Rule One}
Simple rule, displayed in \lstref{seattleempl}. The \texttt{:-} symbol
is reminiscent of ``yields'' from mathematics. I use the underscore to
discard the variables from the \texttt{people} predicate that I don't
need, and just check if a person exists with that name and is from
Seattle.
\begin{listing}
\inputminted[label={is\_seattle\_employee},firstline=10,lastline=10]{prolog}{Assign4.pl}
\caption{Rule to determine if an employee is from Seattle.}
\label{lst:seattleempl}
\end{listing}
\subsection{Rule Two}
Very similar to rule one, displayed in \lstref{seniormangerit}. Now, we
just have more filters to apply, which is done by replacing some of the
underscores with concrete values.
\begin{listing}
\inputminted[label={is\_senior\_manager\_in\_it},firstline=12,lastline=12]{prolog}{Assign4.pl}
\caption{Rule to determine if an employee is a senior manager in IT.}
\label{lst:seniormangerit}
\end{listing}
\subsection{Rule Three}
Displayed in \lstref{rule3}. Standard like the previous rules.
\begin{listing}
\inputminted[label={is\_director\_finance\_miami},firstline=14,lastline=14]{prolog}{Assign4.pl}
\caption{Rule to determine if an employee is a Director of Finance
in Miami.}
\label{lst:rule3}
\end{listing}
\subsection{Rule Four}
This rule is a little more complicated. It's displayed in
\lstref{rule4}. Since this predicate has more arguments, I have them
directly wired into the \texttt{person} predicate. Then, I simply have
conditions that specify the values of those predicates. In addition, I
had to do this for \texttt{Age} because I need to check if it's greater
than 40.
\begin{listing}
\inputminted[label={is\_asian\_US\_manufacturing\_40M},firstline=16,lastline=16]{prolog}{Assign4.pl}
\caption{Rule to determine if an employee is a 40 or older Asian
male working in manufacturing in the US.}
\label{lst:rule4}
\end{listing}
\subsection{Rule Five}
This rule is given an employee ID, and should print a greeting for that
employee. The source code is displayed in \lstref{rule5}. I looked in
the SWI-Prolog documentation to figure out how to format a string, and I
found the \texttt{format/2} predicate. It basically uses the ~ symbol to
define placeholders. Then, using the \texttt{person} predicate to get
the data from the employee ID and outputting it with \texttt{format/2}
was pretty easy.
\begin{listing}
\inputminted[label={greet},firstline=19,lastline=19]{prolog}{Assign4.pl}
\caption{Rule to greet an employee from their ID.}
\label{lst:rule5}
\end{listing}
\subsection{Rule Six}
This rule calculates an employee's number of years until retirement. The
retirement age is given as 65. This was fairly simple, I had to use the
\texttt{is} operator to calculate a value and unify it with the
predicate's argument. The source code is displayed in \lstref{rule6}.
\begin{listing}
\inputminted[label={years\_until\_retirement},firstline=21,lastline=21]{prolog}{Assign4.pl}
\caption{Rule to determine an employee's years until retirement.}
\label{lst:rule6}
\end{listing}
\subsection{Rule Seven}
This rule is similar to the previous ones, but now I define two
conditions on the age, using the less than and greater than operators.
Prolog is interesting in this regard because it has the order of the
symbols swapped from what it usually is. Something to remember for sure.
The source code is displayed in \lstref{rule7}.
\begin{listing}
\inputminted[label={is\_rd\_black\_midAge},firstline=23,lastline=23]{prolog}{Assign4.pl}
\caption{Rule to determine if an employee is middle-aged, black, and
in research and development.}
\label{lst:rule7}
\end{listing}
\subsection{Rule Eight}
This rule introduced me to the semicolon operator, and how to use it to
make more complicated rules. Basically, the semicolon operator tells
Prolog to backtrack and try other possibilities if one of the dependent
propositions is false. I can wrap stuff in parentheses and use the
semicolon to define ``or''-like statements. The source code is displayed in
\lstref{rule8}.
\begin{listing}
\inputminted[label={is\_ITorFin\_PHXorMIAorAUS},firstline=25,lastline=27]{prolog}{Assign4.pl}
\caption{Rule to determine if an employee is in IT or finance, and
in Phoenix or Miami or Austin.}
\label{lst:rule8}
\end{listing}
\subsection{Rule Nine}
This rule introduced me to the \texttt{atom\_concat} predicate. It was
really interesting to learn that in Prolog, predicates like these can
work in different ways based on which variable needs to be unified.
