tvdb/tvdb.c

#include <stdio.h>
#include <sys/types.h>
#include <sys/stat.h>
#include <fcntl.h>
#include <string.h>
#include <unistd.h>
#include <stdlib.h>
#include <sqlite3.h>

#define CACHE "/mod/var/tvdb"
#define SERIESDB CACHE "/series.db"

#define EXEC(xx) if (sqlite3_exec(db, xx, NULL, NULL, &error) != SQLITE_OK) \
        do { \
                printf("Failed: %s - %s (%s)\n", xx, error, sqlite3_errmsg(db)); \
                sqlite3_free(error); \
                exit(0); \
        } while (0)

struct series {
	unsigned long id;
	char *imdb;
	char *name;
	char *overview;
	char *banner;
};

struct episode {
	unsigned long id;
	char *name;
	char *overview;
	char *thumb;
	unsigned long series;
	unsigned long episode;
};

#define HANDLE(str, c, n) \
    do { \
	if (!memcmp(p, str, (n + 1))) \
	{ \
		*p = c; \
		memmove(p + 1, p + (n + 1), l - (p - txt) - (n + 1)); \
		l -= n; \
		txt[l] = '\0'; \
	} \
    } while (0)

void
unescape(char *txt)
{
	char *p = txt;
	int l = strlen(txt);

	for (; (p = strchr(p, '&')); p++)
	{
		int ll = 0;
		unsigned char icode;
		/* sscanf -> 1: the code was read; ll>0: ';' came next */
		if ((1 == sscanf( p, "&#%hhu;%n", &icode, &ll) ||
		     1 == sscanf( p, "&#%*[xX]%hhx;%n", &icode, &ll)) && 
		     ll > 0) {
			/*  &#x<hex>;, &#<decimal>; */
			HANDLE(p, (char)icode, ll-1);
		} else {
			HANDLE("&amp;", '&', 4);
			HANDLE("&quot;", '"', 5);
			HANDLE("&apos;", '\'', 5);
			HANDLE("&lt;", '<', 3);
			HANDLE("&gt;", '>', 3);
		}
	}

	for (p = txt; (p = memchr(p, '\xe2', l - (p - txt))); p++)
	{	/* curly apostrophe, en dash, curly quotes */
		HANDLE("\xe2\x80\x99", '\'', 2);
		HANDLE("\xe2\x80\x93", '-', 2);
		HANDLE("\xe2\x80\x9c", '"', 2);
		HANDLE("\xe2\x80\x9d", '"', 2);
	}

	if ((p = strpbrk(txt, "\n\r")))
		*p = '\0';
}

int
main(int argc, char **argv)
{
	char *buf, *p, *q;
	struct series s;
	struct episode e;
	struct stat st;
	char *error;
	sqlite3_stmt *stmt;
	sqlite3 *db;
	int fd, epcnt = 0;
	int debug = 0;

	if (argc < 2)
	{
		printf("Syntax: %s [-d] <xml file>\n", argv[0]);
		return 0;
	}

	if (!strcmp(argv[1], "-d"))
	{
		debug = 1;
		argc--, argv++;
	}

	if (stat(argv[1], &st) == -1)
	{
		perror("stat");
		return 0;
	}

	if (!(buf = malloc(st.st_size)))
	{
		perror("malloc");
		return 0;
	}

	if ((fd = open(argv[1], O_RDONLY)) == -1)
	{
		perror("open");
		return 0;
	}

	if (read(fd, buf, st.st_size) != st.st_size)
	{
		perror("read");
		return 0;
	}

	if (debug)
		printf("Read - %lu bytes\n", st.st_size);

	sqlite3_config(SQLITE_CONFIG_SINGLETHREAD);

	if (sqlite3_open_v2(SERIESDB, &db,
	    SQLITE_OPEN_READWRITE | SQLITE_OPEN_CREATE, NULL)
	    != SQLITE_OK)
	{
		printf("Failed to open series database: %s\n",
		    sqlite3_errmsg(db));
		return 0;
	}

	EXEC("pragma journal_mode = off");
	EXEC("pragma synchronous = off");

	EXEC(
		"create table if not exists series("
		"[series_id] integer primary key, "
		"[imdb_id] text, "
		"[name] text, "
		"[overview] text, "
		"[banner] text, "
		"[dat] text"
		")"
	);

	if (sqlite3_prepare_v2(db,
	    "insert or replace into series "
	    "values(?,?,?,?,?,date('now'))",
	    -1, &stmt, NULL) != SQLITE_OK)
	{
		fprintf(stderr,
		    "Problem preparing series statement %s",
		    sqlite3_errmsg(db));
		exit(1);
	}

	p = strchr(buf, '<');
	for (;;)
	{
		if (!p)
			break;
		p++;

