/* 

elecpred.c

C program for helping to predict Canadian election results

by Jeffrey Rosenthal (probability.ca), 2011

Compile with "cc -lm elecpred.c -o elecpred".

Requires five auxiliary files:  partynames, provincenames,
results2008.csv, groundvals, and shiftvals.

Note: results2008.csv is copied from:
http://www.elections.ca/scripts/OVR2008/31/data/table_tableau12.csv

After compiling, run the program with "elecpred a b c" where:
a=0 for model (1), or a=11 for model (2)
b=0 to not include the "ground" (over-performing) adjustments,
					or b=1 to include them
c=0 to not include the "shift" adjustments, or c=1 to include them

*/

#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <math.h>

#define verboselevel 0

#define MAXLINELENGTH 160
#define MAXNUMFIELDS 20
#define MAXFIELDLENGTH 160
#define EFILENAME "results2008.csv"

/* Global variables for storing vote results. */
#define MAXRIDINGNUM 99999
#define NUMREGIONS 5
#define NUMPARTIES 5
#define NUMPROVINCES 13
int ridingtot[MAXRIDINGNUM];
int ridingwinner[MAXRIDINGNUM];
int regiontot[NUMREGIONS];
int votecount[MAXRIDINGNUM][NUMPARTIES];
int regionvote[NUMREGIONS][NUMPARTIES];
double regionperc[NUMREGIONS][NUMPARTIES];
double totperc[NUMPARTIES];
int seattot[NUMPARTIES];
int votetot[NUMPARTIES];
int seatregion[NUMREGIONS][NUMPARTIES];
int ridingregion[MAXRIDINGNUM];
int tottot, numseats;
int regionseats[NUMREGIONS];
double adjustedcount[MAXRIDINGNUM][NUMPARTIES];
int adjustedwinner[MAXRIDINGNUM];
int adjustedseatregion[NUMREGIONS][NUMPARTIES];
int adjustedseattot[NUMPARTIES];
double groundval[NUMPARTIES];
double shiftval[NUMPARTIES];
int USERATIO, USEGROUND, USESHIFT;

/* Global variables for parsing input lines etc. */
char linedata[MAXNUMFIELDS][MAXFIELDLENGTH];
char theline[MAXLINELENGTH];
char prevline[MAXLINELENGTH];
char provincename[NUMPROVINCES][80];
char partyname[NUMPARTIES][80];
int totnumfields, filedone, maxnumcourses;
int regionmap[] = {0,0,0,0,1,2,3,3,3,3,3,3,4};
char partytag[][20] = {"Con", "Lib", "NDP", "Bloc", "Green"};
char regiontag[][20] = {"Atlantic", "Quebec", "Ontario", "Praries", "B.C."};

/* BEGIN MAIN PROGRAM. */

int main(int argc, char **argv)
{

    /* Initial declarations. */
    FILE *fp;
    char tmpstring[MAXLINELENGTH];
    int i, j, tmptot;
    int riding, party, votes, province, region, winner, ridingmax;
    double pollreg[NUMREGIONS][NUMPARTIES];
    double polltot[NUMPARTIES];

    /* Read in the province names. */
    if ((fp = fopen("provincenames","r")) == NULL) {
        printf("OOPS, unable to read file %s.\n", "provincenames");
        exit(1);
    }
    for (i=0; i<NUMPROVINCES; i++) {
      fgets(theline, MAXLINELENGTH,fp);
      theline[strlen(theline)-1] = '\0';
      strcpy(provincename[i], theline);
      /* printf("**%s**\n", theline); */
      printf("**%s**\n", provincename[i]);
    }
    fclose(fp);

    /* Read in the party names. */
    if ((fp = fopen("partynames","r")) == NULL) {
        printf("OOPS, unable to read file %s.\n", "partynames");
        exit(1);
    }
    for (i=0; i<NUMPARTIES; i++) {
      fgets(theline, MAXLINELENGTH,fp);
      theline[strlen(theline)-1] = '\0';
      strcpy(partyname[i], theline);
      /* printf("**%s**\n", partyname[i]); */
    }
    fclose(fp);
    /* printf("provincename[0]: **%s**\n", provincename[0]); */

    /* Read in the party "ground" values (could be zero). */
    if ((fp = fopen("groundvals","r")) == NULL) {
        printf("OOPS, unable to read file %s.\n", "groundvals");
        exit(1);
    }
    for (i=0; i<NUMPARTIES; i++) {
      fscanf(fp, "%s", tmpstring);
      groundval[i] = atof(tmpstring);
      printf("**%f**\n", groundval[i]);
    }
    fclose(fp);

