Team:Paris Liliane Bettencourt/Project/SIP

From 2010.igem.org

Revision as of 20:39, 23 October 2010 by Theotime (Talk | contribs)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)

SIP Wiki Analyser : Find iGEM winner with satistics

So this year, iGEM team Paris try to find who will win the iGEM competition. We have several aproach to find that, using data on the wikis.
To analyse the wiki, we have implemented an algorithm called SIP (Statistically Improbable Phrases), used by the Amazon website to caracterize their books. We try to find which are the most improbable words in a large sample (we use all iGEM wiki) but the most probable in the context.

In Math language :

We have f/F
f the frequence of the word in the context : f = o/n (o is the occurence of the word, and n the number of words in the wiki).
F the frequence of the same word in a large sample : F = O/N (O, occurence in the sample, and N the number of words in the sample).

In this code, we use wget software to get all wiki of one year, then we remove all character non-alpha-numeric, and convert to lower case. New we compute the frenquencies of each words in a sqlite3 database.
We have extract and calculate this for the last 3 years. Available on the bottom of the page.

We will try next to display these words with graphical effects using [http://www.wordle.net/create this software written in java] and also we want to compare this words with words used in Instructor's papers (using ENTREZ API), to determine how Instructors influence the team. We think in the last part that an influence by the Instructor has positive results on the team, a better cooperation between memberships.

SIP.C

/* SIP.C *** Make SIP database and dictionary

PUBLIC DOMAIN
From iGEM team 2010 Paris

comment:

 This code permit you to calculate SIP words (most improbable words) in each
 wiki team.
 You need sqlite3, wget and links to run this program.

build:

 $ gcc -o sip sip.c -lsqlite3

usage :

 $ ./sip [list of team name] [name of database] [year]


Sqlite3 database :

Table for team.
+--------+-----------+------------+-----------+
| Words  | local occ | local freq | SIP value |
+--------+-----------+------------+-----------+
| string |   u_long  |  float     |   float   |  
|        |           |            |           |
+--------+-----------+------------+-----------+


Table name : Dictionary.
+--------+------------+-------------+
| word   | global_occ | global_freq |
+--------+------------+-------------+
| string |   u_long   |  float      |
|        |            |             |
+--------+------------+-------------+

*/

#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <dirent.h>
#include <sqlite3.h>

#define MAXBUFFER	256
#define LEN_CMD		96

#define TEAM_NAME 	1
#define DATABASE_NAME	2
#define YEAR		3

#define FALSE 		0
#define TRUE		1

#define IS_REP		1
#define NO_REP		2

/*
isRep say if the stream is a directory or a file.
@entryname : name of the file
return : IS_REP it's a directory, NO_REP it's a file
*/
int isRep( char *entryname )
{
	FILE *fp;
	fp = fopen( entryname, "rb" );

	if ( fp == 0 )
		return IS_REP;
	else {
		fclose( fp );
		return NO_REP; }
}

void chgchar( char *string, char c, char r, int len )
{
	int i = 0;

	while ( i < len )
	{
		if ( string[i] == c )
		{
			
			string[i] = r;
			i++;
		}
		else
			i++;
	}
}

int close_sqlitedb( sqlite3*** conn )
{
	if ( sqlite3_close(**conn) != SQLITE_OK )
	{
		printf("Error closing the db: %s\n", sqlite3_errmsg(*conn));
	}
}

void read_next_word( char word[MAXBUFFER], FILE *fp )
{
	char buffer[MAXBUFFER];
	char ch;
	int n = 0;

	ch = fgetc(fp);

	if( ch == EOF ) {
		return;
	}

	while ( ch != 0x20 && ch != EOF && n < MAXBUFFER )
	{
		buffer[n] = ch;
		n++;
		ch = fgetc(fp);
	}
	buffer[n] = '\0';
	strcpy( word, buffer );
}

int is_word_valid( char word )
{
// there's no filter at the moment.
// We can compare with MeSH white list ie.
}

int is_word_exists( char *word, sqlite3_stmt **statement, sqlite3** conn )
{
	if ( sqlite3_bind_text( *statement, 1, word, -1, SQLITE_STATIC) != SQLITE_OK ) {
		printf("Error binding to query select: %s\n", sqlite3_errmsg(*conn));
		close_sqlitedb( &conn );
		exit(1);
	}

	if ( sqlite3_step( *statement ) != SQLITE_DONE ) {
		sqlite3_reset( *statement );
		return TRUE;
	} else {
		sqlite3_reset( *statement );
		return FALSE;
	}

