index.erl

-module(index).
-export([get_file_contents/1,show_file_contents/1
        ,split/1
]).

% Used to read a file into a list of lines.
% Example files available in:
%   gettysburg-address.txt (short)
%   dickens-christmas.txt  (long)
  

% Get the contents of a text file into a list of lines.
% Each line has its trailing newline removed.

get_file_contents(Name) ->
    {ok,File} = file:open(Name,[read]),
    Rev = get_all_lines(File,[]),
lists:reverse(Rev).

% Auxiliary function for get_file_contents.
% Not exported.

get_all_lines(File,Partial) ->
    case io:get_line(File,"") of
        eof -> file:close(File),
               Partial;
        Line -> {Strip,_} = lists:split(length(Line)-1,Line),
                get_all_lines(File,[Strip|Partial])
    end.

% Show the contents of a list of strings.
% Can be used to check the results of calling get_file_contents.

show_file_contents([L|Ls]) ->
    io:format("~s~n",[L]),
    show_file_contents(Ls);
 show_file_contents([]) ->
    ok.    
     
split(Name) ->
    Ls = get_file_contents(Name),
    split_lines(Ls).

split_lines(Ls) ->
    split_lines(1,noemptylines(nopunct(Ls))).

split_lines(_,[]) ->
    [];
split_lines(N,[L|Ls]) -> 
    [{N,string:tokens(L," ")}|split_lines(N+1,Ls)].

noemptylines([]) ->
    [];
noemptylines([L|Ls]) ->
        case L of
        [] -> noemptylines(Ls);
        _ -> [L|noemptylines(Ls)]
    end.

nopunct([]) ->
    [];
nopunct([X|Xs]) ->
    case lists:member(X,".,;:\t\n\'\"") of
	true ->
	    nopunct(Xs);
	false ->
	    [ X | nopunct(Xs) ]
    end.