    /* Read in the party "shift" values (could be zero). */
    if ((fp = fopen("shiftvals","r")) == NULL) {
        printf("OOPS, unable to read file %s.\n", "shiftvals");
        exit(1);
    }
    for (i=0; i<NUMPARTIES; i++) {
      fscanf(fp, "%s", tmpstring);
      shiftval[i] = atof(tmpstring);
      printf("**%f**\n", shiftval[i]);
    }
    fclose(fp);

    /* Check command options. */
    USERATIO = USEGROUND = USESHIFT = 0;
    if (argc >= 2)
      USERATIO = atoi(argv[1]);
    if (argc >= 3)
      USEGROUND = atoi(argv[2]);
    if (argc >= 4)
      USESHIFT = atoi(argv[3]);
    if (USEGROUND==0)
      for (i=0; i<NUMPARTIES; i++)
        groundval[i] = 0;
    if (USESHIFT==0)
      for (i=0; i<NUMPARTIES; i++)
        shiftval[i] = 0;

    /* Initialise various counters. */
    for (i=0; i<MAXRIDINGNUM; i++) {
      ridingtot[i] = 0;
      ridingregion[i] = -1;
      for (j=0; j<NUMPARTIES; j++) {
        votecount[i][j] = 0;
      }
    }
	  printf("provincename[0]: **%s**\n", provincename[0]);
    for (j=0; j<NUMPARTIES; j++)
      seattot[j] = adjustedseattot[j] = 0;
    for (i=0; i<NUMREGIONS; i++) {
      regiontot[i] = regionseats[i] = 0;
      for (j=0; j<NUMPARTIES; j++)
        seatregion[i][j] = adjustedseatregion[i][j] = regionvote[i][j] = 0;
    }
    numseats = 0;
	  printf("provincename[0]: **%s**\n", provincename[0]);

    /* Open election results file. */
    if ((fp = fopen(EFILENAME,"r")) == NULL) {
        printf("OOPS, unable to read file %s.\n", EFILENAME);
        exit(1);
    }

    /* Read through file, collecting data. */
    printf("Reading through %s text data file ... \n", EFILENAME);
    if (verboselevel>0)
      printf("\n\n");
    fflush(stdout);
    nextline(fp); /* Omit headers. */
    filedone = 0;
    while (filedone==0) {

      /* Read a line of the file. */
      nextline(fp);

      /* Process the line. */
      if (filedone==0) {
        parseline();
	/* printf("**%s**\n", linedata[0]); */
	riding = atoi(linedata[2]);
	votes = atoi(linedata[6]);

	/* Identify province. */
	province = -1;
        for (i=0; i<NUMPROVINCES; i++) {
	  if (!strcmp(linedata[0], provincename[i]))
	    province = i;
	}
	if (province == -1) {
	  printf("OOPS, unable to identify province in input file.\n");
	  printf("field: **%s**\n", linedata[0]);
	  printf("provincename[0]: **%s**\n", provincename[0]);
	  exit(1);
	}
	region = regionmap[province];
	ridingregion[riding] = region;
	if (verboselevel>=2)
	  printf("province=%d, region=%d\n", province, region);

	/* Identify party. */
	party = -1;
        for (i=0; i<5; i++) {
	  if (!strcmp(linedata[3]+strlen(linedata[3])-strlen(partyname[i]),
	  					partyname[i]))
	    party = i;
	}

	if (verboselevel>=2)
	  printf("riding=%d, votes=%d, province=%d, party=%d\n",
		riding, votes, province, party);
	ridingtot[riding] = ridingtot[riding] + votes;
	regiontot[region] = regiontot[region] + votes;
	if (party >= 0) {
	  votecount[riding][party] = votes;
	  regionvote[region][party] = regionvote[region][party] + votes;
	  votetot[party] = votetot[party] + votes;
	}

      } /* End of "if (filedone==0)" branch. */
    }  /* End of "while" loop. */
    printf("done.\n");
    fclose(fp);

    /* Compute total number of votes cast. */
    tottot = 0;
    for (i=0; i<MAXRIDINGNUM; i++)
      tottot = tottot + ridingtot[i];
    if (verboselevel >= 1)
      printf("tottot=%d\n", tottot);

    /* Compute regional percentages. */
    for (i=0; i<NUMREGIONS; i++) {
      for (j=0; j<NUMPARTIES; j++) {
        regionperc[i][j] = 100.0 * regionvote[i][j] / regiontot[i];
	if (verboselevel >= 1)
	  printf("regionperc[%d][%d] = 100 * %d / %d = %f\n",
		i, j, regionvote[i][j], regiontot[i], regionperc[i][j]);
      }
    }
    for (j=0; j<NUMPARTIES; j++)
      totperc[j] = 100.0 * votetot[j] / tottot;

if (verboselevel>2) {
    for (j=0; j<NUMPARTIES; j++)
      printf("votecount[10001][%d] = %d\n", j, votecount[10001][j]);
    for (j=0; j<NUMPARTIES; j++)
      printf("votecount[10002][%d] = %d\n", j, votecount[10002][j]);
    for (j=0; j<NUMPARTIES; j++)
      printf("votecount[10003][%d] = %d\n", j, votecount[10003][j]);
}