Thus, even though the predicate is called \texttt{concat}, it can also
split. That's what I did here to remove the `Sr.' from the employee's
title. The predicate is only true (can be unified) when the title starts
with Sr. The source code is displayed in \lstref{rule9}.
\begin{listing}
\inputminted[label={is\_female\_senior\_role},firstline=29,lastline=29]{prolog}{Assign4.pl}
\caption{Rule to determine if an employee is both female and in a
senior role.}
\label{lst:rule9}
\end{listing}
\subsection{Rule Ten}
For this rule, I had to create a helper predicate
\texttt{convert\_salary/2}, which is shown in \lstref{convertsalary}.
Now, I had to write predicates that reasoned with numbers instead of
just atom equality. The source code is shown in \lstref{rule10}. This
can be done with typical symbols like less than, greater than. I also
have to write code to convert the string representation of the salary
into a number; this was done by first removing the dollar sign, then
removing the comma and parsing it into a number. The SWI Prolog
documentation helped here.
\begin{listing}
\inputminted[label={convert\_salary},firstline=33,lastline=39]{prolog}{Assign4.pl}
\caption{Rule to convert a salary string to a number.}
\label{lst:convertsalary}
\end{listing}
\begin{listing}
\inputminted[label={is\_highly\_paid\_senior\_manager},firstline=41,lastline=43]{prolog}{Assign4.pl}
\caption{Rule to determine if an employee is a highly paid senior
manager.}
\label{lst:rule10}
\end{listing}
\subsection{Rule Eleven}
This rule determines if an employee has an age that's prime. This one
was difficult. I thought I could use an algorithm similar to Haskell,
but that was a generator. I ended up using a naive recursive algorithm
that checks if all the numbers less than the age do NOT divide the age.
Probably inefficient, but it works. The source code is shown in
\lstref{rule11}.
\begin{listing}
\inputminted[label={is\_prime\_age},firstline=45,lastline=59]{prolog}{Assign4.pl}
\caption{Rule to determine if an employee has a prime age.}
\label{lst:rule11}
\end{listing}
\subsection{Rule Twelve}
This rule calculates an employee's salary after bonuses, but before
taxes. Now that I wrote \texttt{convert\_salary/2}, this predicate has a
fairly simple and easy to read implementation. It is displayed in
\lstref{rule12}.
\begin{listing}
\inputminted[label={total\_salary},firstline=62,lastline=64]{prolog}{Assign4.pl}
\caption{Rule to determine an employee's total salary.}
\label{lst:rule12}
\end{listing}
\subsection{Rule Thirteen}
Now, this rule calculates the take-home salary of an employee. This is
done by using a tax bracketed system based on the employee's total
salary. Again, I had to use the semicolon to provide backtracking and
alternatives for Prolog, which worked as a conditional to set the tax
level. The code is displayed in \lstref{rule13}.
\begin{listing}
\inputminted[label={takehome\_salary},firstline=66,lastline=77]{prolog}{Assign4.pl}
\caption{Rule to determine an employee's take-home salary.}
\label{lst:rule13}
\end{listing}
\subsection{Rule Fourteen}
This was a pretty complicated predicate to write, but it was fun. The
source code is displayed in \lstref{rule14}. I used the SWI-Prolog
documentation to figure out how to use the date-time objects and
predicates. One of the difficult parts is the strange date format in the
CSV, which I had to parse using \texttt{split\_string/4} and some pattern
matching. Then, I had to convert the short year number into a year in
the 20th or 21st century.
\begin{listing}
\inputminted[label={total\_years},firstline=79,lastline=94]{prolog}{Assign4.pl}
\caption{Rule to determine an employee's years of service.}
\label{lst:rule14}
\end{listing}
\subsection{Rule Fifteen}
Finally, I had to write a rule to determine the average salary for an
entire job title. The source code is displayed in \lstref{rule15}.
Using the ways of Prolog backtracking, this actually
wasn't too bad. Essentially, I wrote a predicate that simply gets the
salary for a specific person with a title. There are many ways to unify
the Salary variable that make the predicate true. Then, I use the Prolog
\texttt{bagof/3} predicate to collect every possible value that
satisfies the predicate into a list. Then, getting the average is easy.
\begin{listing}
\inputminted[label={average\_salary},firstline=97,lastline=102]{prolog}{Assign4.pl}
\caption{Rule to determine the average salary for a job title.}
\label{lst:rule15}
\end{listing}
\end{document}