trunk/proteinstructure/PdbaseInfo.java

package proteinstructure;

import tools.MySQLConnection;

import java.sql.Statement;
import java.sql.ResultSet;
import java.sql.SQLException;
import java.util.ArrayList;
import java.util.HashMap;
import java.util.Collections;

public class PdbaseInfo {
        public final static String MYSQLSERVER="white";
        public final static String MYSQLUSER=getUserName();
        public final static String MYSQLPWD="nieve";
        public final static String mymsdsdDB="my_msdsd_00_07_a";
        public final static String msdsdDB="msdsd_00_07_a";
        public final static String pdbaseDB="pdbase";
        
        MySQLConnection conn;
        String accode="";
        String chaincode="";
        int model=DEFAULT_MODEL;
        String chain=""; // the internal pdbase chain identifier (asym_id)
        int entrykey;
        String asymid;
        int entitykey;
        String alt_locs_sql_str;
        
        static int DEFAULT_MODEL=1;
        
        PdbaseInfo (String accode, String chaincode, int model_serial, String db) throws PdbaseInconsistencyError, PdbaseAcCodeNotFoundError{
                this.accode=accode;
                this.chaincode=chaincode;
                this.model=model_serial;
                this.conn = new MySQLConnection(MYSQLSERVER,MYSQLUSER,MYSQLPWD,db);
        this.entrykey=get_entry_key();
        this.asymid=get_asym_id();
        this.entitykey=get_entity_key();
        this.alt_locs_sql_str=get_atom_alt_locs();

        }
        
        PdbaseInfo (String accode, String chaincode, String db) throws PdbaseInconsistencyError, PdbaseAcCodeNotFoundError {
                this(accode,chaincode,DEFAULT_MODEL,db);
        }
        PdbaseInfo (String accode, String chaincode, int model_serial) throws PdbaseInconsistencyError, PdbaseAcCodeNotFoundError {
                this(accode,chaincode,model_serial,pdbaseDB);
        }
        
        PdbaseInfo (String accode, String chaincode) throws PdbaseInconsistencyError, PdbaseAcCodeNotFoundError {
                this(accode,chaincode,DEFAULT_MODEL,pdbaseDB);
        }

        /** get user name from operating system (for use as database username) */
        private static String getUserName() {
                String user = null;
                user = System.getProperty("user.name");
                if(user == null) {
                        System.err.println("Could not get user name from operating system. Exiting");
                        System.exit(1);
                }
                return user;
        }
        
        public void close() {
                conn.close();
        }
        
        public int get_entry_key() throws PdbaseAcCodeNotFoundError {
                String sql="SELECT entry_key FROM struct WHERE entry_id='"+accode.toUpperCase()+"'";
                try {
                        Statement stmt = conn.createStatement();
                        ResultSet rsst = stmt.executeQuery(sql);
                        if (rsst.next()) {
                                entrykey = rsst.getInt(1);
                                if (! rsst.isLast()) {
                                        System.err.println("More than 1 entry_key match for accession_code="+accode+", chain_pdb_code="+chaincode);
                                        throw new PdbaseAcCodeNotFoundError();                                  
                                }
                        } else {
                                System.err.println("No entry_key match for accession_code="+accode+", chain_pdb_code="+chaincode);
                                throw new PdbaseAcCodeNotFoundError();
                        }
                        rsst.close();
                        stmt.close();
                } catch (SQLException e) {
                        e.printStackTrace();
                }
                return entrykey;
        }
        
        public String get_asym_id() throws PdbaseInconsistencyError {
                String pdbstrandid=chaincode;
                if (chaincode.equals("NULL")){
                        pdbstrandid="A";
                }
                String sql="SELECT asym_id " +
                                " FROM pdbx_poly_seq_scheme " +
                                " WHERE entry_key=" + entrykey +
                                " AND pdb_strand_id='"+pdbstrandid+"' " +
                                " LIMIT 1";
                try {
                        Statement stmt = conn.createStatement();
                        ResultSet rsst = stmt.executeQuery(sql);
                        if (rsst.next()) {
                                asymid = rsst.getString(1);
                        } else {
                                System.err.println("No asym_id match for entry_key="+entrykey+", pdb_strand_id="+chaincode);
                                throw new PdbaseInconsistencyError("No asym_id match for entry_key="+entrykey+", pdb_strand_id="+chaincode);
                        }
                        rsst.close();
                        stmt.close();
                } catch (SQLException e) {
                        e.printStackTrace();
                }
                // we set the internal chain identifier self.chain from asymid
                chain = asymid;
                return asymid;  
        }
        