}

void add_to_the_list( char *word, sqlite3_stmt **statement, sqlite3** conn )
{
	if ( sqlite3_bind_text( *statement, 1, word, -1, SQLITE_STATIC) != SQLITE_OK ) {
		printf("Error binding to query insert: %s\n", sqlite3_errmsg(*conn));
		close_sqlitedb( &conn );
		exit(1);
	}

	sqlite3_step( *statement );
	sqlite3_reset( *statement );
}

void inc_occ( char *word, sqlite3_stmt **statement, sqlite3** conn )
{
// UPDATE dictionary SET global_occ = global_occ + 1 WHERE word=word
// INSERT into dictionary ...
	if ( sqlite3_bind_text( *statement, 1, word, -1, SQLITE_STATIC) != SQLITE_OK ) {
		printf("Error binding to query update: %s\n", sqlite3_errmsg(*conn));
		close_sqlitedb( &conn );
		exit(1);
	}

	sqlite3_step( *statement );
	sqlite3_reset( *statement );
}

void compute_freq( sqlite3** conn, float nbr_of_words, char *table )
{
// UPDATE dictionary SET global_freq = global_occ / nbr_of_word
// note that nbr_of_word must be a REAL type.
/* For each words we compute : F = occ_of_word / nbr_of_words */
	char sqlite3_query[MAXBUFFER];
	sqlite3_stmt *statement;

	sprintf( sqlite3_query, "UPDATE %s SET global_freq = global_occ / %f",
		table, nbr_of_words);

	if ( sqlite3_prepare_v2(*conn, sqlite3_query, -1, &statement, NULL ) != SQLITE_OK )
	{
		printf("Error compiling the request: %s\n", sqlite3_errmsg(*conn) );
		close_sqlitedb( &conn );
		exit(1);
	}
	sqlite3_step( statement );
	sqlite3_finalize( statement );
}

make_dictionary( char *team_list, char *database_name)
{
	FILE *fp, *fp_wiki;
	float nbr_of_words = 0;
	char word[MAXBUFFER], digest_name[MAXBUFFER-LEN_CMD], team_name[MAXBUFFER-LEN_CMD];
	int size_name;

	sqlite3_stmt *stmt_select, *stmt_insert, *stmt_update;

	sqlite3 *conn;
	if ( sqlite3_open( database_name, &conn ) != SQLITE_OK )
	{
		printf("Error opening the db: %s\n", sqlite3_errmsg(conn));
		exit(1);
	}

	fp = fopen(team_list, "r");

	if ( fp == 0 ) {
		printf("Can't open %s file\n", team_list);
		exit(1);
	}

	// pre-compile the query to optimize the process
	if ( sqlite3_prepare_v2(conn, "SELECT * FROM dictionary WHERE word=?", -1,
		&stmt_select, NULL ) != SQLITE_OK )
	{
		printf("Error compiling the request: %s\n", sqlite3_errmsg(conn) );
		goto CLOSE_DB;
		exit(1);
	}

	if ( sqlite3_prepare_v2(conn, "INSERT INTO dictionary VALUES (?, 1, NULL)", -1,
		&stmt_insert, NULL ) != SQLITE_OK )
	{
		printf("Error compiling the request: %s\n", sqlite3_errmsg(conn) );
		goto CLOSE_DB;
		exit(1);
	}

	if ( sqlite3_prepare_v2(conn, "UPDATE dictionary \
		SET global_occ = global_occ + 1 WHERE word=?",
		-1, &stmt_update, NULL ) != SQLITE_OK )
	{
		printf("Error compiling the request: %s\n", sqlite3_errmsg(conn) );
		goto CLOSE_DB;
		exit(1);
	}

	while( !feof(fp) )
	{
		fgets( team_name, (MAXBUFFER - LEN_CMD), fp );