		if (!strncmp(p, "Series>", 7))
		{
			memset(&s, '\0', sizeof(s));
		}
		if (!strncmp(p, "id>", 3))
		{
			p += 3;
			s.id = strtoul(p, (char **)NULL, 10);
		}
		if (!strncmp(p, "IMDB_ID>", 8))
		{
			p += 8;
			if ((q = strchr(p, '<')))
				*q = '\0';
			s.imdb = strdup(p);
			p = q;
			continue;
		}
		if (!strncmp(p, "SeriesName>", 11))
		{
			p += 11;
			if ((q = strchr(p, '<')))
				*q = '\0';
			s.name = strdup(p);
			unescape(s.name);
			printf("Series: %s\n", s.name);
			p = q;
			continue;
		}
		if (!strncmp(p, "Overview>", 9))
		{
			p += 9;
			if ((q = strchr(p, '<')))
				*q = '\0';
			s.overview = strdup(p);
			unescape(s.overview);
			if (debug)
				printf("        %s\n", s.overview);
			p = q;
			continue;
		}
		if (!strncmp(p, "banner>", 7))
		{
			p += 7;
			if ((q = strchr(p, '<')))
				*q = '\0';
			s.banner = strdup(p);
			unescape(s.banner);
			p = q;
			continue;
		}

		if (!strncmp(p, "/Series>", 8))
		{
			if (s.id && s.name && s.overview)
			{
				sqlite3_bind_int(stmt, 1, s.id);
				sqlite3_bind_text(stmt, 2,
				    s.imdb ? s.imdb : "", -1, NULL);
				sqlite3_bind_text(stmt, 3, s.name, -1, NULL);
				sqlite3_bind_text(stmt, 4, s.overview,
				    -1, NULL);
				sqlite3_bind_text(stmt, 5, s.banner, -1, NULL);
				sqlite3_step(stmt);
			}
			break;
		}
		p = strchr(p, '<');
	}
	sqlite3_finalize(stmt);

	if (!s.id)
	{
		printf("No series found.\n");
		return 0;
	}

	/* Attach the Episode database for this series. */
	sprintf(buf, "attach '%s/%lu.db' as episode", CACHE, s.id);
	EXEC(buf);

	EXEC("drop table if exists episode.episode");
	EXEC(
		"create table if not exists episode.episode("
		"[episode_id] integer primary key, "
		"[series] integer, "
		"[episode] integer, "
		"[name] text, "
		"[overview] text, "
		"[thumb] text"
		")"
	);

	if (sqlite3_prepare_v2(db,
	    "insert or replace into episode.episode "
	    "values(?,?,?,?,?,?)",
	    -1, &stmt, NULL) != SQLITE_OK)
	{
		fprintf(stderr,
		    "Problem preparing statement %s", sqlite3_errmsg(db));
		exit(1);
	}

	for (;;)
	{
		if (!p)
			break;
		p++;

		if (!strncmp(p, "Episode>", 8))
		{
			/* New episode */
			memset(&e, '\0', sizeof(e));
		}

		if (!strncmp(p, "id>", 3))
		{
			p += 3;
			e.id = strtoul(p, (char **)NULL, 10);
		}
		if (!strncmp(p, "Combined_episodenumber>", 23)) {
			p += 23;
			e.episode = strtoul(p, (char **)NULL, 10);
		}
		if (!strncmp(p, "Combined_season>", 16)) {
			p += 16;
			e.series = strtoul(p, (char **)NULL, 10);
		}
		if (!strncmp(p, "EpisodeName>", 12))
		{
			p += 12;
			if ((q = strchr(p, '<')))
				*q = '\0';
			e.name = strdup(p);
			unescape(e.name);
			printf("Episode: %s\n", e.name);
			p = q;
			continue;
		}
		if (!strncmp(p, "Overview>", 9))
		{
			p += 9;
			if ((q = strchr(p, '<')))
				*q = '\0';
			e.overview = strdup(p);
			unescape(e.overview);
			if (debug)
				printf("         %s\n", e.overview);
			p = q;
			continue;
		}
		if (!strncmp(p, "filename>", 9))
		{
			p += 9;
			if ((q = strchr(p, '<')))
				*q = '\0';
			e.thumb = strdup(p);
			unescape(e.thumb);
			p = q;
			continue;
		}

		if (!strncmp(p, "/Episode>", 9))
		{
			/* End of episode */
			if (e.id && e.name && e.overview)
			{
				epcnt++;
				sqlite3_bind_int(stmt, 1, e.id);
				sqlite3_bind_int(stmt, 2, e.series);
				sqlite3_bind_int(stmt, 3, e.episode);
				sqlite3_bind_text(stmt, 4, e.name, -1, NULL);
				sqlite3_bind_text(stmt, 5, e.overview,
				    -1, NULL);
				sqlite3_bind_text(stmt, 6, e.thumb, -1, NULL);
				sqlite3_step(stmt);
				sqlite3_reset(stmt);
				sqlite3_clear_bindings(stmt);
			}
			if (e.name) free(e.name);
			if (e.overview) free(e.overview);
			if (e.thumb) free(e.thumb);
			memset(&e, '\0', sizeof(e));
		}

		p = strchr(p, '<');
	}

	close(fd);
	free(buf);
	sqlite3_finalize(stmt);
	sqlite3_close(db);

	printf("Episodes: %d\n", epcnt);

	return 0;
}