Read a file line by line in Prolog
You can use read
to read the stream. Remember to invoke at_end_of_stream
to ensure no syntax errors.
Example:
readFile.pl
main :-
open('myFile.txt', read, Str),
read_file(Str,Lines),
close(Str),
write(Lines), nl.
read_file(Stream,[]) :-
at_end_of_stream(Stream).
read_file(Stream,[X|L]) :-
\+ at_end_of_stream(Stream),
read(Stream,X),
read_file(Stream,L).
myFile.txt
'line 0'.
'line 1'.
'line 2'.
'line 3'.
'line 4'.
'line 5'.
'line 6'.
'line 7'.
'line 8'.
'line 9'.
Thus by invoking main
you will recieve the output:
?- main.
[line 0,line 1,line 2,line 3,line 4,line 5,line 6,line 7,line 8,line 9]
true
Just configure main
. The output here is an example by using write
, of course. Configure to match your request.
I assume that this principle can be applied to answer your question. Good luck.
In SWI-Prolog, the cleanest solution is to write a DCG that describes what a "line" is, then call a predicate for each line. Use library(pio) to apply the DCG to a file.
EDIT: As requested, consider:
:- use_module(library(pio)).
lines([]) --> call(eos), !.
lines([Line|Lines]) --> line(Line), lines(Lines).
eos([], []).
line([]) --> ( "\n" ; call(eos) ), !.
line([L|Ls]) --> [L], line(Ls).
Sample usage: ?- phrase_from_file(lines(Ls), 'your_file.txt').
There are kind of more possible in number and more reasonable in performance solutions, to get uninterpreted i.e plain text lines from a file:
SWI-Prolog:
read_line(S, X) :-
read_line_to_codes(S, L),
read_line2(L, X).
read_line2(end_of_file, _) :- !, fail.
read_line2(L, X) :-
atom_codes(X, L).
Jekejeke Prolog:
:- use_module(library(stream/console)).
Here are some timings, reading a file of 655 lines:
test :-
open('<path>', read, Stream),
test(Stream),
close(Stream).
test(Stream) :-
read_line(Stream, _), !,
test(Stream).
test(_).
SWI-Prolog:
̀?- time((between(1,100,_), test, fail; true)).
% 328,300 inferences, 0.125 CPU in 0.143 seconds (88% CPU, 2626400 Lips)
true.
Jekejeke Prolog:
?- time((between(1,100,_), test, fail; true)).
% Up 121 ms, GC 2 ms, Thread Cpu 94 ms (Current 05/07/19 17:19:05)
Yes
I guess a SWI-Prolog solution that reads into a string instead into an atom could be faster. But in the above we compare atom against atom reading.