    /* Compute riding winners. */
    for (i=0; i<MAXRIDINGNUM; i++) {
      if (ridingtot[i]>0) {
	numseats++;
	regionseats[ridingregion[i]] = regionseats[ridingregion[i]] + 1;
        winner = ridingmax = -1;
	for (j=0; j<NUMPARTIES; j++) {
	  if (votecount[i][j] > ridingmax) {
	    winner = j;
	    ridingmax = votecount[i][j];
	  }
	}
	if (verboselevel >= 2)
	  printf("winner of riding %d is %d\n", i, winner);
	if (winner<0) {
	  printf("OOPS, unable to identify winner in riding %d.\n", i);
	  exit(1);
	}
	ridingwinner[i] = winner;
	seatregion[ridingregion[i]][winner] =
			seatregion[ridingregion[i]][winner] + 1;
	seattot[winner] = seattot[winner] + 1;
      }
    }

    /* Output national election results. */
    printf("\n=============================== ELECTION RESULTS ============\n");
    printf("\nCANADA:\n\t");
    for (j=0; j<NUMPARTIES; j++)
      printf("\t%s", partytag[j]);
    printf("\tTotal");
    printf("\nVotes\t");
    for (j=0; j<NUMPARTIES; j++)
      printf("\t%d", votetot[j]);
    printf("\t%d", tottot);
    printf("\nPerc\t");
    for (j=0; j<NUMPARTIES; j++)
      printf("\t%.2f", totperc[j]);
    printf("\t100.00");
    printf("\nSeats\t");
    for (j=0; j<NUMPARTIES; j++)
      printf("\t%d", seattot[j]);
    printf("\t%d", numseats);
    printf("\n\n");

    /* Output regional election results. */
for (i=0; i<NUMREGIONS; i++) {
    printf("\n%s:\n\t", regiontag[i]);
    for (j=0; j<NUMPARTIES; j++)
      printf("\t%s", partytag[j]);
    printf("\tTotal");
    printf("\nVotes\t");
    for (j=0; j<NUMPARTIES; j++)
      printf("\t%d", regionvote[i][j]);
    printf("\t%d", regiontot[i]);
    printf("\nPerc\t");
    for (j=0; j<NUMPARTIES; j++)
      printf("\t%.2f", regionperc[i][j]);
    printf("\t100.00");
    printf("\nSeats\t");
    for (j=0; j<NUMPARTIES; j++)
      printf("\t%d", seatregion[i][j]);
    printf("\t%d", regionseats[i]);
    printf("\n");
}

    /* Read in the latest poll. */
    if ((fp = fopen("pollresults","r")) == NULL) {
        printf("OOPS, unable to read file %s.\n", "pollresults");
        exit(1);
    }
    for (i=0; i<NUMPARTIES; i++) {
      fscanf(fp, "%s", tmpstring);  /* Read in and ignore party name. */
      fscanf(fp, "%s", tmpstring); /* National poll number. */
      polltot[i] = atof(tmpstring);
      for (j=0; j<NUMREGIONS; j++) {
        fscanf(fp, "%s", tmpstring); /* Regional poll number. */
	pollreg[j][i] = atof(tmpstring);
	if (verboselevel >= 1)
	  printf("pollreg[%d][%d]=%f\n", j, i, pollreg[j][i]);
      }
    }
    fclose(fp);

    /* Output the election results summary. */
    printf("\n\n=============================== ELECTION RESULTS ======================\n");
    printf("\n\t");
    for (j=0; j<NUMPARTIES; j++)
      printf("\t%s", partytag[j]);
    printf("\n");
    printf("Canada:   ");
    for (j=0; j<NUMPARTIES; j++) {
      printf("\t%.2f", totperc[j]);
    }
    printf("\n");
    for (i=0; i<NUMREGIONS; i++) {
      printf("%s:   ", regiontag[i]);
      for (j=0; j<NUMPARTIES; j++) {
        printf("\t%.2f", regionperc[i][j]);
      }
      printf("\n");
    }

    /* Output the poll results summary. */
    printf("\n\n=============================== POLL RESULTS ======================\n");
    printf("\n\t");
    for (j=0; j<NUMPARTIES; j++)
      printf("\t%s", partytag[j]);
    printf("\n");
    printf("Canada:   ");
    for (j=0; j<NUMPARTIES; j++) {
      printf("\t%.1f", polltot[j]);
    }
    printf("\n");
    for (i=0; i<NUMREGIONS; i++) {
      printf("%s:   ", regiontag[i]);
      for (j=0; j<NUMPARTIES; j++) {
        printf("\t%.1f", pollreg[i][j]);
      }
      printf("\n");
    }