		if ( feof(fp) ) {
			break;
		}

		size_name = strlen( team_name );
		team_name[size_name - 1] = '\0';
		chgchar( team_name, '-', '_', size_name);
		printf("On the %s Team\n", team_name);

		if ( chdir(team_name) == -1 ) {
			printf("error changing directory to %s\n", team_name);
			exit(1);
		}

		strncpy( digest_name, team_name, (MAXBUFFER-LEN_CMD-4) );
		strcat( digest_name, ".xtr");

		fp_wiki = fopen( digest_name, "r" );
		if( fp_wiki == 0 ) {
			printf("error when you open %s file", digest_name);
			exit(1);
		}

		if ( chdir("..") == -1 ) {
			printf("error changing directory to ..\n");
			exit(1);
		}


		sqlite3_exec(conn, "begin", NULL, NULL, NULL );
		while ( !feof(fp_wiki) )
		{
			read_next_word( word, fp_wiki );
			if ( is_word_valid( word ) == FALSE ) {
				continue;
			}
			if ( is_word_exists( word, &stmt_select, &conn ) == FALSE ) {
				add_to_the_list( word, &stmt_insert, &conn );
			}
			else {
				inc_occ( word, &stmt_update, &conn );
			}
		
			nbr_of_words++;
		}
		sqlite3_exec(conn, "commit", NULL, NULL, NULL );
		fclose(fp_wiki);
	}
	fclose(fp);
	sqlite3_finalize( stmt_select );
	sqlite3_finalize( stmt_insert );
	sqlite3_finalize( stmt_update );

	compute_freq( &conn, nbr_of_words, "dictionary" );

CLOSE_DB :
	if ( sqlite3_close(conn) != SQLITE_OK ) {
		printf("Error closing the db: %s\n", sqlite3_errmsg(conn));
	}
}

void make_database( char *team_list, char *database_name )
{
	sqlite3 *conn;
	sqlite3_stmt *statement;
	FILE* fp;
	int size_name;
	char team_name[MAXBUFFER-LEN_CMD], sqlite_query[MAXBUFFER];

	if ( sqlite3_open( database_name, &conn ) != SQLITE_OK )
	{
		printf("Error opening the db: %s\n", sqlite3_errmsg(conn));
		exit(1);
	}

	if ( sqlite3_prepare_v2(conn, "CREATE TABLE dictionary ( word text,  \
		global_occ int, global_freq real )", -1, &statement, NULL ) != SQLITE_OK )
	{
		printf("Error compiling the request: %s\n", sqlite3_errmsg(conn) );
		goto CLOSE_DB;
		exit(1);
	}
	
	sqlite3_step( statement );
	sqlite3_finalize( statement );

	fp = fopen(team_list, "r");

	if ( fp == 0 ) {
		printf("Can't open %s file\n", team_list);
		exit(1);
	}

	while( !feof(fp) )
	{
		fgets( team_name, (MAXBUFFER - LEN_CMD), fp );

		if ( feof(fp) ) {
			break;
		}

		size_name = strlen( team_name );
		team_name[size_name - 1] = '\0';
		chgchar( team_name, '-', '_', size_name);
		printf("Create %s Team Table\n", team_name);

		sprintf( sqlite_query, "CREATE TABLE %s ( word text, global_occ int, \
			global_freq real, sip real )", team_name);
		if ( sqlite3_prepare_v2(conn, sqlite_query, -1, &statement, NULL ) != SQLITE_OK )
		{
			printf("Error compiling the request: %s\n", sqlite3_errmsg(conn) );
			goto CLOSE_DB;
			exit(1);
		}

		sqlite3_step( statement );
		sqlite3_reset( statement );

	}
	fclose( fp);
	sqlite3_finalize( statement );

CLOSE_DB :
	if ( sqlite3_close(conn) != SQLITE_OK ) {
		printf("Error closing the db: %s\n", sqlite3_errmsg(conn));
	}
	
