%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% wordcloud %% %% drawing wordclouds %% %% with METAPOST and Lua %% %% chupin@ceremade.dauphine.fr %% %% Version 0.1 (aout 2023) %% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% % % This work may be distributed and/or modified under the conditions of % the LaTeX Project Public License, either version 1.3c of this license % or (at your option) any later version. The latest version of this % license is in http://www.latex-project.org/lppl.txt % %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% if not known mplib: input latexmp ; show"wordcloud Infos: latexmp loaded";fi; input colorbrewer-rgb % function to get all the path of letters (too slow) vardef build_pathpicture(expr pict)= save tex_pct,glp_pct,glp_idx,glp_num,pth_num,glp_pth,fnt_str, txt_str, sub_str,txt_wd; % text material and glyphs contours picture tex_pct, glp_pct; numeric glp_num, pth_num[]; path glp_pth[][]; % tex_pct btex .. etex material % glp_num number of glyphs within tex_pct % pth_num[i] number of paths defining the rank i (0 .. glp_num-1) glyph % glp_pth[i][j] rank j (0 .. pth_num[i]-1) path of the rank i glyph % glp_pct glyphs picture tex_pct:=pict; glp_pct:=nullpicture; string fnt_str, txt_str, sub_str; numeric txt_wd; glp_num:=0; for tkn within pict: %show tkn; if textual tkn: %show "textual"; fnt_str:=fontpart tkn; txt_str:=textpart tkn; txt_wd:=0; for glp_idx=0 upto (length txt_str-1): sub_str:=substring (glp_idx, glp_idx+1) of txt_str; for sub_tkn within glyph ASCII sub_str of fnt_str scaled (fontsize fnt_str/1000) xscaled xxpart tkn yscaled yypart tkn shifted (txt_wd+xpart tkn, ypart tkn): addto glp_pct doublepath (pathpart sub_tkn); endfor txt_wd:=txt_wd+ (xxpart tkn)*xpart urcorner (sub_str infont fnt_str); endfor fi endfor glp_pct enddef; % path intersection (trop gourmant en temps, too slow...) vardef word_intersection(expr tic, tac)= save _p, _w, _v,_output; boolean _output; _output:=false; path _p,_w,_v; for _p within tic: _w := pathpart _p; for _q within tac: _v := pathpart _q; if((_w intersectiontimes _v)<>(-1,-1)): % intersection _output:=true; fi endfor endfor _output enddef; % global variable _wordcloud_box_margin:=0.3pt; % setting the box margin def set_box_margin(expr s)= _wordcloud_box_margin:=s; enddef; % extrenal box intersection (horizontal or vertical) vardef extbb_intersection(expr tic, tac)= save xmax_, xmin_, ymax_, ymin_, xmax__, xmin__, ymax__, ymin__, _output; boolean _output; _output:=false; %show _p; xmin_:= (xpart llcorner tic) -_wordcloud_box_margin; ymin_ := (ypart llcorner tic)-_wordcloud_box_margin; xmax_:= (xpart urcorner tic)+_wordcloud_box_margin; ymax_ := (ypart urcorner tic)+_wordcloud_box_margin; xmin__:= (xpart llcorner tac)-_wordcloud_box_margin; ymin__ := (ypart llcorner tac)-_wordcloud_box_margin; xmax__:= (xpart urcorner tac)+_wordcloud_box_margin; ymax__ := (ypart urcorner tac)+_wordcloud_box_margin; if ((xmin_<= xmax__) and (xmax_ >= xmin__) and (ymin_<= ymax__) and (ymax_ >= ymin__)): _output:=true; fi _output enddef; % subbb_entersection of words (from textext) %%%!!!!!! not working !!!!!!!!! vardef bb_intersection(expr tic, tac)= save xmax_, xmin_, ymax_, ymin_, xmax__, xmin__, ymax__, ymin__, _output; save _p, _w, _v,_output,_q; boolean _output; _output:=false; path _p,_w,_v; i:=0; for _v within tic: for _p within _v: i:=i+1; %show "i"&decimal(i); xmin_:= (xpart llcorner _p)-_wordcloud_box_margin; ymin_ := (ypart llcorner _p)-_wordcloud_box_margin; xmax_:= (xpart urcorner _p)+_wordcloud_box_margin; ymax_ := (ypart urcorner _p)+_wordcloud_box_margin; j:=0; for _w within tac: for _q within _w: j:=j+1; %show "j"&decimal(j); xmin__:= (xpart llcorner _q)-_wordcloud_box_margin; ymin__ := (ypart llcorner _q)-_wordcloud_box_margin; xmax__:= (xpart urcorner _q)+_wordcloud_box_margin; ymax__ := (ypart urcorner _q)+_wordcloud_box_margin; if ((xmin_<= xmax__) and (xmax_ >= xmin__) and (ymin_<= ymax__) and (ymax_ >= ymin__)): _output:=true; fi exitif(_output); endfor endfor exitif(_output); endfor endfor _output enddef; % build the spiral to put words path _wordcloud_spiral; _wordcloud_spiral:=(0,0); _wordcloud_N:= 10*360; _tours:=20; _scale:=0.3; for j:=1 upto _tours: _rapj:=j/_tours; for i:=1 step 10 until 360: _rapi:=((i)+(j)*360)/360; %show (_rapj*_rapi); %show (6*_rapj*_rapi)*(cosd(i)*1cm,sind(i)*1cm); _wordcloud_spiral:=_wordcloud_spiral--((_scale*_rapi)*(cosd(i)*1cm,sind(i)*1cm)); endfor; endfor picture _wordcloud_words[], _wordcloud_pathwords[]; path _wordcloud_words_oriented_bb[]; numeric _wordcloud_weights[]; %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%% FONCTIONS USUELLES %%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% % from repere.mp Olivier Péault pi:=3.141592654; e:=2.718281828; vardef sin(expr x) = sind(180*x/pi) enddef; vardef cos(expr x) = cosd(180*x/pi) enddef; vardef tan(expr x) = sin(x)/cos(x) enddef; vardef exp(expr x) = mexp(x*256) enddef; vardef ln(expr x) = mlog(x)/256 enddef; vardef ch(expr x)=(exp(x)+exp(-x))/2 enddef; vardef sh(expr x)=(exp(x)-exp(-x))/2 enddef; vardef Arctan(expr x)= pi*angle(1,x)/180 enddef; vardef Arctand(expr x)= angle(1,x) enddef; vardef Arcsin(expr x)= Arctan(x/sqrt(1-x**2)) enddef; vardef Arcsind(expr x)= Arctand(x/sqrt(1-x**2)) enddef; vardef Arccos(expr x)= pi/2-Arcsin(x) enddef; vardef Arccosd(expr x)= Arctand((sqrt(1-x**2))/x)enddef; % general scaling _wordcloud_gen_scale:=1; def set_wordcloud_scale(expr s)= _wordcloud_gen_scale:=s; enddef; % fonction d'agrandissement en fonction du poids du mot vardef _wordcloud_scale_fct(expr e,_max,_min)= save _t,_N,_a,_b; _a*_min+_b=0.7; _a*_max+_b=2.5; _t:=_a*e+_b; _t:=ln(_t)-ln(0.7)+0.295; _wordcloud_gen_scale*(1.1*_t) enddef; % fabrication des mots avec la taille requise vardef _build_tabs(suffix _words, _weights)(expr _N)= save _tmp_pict,_width,_height; picture _tmp_pict; for i=1 upto _N: _tmp_pict:=textext(_words[i]) scaled (_wordcloud_scale_fct(_weights[i],_weights[1],_weights[_N])); _width:=abs(ulcorner _tmp_pict - urcorner _tmp_pict); _height:=abs(ulcorner _tmp_pict - llcorner _tmp_pict); _wordcloud_words[i]:= _tmp_pict shifted (-0.5_width,-0.5_height); _wordcloud_weights[i]:=_weights[i]; _bbox_old:=bboxmargin; bboxmargin:=_wordcloud_box_margin; _wordcloud_words_oriented_bb[i]:= bbox _wordcloud_words[i]; bboxmargin:=_bbox_old; _wordcloud_pathwords[i]:=build_pathpicture(_wordcloud_words[i]) ; endfor; enddef; % function to compute angle def _wordcloud_rotation(expr a,b)= % a: angle % b: 0 or 1 if b=0: a else: if(a=0): 90 else: -a fi fi enddef; %%% colors boolean _wordcloud_use_color; _wordcloud_use_color:=false; numeric wordcloud_colors_number; color wordcloud_colors[]; % default colors wordcloud_colors[1]:=Reds[3][3]; wordcloud_colors[2]:=Greens[3][3]; wordcloud_colors[3]:=Blues[3][3]; wordcloud_colors[4]:=Oranges[3][3]; wordcloud_colors[5]:=black; wordcloud_colors_number:=5; def wordcloud_use_color(expr b)= _wordcloud_use_color:= b; enddef; % main function to draw the picture vardef draw_wordcloud(suffix _words, _weights, rotation)(expr _N)= save _rotation_bool,_position,_ok_draw,_tmp_word,_random_pt,_position_pt,_pt; pair _random_pt,_position_pt,_pt; path _tmp_word; boolean _ok_draw; _position:=0; _build_tabs(_words,_weights,_N); for i=1 upto _N: %show i; _position:=0; if(i mod 2) = 0: _rotation_bool:=1; else: _rotation_bool:=0; fi if(i=1): draw _wordcloud_words[i] rotated (_wordcloud_rotation(rotation,_rotation_bool)) shifted (point _position of _wordcloud_spiral) if(_wordcloud_use_color): withcolor wordcloud_colors[i] fi ; _position_pt:=(0,0); _random_pt:=(0,0); else: _count:=0; forever: _position_pt:=point _position of _wordcloud_spiral; _random_pt:=(uniformdeviate(1)*0.01cm,uniformdeviate(1)*0.01cm); _pt:=_position_pt+_random_pt; _count:=_count+1; _ok_draw:= true; _tmp_word:=_wordcloud_words_oriented_bb[i] rotated (_wordcloud_rotation(rotation,_rotation_bool)) shifted (_position_pt + _random_pt); %_tmp_word:=_wordcloud_words[i] rotated (_rotation*90) shifted (point %_position % of _wordcloud_spiral); for j=1 upto i-1: if(rectangles_intersection(_tmp_word,_wordcloud_words_oriented_bb[j])): %if(bb_intersection(_tmp_word,_wordcloud_words[j])): _ok_draw:= false; %fi fi endfor; if(_ok_draw): % if there is no intersections draw _wordcloud_words[i] rotated (_wordcloud_rotation(rotation,_rotation_bool)) shifted (_position_pt + _random_pt) if(_wordcloud_use_color): withcolor wordcloud_colors[((i-1) mod wordcloud_colors_number)+1] fi; else: % else, we move along the spiral _position:=_position+1; fi exitif(_ok_draw); %exitif(_count>30); endfor; fi _wordcloud_words[i]:=_wordcloud_words[i] rotated (_wordcloud_rotation(rotation,_rotation_bool)) shifted (_position_pt + _random_pt); _wordcloud_words_oriented_bb[i]:=_wordcloud_words_oriented_bb[i] rotated (_wordcloud_rotation(rotation,_rotation_bool)) shifted (_position_pt + _random_pt); _wordcloud_pathwords[i]:=_wordcloud_pathwords[i] rotated (_wordcloud_rotation(rotation,_rotation_bool)) shifted (_position_pt + _random_pt); endfor; enddef; vardef dotproduct(expr a,b)= save _output; _output := (xpart a)*(xpart b)+(ypart a)*(ypart b); _output enddef; vardef normvec(expr p)= p/(sqrt((xpart p)**2+(ypart p)**2)) enddef; vardef bb_projection(expr p,a)= % p : edge of a rectangle (pair) % a : axis of proj dotproduct(p,a) enddef; % intersection of oriented rectangel vardef rectangles_intersection(expr tic,tac)= % tic, tac, two rectangular path (closed) save x,y,a,b,normedge,edge,_min,_max,_dotprod,_output,_minEdge,_maxEdge; pair normedge,edge; boolean _output; _output:=false; numeric x[],y[],a[],b[]; (x1,y1) = point 1 of tic; (x2,y2) = point 2 of tic; (x3,y3) = point 3 of tic; (x4,y4) = point 4 of tic; (a1,b1) = point 1 of tac; (a2,b2) = point 2 of tac; (a3,b3) = point 3 of tac; (a4,b4) = point 4 of tac; for i=1 upto 4: %show (x[i],y[i]); endfor; for i=1 upto 4: %show (a[i],b[i]); endfor; %draw (a1,b1)--(a2,b2)--(a3,b3)--(a4,b4); for i=1 upto 2: % on prend deux côtés perpendiculaire du rectangle 1 %show "premier"; %show i; edge:=(x[i+1],y[i+1])-(x[i],y[i]); %show edge; normedge:=normvec(edge); for j=1 upto 4: %pour chaque point de deuxième triangle on projette sur % l’axe du côté du rectangle 1 qu’on considère _dotprod:=dotproduct((x[j],y[j]),normedge); if(j=1): _minEdge:=_dotprod; _maxEdge:=_dotprod; fi if(_dotprod<_minEdge): _minEdge:=_dotprod; else: if(_dotprod>_maxEdge): _maxEdge:=_dotprod; fi fi %show _minEdge; %show _maxEdge; endfor; %show _minEdge; %show _maxEdge; for j=1 upto 4: %pour chaque point de deuxième triangle on projette sur % l’axe du côté du rectangle 1 qu’on considère _dotprod:=dotproduct((a[j],b[j]),normedge); if(j=1): _min:=_dotprod; _max:=_dotprod; fi if(_dotprod<_min): _min:=_dotprod; else: if(_dotprod>_max): _max:=_dotprod; fi fi endfor; %show _min; %show _max; if (_min>_maxEdge) or (_max<_minEdge): %show "OK"; _output:=true; fi endfor; %the other way for i=1 upto 2: % on prend deux côtés perpendiculaire du rectangle 1 %show "deuxième"; %show i; edge:=(a[i+1],b[i+1])-(a[i],b[i]); %show edge; normedge:=normvec(edge); for j=1 upto 4: %pour chaque point de deuxième triangle on projette sur % l’axe du côté du rectangle 1 qu’on considère _dotprod:=dotproduct((a[j],b[j]),normedge); if(j=1): _minEdge:=_dotprod; _maxEdge:=_dotprod; fi if(_dotprod<_minEdge): _minEdge:=_dotprod; else: if(_dotprod>_maxEdge): _maxEdge:=_dotprod; fi fi %show _minEdge; %show _maxEdge; endfor; %show _minEdge; %show _maxEdge; for j=1 upto 4: %pour chaque point de deuxième triangle on projette sur % l’axe du côté du rectangle 1 qu’on considère _dotprod:=dotproduct((x[j],y[j]),normedge); if(j=1): _min:=_dotprod; _max:=_dotprod; fi if(_dotprod<_min): _min:=_dotprod; else: if(_dotprod>_max): _max:=_dotprod; fi fi endfor; if (_min>_maxEdge) or (_max<_minEdge): %show "OK"; _output:=true; fi endfor; not _output enddef;