    /* Output the adjustments. */
    printf("\n\n==================== ADJUSTMENTS & PREDICTIONS =================\n");
    printf("\nUSEGROUND=%d,  USESHIFT=%d\n", USEGROUND, USESHIFT);
    printf("\n\t");
    for (j=0; j<NUMPARTIES; j++)
      printf("\t%s", partytag[j]);
    printf("\n");
    printf("Ground:   ");
    for (j=0; j<NUMPARTIES; j++) {
      printf("\t%.2f", groundval[j]);
    }
    printf("\n");
    printf("Shift:   ");
    for (j=0; j<NUMPARTIES; j++) {
      printf("\t%.2f", shiftval[j]);
    }

    /* Compute adjusted riding vote counts. */
    for (i=0; i<MAXRIDINGNUM; i++) {
      if (ridingtot[i]>0) {
	region = ridingregion[i];
	for (j=0; j<NUMPARTIES; j++) {
	  if (USERATIO)
	    adjustedcount[i][j] = votecount[i][j]
		* (pollreg[region][j]+groundval[j]+shiftval[j])
		/ regionperc[region][j];
	  else
	    adjustedcount[i][j] = votecount[i][j] + 0.01 *
	(pollreg[region][j]+groundval[j]+shiftval[j]-regionperc[region][j])
		* ridingtot[i];
	}
        winner = ridingmax = -1;
	for (j=0; j<NUMPARTIES; j++) {
	  if (adjustedcount[i][j] > ridingmax) {
	    winner = j;
	    ridingmax = adjustedcount[i][j];
	  }
	}
	if (verboselevel >= 2)
	  printf("adjusted winner of riding %d is %d\n", i, winner);
	if (winner<0) {
	  printf("OOPS, unable to identify adjusted winner in riding %d.\n", i);
	  exit(1);
	}
	adjustedwinner[i] = winner;
	adjustedseatregion[region][winner] =
			adjustedseatregion[region][winner] + 1;
	adjustedseattot[winner] = adjustedseattot[winner] + 1;
      }
    }

    /* Output predicted seat counts summary. */
    printf("\n\n=============================== PREDICTED SEAT COUNTS ================\n");
    printf("\n\t");
    for (j=0; j<NUMPARTIES; j++)
      printf("\t%s", partytag[j]);
    printf("\n");
    printf("Canada:   ");
    for (j=0; j<NUMPARTIES; j++) {
      printf("\t%d", adjustedseattot[j]);
    }
    printf("\n");
    for (i=0; i<NUMREGIONS; i++) {
      printf("%s:   ", regiontag[i]);
      for (j=0; j<NUMPARTIES; j++) {
        printf("\t%d", adjustedseatregion[i][j]);
      }
      printf("\n");
    }

    /* Finish up. */
    printf("\n");
    return(0);

}

/* nextline: read a file line into the "theline" variable */
nextline(FILE *fp)
{
    if (fgets(theline, 9999, fp)==NULL)
      filedone = 1;
    return(0);
}

/* parseline: parse a csv file line from "theline" */
parseline()
{
    int i, j;
    int fieldnum = 0;
    int fieldlength = 0;
    int done = 0;
    int linespot = 0;
    int quoted = 0;

    while (!done) {
      if (theline[linespot]=='"') {
	  /* Deal with opening/closing quotation marks. */
	  quoted = !quoted;
      } else if ((!quoted) && (theline[linespot]==',')) {
	  linedata[fieldnum][fieldlength] = '\0';
	  fieldnum++;
	  /* fieldlength++; */
	  fieldlength = 0;
	  if (fieldnum >= MAXNUMFIELDS) {
	    printf("ERROR: maximum number of fields exceeded.\n");
	    exit(1);
          }
      } else if ( (theline[linespot]=='\n') || (theline[linespot]=='\0') ) {
	  /* Deal with end of line. */
	  done = 1;
      } else {
	  /* Deal with ordinary characters. */
	  linedata[fieldnum][fieldlength] = theline[linespot];
	  fieldlength++;
	  if (fieldlength >= MAXFIELDLENGTH) {
	    linedata[fieldnum][fieldlength] = '\0';
	    printf("ERROR: field length %d >= %d: %s\n",
	    		fieldlength, MAXFIELDLENGTH, linedata[fieldnum]);
	    exit(1);
	  }
      }
      linespot++;
    }
    totnumfields = fieldnum;
    if (verboselevel>2)
      printf("parseline: totnumfields=%d\n", totnumfields);
    return(0);
}


int divides(int aa, int bb)
{
    return( (bb/aa)*aa == bb );
}