}

void download_wiki( char *team_list, char* year )
{
	FILE *fp_src, *fp_tar, *fp;
	char ch, last_ch;
	char buffer[MAXBUFFER], team_name[MAXBUFFER-LEN_CMD], digest_name[MAXBUFFER-LEN_CMD];
	int size_name;
	DIR *dp;
	struct dirent *entry;

	fp = fopen(team_list, "r");

	if ( fp == 0 ) {
		printf("Can't open %s file\n", team_list);
		exit(1);
	}

	while( !feof(fp) )
	{
		fgets( team_name, (MAXBUFFER - LEN_CMD), fp );

		if ( feof(fp) ) {
			break;
		}

		size_name = strlen( team_name );
		team_name[size_name - 1] = '\0';

		chgchar( team_name, '-', '_', size_name);

		if ( mkdir(team_name, 0777) == -1 ) {
			printf("error in the mkdir\n");
			exit(1);
		}

		if ( chdir(team_name) == -1 ) {
			printf("error changing the directory to %s\n", team_name);
			exit(1);
		}

		if ( strcmp( year, "2007") == 0 ) {
			snprintf(buffer, MAXBUFFER, "wget -R.jpg -R.png -R.gif \
				-R.jpeg -RUser:* -E -l1 -nd -r \
				http://parts.mit.edu/igem07/index.php/%s", team_name);
			printf("CMD : %s\n", buffer);
			system(buffer);
		} else {
			snprintf(buffer, MAXBUFFER, "wget -E -ITeam:%s -nd -r \
				http://%s.igem.org/Team:%s", team_name, year, team_name);
			printf("CMD : %s\n", buffer);
			system(buffer);
		}
	
		if((dp = opendir(".")) == NULL) {
			printf("can't open the dir");
			return;
		}
        	   
		while((entry = readdir(dp)) != NULL) {
			if ( isRep(entry->d_name) != IS_REP ) {
				if ( (strstr(entry->d_name, ".html") != NULL)
					&& (strstr(entry->d_name, ".dumpx") == NULL) ) {
				snprintf(buffer, MAXBUFFER, "links -dump %s > %s.dumpx",
					entry->d_name, entry->d_name);
				system(buffer);
				}	
			}
		}
		closedir(dp);
		snprintf(buffer, MAXBUFFER, "cat *.dumpx > %s", team_name);
		system(buffer);

		strncpy( digest_name, team_name, (MAXBUFFER-LEN_CMD-4) );
		strcat( digest_name, ".xtr");
	
		fp_src = fopen(team_name, "r");
		if ( fp == 0 ) {
			printf("Can't open %s file\n", team_name);
			exit(1);
		}

		fp_tar = fopen(digest_name, "w");
		if ( fp == 0 ) {
			printf("Can't open %s file\n", digest_name);
			exit(1);
		}

		while( !feof(fp_src) )
		{
			ch = fgetc(fp_src);
	
			if ( (ch >= 0x41 && ch <= 0x5A) ||
			(ch >= 0x61 && ch <= 0x7A) ||
			(ch == 0x20 && last_ch != 0x20) ||
			ch == 0x2D ||
			(ch >= 0x30 && ch <= 0x39) ) {
				if (ch >= 0x41 && ch <= 0x5A) { //put in lowercase
					ch += 0x20;
				}
				fputc(ch, fp_tar);
			}
			else {
				ch = 0x20;
	
				if ( last_ch != 0x20 ) {
					fputc(ch, fp_tar);
				}
			}
			last_ch = ch;
		}
		
		fclose( fp_src );
		fclose( fp_tar );

		if ( chdir("..") == -1 ) {
			printf("error changing the directory to %s\n", team_name);
			exit(1);
		}
	}
	fclose(fp);
}

void compute_sip( sqlite3 **conn, char *team_name )
{
// SELECT * from TEAM_NAME;
// pour chaque result -> get the freq | get the word => get the freq of word in all wiki
// f/F
// UPDATE where word='yourword'

	sqlite3_stmt *statement, *stmt_select, *stmt_update;
	char sqlite_query[MAXBUFFER];
	float local_freq, global_freq;
	int res;

	snprintf( sqlite_query, MAXBUFFER, "SELECT * from %s", team_name);
	if ( sqlite3_prepare_v2(*conn, sqlite_query, -1, &statement, NULL ) != SQLITE_OK ) {
		printf("compute_sip() : Error compiling the request 1: %s\n",
			sqlite3_errmsg(*conn) );
		close_sqlitedb( &conn );
		exit(1);
	}

	if ( sqlite3_prepare_v2(*conn, "SELECT * from dictionary WHERE word=?", -1,
				&stmt_select, NULL ) != SQLITE_OK )
	{
		printf("compute_sip() : Error compiling the request 2: %s\n",
			sqlite3_errmsg(*conn) );
		close_sqlitedb( &conn );
		exit(1);
	}

	snprintf( sqlite_query, MAXBUFFER, "UPDATE %s SET sip = ?/? WHERE word=?",
		team_name);
	if ( sqlite3_prepare_v2(*conn, sqlite_query, -1, &stmt_update, NULL ) != SQLITE_OK ) {
		printf("compute_sip() : Error compiling the request 3: %s\n",
			sqlite3_errmsg(*conn) );
		close_sqlitedb( &conn );
		exit(1);
	}

	sqlite3_exec(*conn, "begin", NULL, NULL, NULL );
	while( (res = sqlite3_step(statement)) == SQLITE_ROW)
	{
		local_freq = (float)sqlite3_column_double(statement, 2);

		if ( sqlite3_bind_text( stmt_select, 1, sqlite3_column_text(statement, 0), -1,
					SQLITE_STATIC) != SQLITE_OK ) {
			printf("compute_sip() : Error binding 1: %s\n", sqlite3_errmsg(*conn));
			close_sqlitedb( &conn );
			exit(1);
		}
		sqlite3_step( stmt_select );
		global_freq = (float)sqlite3_column_double(stmt_select, 2);

		if ( sqlite3_bind_double( stmt_update, 1, (double)local_freq) != SQLITE_OK ) {
			printf("compute_sip() : Error binding 2: %s\n", sqlite3_errmsg(*conn));
			close_sqlitedb( &conn );
			exit(1);
		}

		if ( sqlite3_bind_double( stmt_update, 2, (double)global_freq) != SQLITE_OK ) {
			printf("compute_sip() : Error binding 3: %s\n", sqlite3_errmsg(*conn));
			close_sqlitedb( &conn );
			exit(1);
		}

		if ( sqlite3_bind_text( stmt_update, 3, sqlite3_column_text(statement, 0), -1,
					SQLITE_STATIC) != SQLITE_OK ) {
			printf("compute_sip() : Error binding 4: %s\n", sqlite3_errmsg(*conn));
			close_sqlitedb( &conn );
			exit(1);
		}

		sqlite3_step( stmt_update );

		sqlite3_reset( stmt_select );
		sqlite3_reset( stmt_update );
	}
	sqlite3_exec(*conn, "commit", NULL, NULL, NULL );

	sqlite3_finalize( statement );
	sqlite3_finalize( stmt_select );
	sqlite3_finalize( stmt_update );
}

void make_sipword( char *team_list, char *database_name)
{
	FILE *fp_wiki, *fp;
	float nbr_of_words = 0;
	char word[MAXBUFFER], digest_name[MAXBUFFER-LEN_CMD], team_name[MAXBUFFER-LEN_CMD];
	char sqlite_query[MAXBUFFER];
	int size_name;

	sqlite3_stmt *stmt_select, *stmt_insert, *stmt_update;

	sqlite3 *conn;
	if ( sqlite3_open( database_name, &conn ) != SQLITE_OK )
	{
		printf("Error opening the db: %s\n", sqlite3_errmsg(conn));
		exit(1);
	}
	
	fp = fopen(team_list, "r");

	if ( fp == 0 ) {
		printf("Can't open %s file\n", team_list);
		exit(1);
	}

	while( !feof(fp) )
	{
		fgets( team_name, (MAXBUFFER - LEN_CMD), fp );

		if ( feof(fp) ) {
			break;
		}

		size_name = strlen( team_name );
		team_name[size_name - 1] = '\0';
		chgchar( team_name, '-', '_', size_name);

		printf("On the %s Team\n", team_name);

		// pre-compile the query to optimize the process
		snprintf( sqlite_query, MAXBUFFER, "SELECT * FROM %s WHERE word=?",
			team_name);

		if ( sqlite3_prepare_v2(conn, sqlite_query , -1,
			&stmt_select, NULL ) != SQLITE_OK )
		{
			printf("Error compiling the request: %s\n", sqlite3_errmsg(conn) );
			goto CLOSE_DB;
			exit(1);
		}

		snprintf( sqlite_query, MAXBUFFER, "INSERT INTO %s \
			VALUES (?, 1, NULL, NULL)", team_name);

		if ( sqlite3_prepare_v2(conn, sqlite_query, -1,
			&stmt_insert, NULL ) != SQLITE_OK )
		{
			printf("Error compiling the request: %s\n", sqlite3_errmsg(conn) );
			goto CLOSE_DB;
			exit(1);
		}
	
		snprintf( sqlite_query, MAXBUFFER, "UPDATE %s \
			SET global_occ = global_occ + 1 WHERE word=?", team_name);

		if ( sqlite3_prepare_v2(conn, sqlite_query, -1,
			&stmt_update, NULL ) != SQLITE_OK )
		{
			printf("Error compiling the request: %s\n", sqlite3_errmsg(conn) );
			goto CLOSE_DB;
			exit(1);
		}
	
		if ( chdir(team_name) == -1 ) {
			printf("error changing directory to %s\n", team_name);
			exit(1);
		}

		strncpy( digest_name, team_name, MAXBUFFER-LEN_CMD-4 );
		strcat( digest_name, ".xtr");

		fp_wiki = fopen( digest_name, "r" );
		if( fp_wiki == 0 ) {
			printf("error when you open %s file", digest_name);
			exit(1);
		}

		if ( chdir("..") == -1 ) {
			printf("error changing directory to ..\n");
			exit(1);
		}


		sqlite3_exec(conn, "begin", NULL, NULL, NULL );
		while ( !feof(fp_wiki) )
		{
			read_next_word( word, fp_wiki );
			if ( is_word_valid( word ) == FALSE ) {
				continue;
			}
			if ( is_word_exists( word, &stmt_select, &conn ) == FALSE ) {
				add_to_the_list( word, &stmt_insert, &conn );
			}
			else {
				inc_occ( word, &stmt_update, &conn );
			}
		
			nbr_of_words++;
		}
		sqlite3_exec(conn, "commit", NULL, NULL, NULL );
		fclose(fp_wiki);
		sqlite3_finalize( stmt_select );
		sqlite3_finalize( stmt_insert );
		sqlite3_finalize( stmt_update );

		compute_freq( &conn, nbr_of_words, team_name );
		compute_sip( &conn, team_name);
	}

CLOSE_DB :
	if ( sqlite3_close(conn) != SQLITE_OK ) {
		printf("Error closing the db: %s\n", sqlite3_errmsg(conn));
	}
}

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

	if ( argv[TEAM_NAME] == 0 || argv[DATABASE_NAME] == 0 ) {
		printf("usage : %s [list of team name] [name of database] [year]\n \
	- don't put [year] if you want to skip the download step.\n", argv[0] );
		exit(1);
	}

	if ( argv[YEAR] == 0 ) {
		printf("skip the downlad step !\n");
	} else {
		printf("start to download wiki !\n");
		download_wiki( argv[TEAM_NAME], argv[YEAR] );
	}

	printf("make the database !\n");
	make_database( argv[TEAM_NAME], argv[DATABASE_NAME] );

	printf("start to make a dictionary !\n");
	make_dictionary( argv[TEAM_NAME], argv[DATABASE_NAME] );

	printf("start calculate SIP words !\n");
	make_sipword( argv[TEAM_NAME], argv[DATABASE_NAME] );

	return 0;
}

Downloads

Team List

Wiki Data

SIP Database

To read databases, use [http://www.sqlite.org/ sqlite3].

Warning : Notice these files are generated using "links -dump" to remove html, to speed the process, but you can do without that, because SIP will remove them later. With links, some pages with special characters like '(' ')' and ':' in their name are not converted, we consider it's not very important, because it's a small number of pages, but you can re-gen the database without the html parse step.
You can also use html2text, but if the software find special character, it don't remove the html.
Also, I know there're some few bugs like in 2009's files : Illinois-tools is not downloaded... But unfortunately, I didn't fix them before the deadline.

References

http://en.wikipedia.org/wiki/Statistically_Improbable_Phrases
http://petewarden.typepad.com/searchbrowser/2008/01/whats-the-secre.html