        public int get_entity_key() throws PdbaseInconsistencyError {
                String sql="SELECT entity_key " +
                                " FROM struct_asym " +
                                " WHERE entry_key="+ entrykey +
                                " AND id='"+asymid+"'";
                try {
                        Statement stmt = conn.createStatement();
                        ResultSet rsst = stmt.executeQuery(sql);
                        if (rsst.next()) {
                                entitykey = rsst.getInt(1);
                                if (! rsst.isLast()) {
                                        System.err.println("More than 1 entity_key match for entry_key="+entrykey+", asym_id="+asymid);
                                        throw new PdbaseInconsistencyError("More than 1 entity_key match for entry_key="+entrykey+", asym_id="+asymid);                                 
                                }
                        } else {
                                System.err.println("No entity_key match for entry_key="+entrykey+", asym_id="+asymid);
                                throw new PdbaseInconsistencyError("No entity_key match for entry_key="+entrykey+", asym_id="+asymid);
                        }
                        rsst.close();
                        stmt.close();
                } catch (SQLException e) {
                        e.printStackTrace();
                }
                return entitykey;
        }
        
        public String get_atom_alt_locs() throws PdbaseInconsistencyError{
                ArrayList<String> alt_ids = new ArrayList<String>();
                HashMap<String,Integer> alt_ids2keys = new HashMap<String,Integer>();
                String alt_loc_field="label_alt_key";
                String sql="SELECT id, atom_sites_alt_key FROM atom_sites_alt WHERE entry_key="+entrykey;
                try {
                        Statement stmt = conn.createStatement();
                        ResultSet rsst = stmt.executeQuery(sql);
                        int count=0;
                        while (rsst.next()) {
                                count++;
                                alt_ids.add(rsst.getString(1));
                                alt_ids2keys.put(rsst.getString(1), rsst.getInt(2));
                        }
                        if (count!=0){
                                if ((! alt_ids.contains(".")) || alt_ids.indexOf(".")!=alt_ids.lastIndexOf(".")){ // second term is a way of finding out if there is more than 1 ocurrence of "." in the ArrayList 
                                        System.err.println("alt_codes exist for entry_key "+entrykey+" but there is either no default value '.' or more than 1 '.'. Something wrong with this entry_key or with "+pdbaseDB+" db!");
                                        throw new PdbaseInconsistencyError("alt_codes exist for entry_key "+entrykey+" but there is either no default value '.' or more than 1 '.'. Something wrong with this entry_key or with "+pdbaseDB+" db!");
                                }
                                alt_ids.remove(".");
                                Collections.sort(alt_ids);
                                String lowest_alt_id = alt_ids.get(0);
                                alt_locs_sql_str = "("+alt_loc_field+"="+alt_ids2keys.get(".")+" OR "+alt_loc_field+"="+alt_ids2keys.get(lowest_alt_id)+")";
                        } else {
                                alt_locs_sql_str=alt_loc_field+" IS NULL";
                        } 

                        rsst.close();
                        stmt.close();
                } catch (SQLException e) {
                        e.printStackTrace();
                }
                
                return alt_locs_sql_str;
        }
        
        public ArrayList<ArrayList> read_atomData() throws PdbaseInconsistencyError{
                ArrayList<ArrayList> resultset = new ArrayList<ArrayList>();
                String sql = "SELECT id, label_atom_id, label_comp_id, label_seq_id, Cartn_x, Cartn_y, Cartn_z " +
                                " FROM atom_site " +
                                " WHERE entry_key="+entrykey +
                                " AND label_asym_id='"+asymid+"' " +
                                " AND label_entity_key="+ entitykey +
                                " AND model_num="+ model +
                                " AND "+alt_locs_sql_str;
                try {
                        Statement stmt = conn.createStatement();
                        ResultSet rsst = stmt.executeQuery(sql);
                        int count=0;
                        while (rsst.next()){
                                count++;
                                ArrayList thisrecord = new ArrayList();
                                thisrecord.add(rsst.getInt(1)); //atomserial
                                thisrecord.add(rsst.getString(2).trim()); // atom
                                thisrecord.add(rsst.getString(3).trim()); // res_type
                                thisrecord.add(rsst.getInt(4)); //res_serial
                                thisrecord.add(rsst.getDouble(5)); // x
                                thisrecord.add(rsst.getDouble(6)); // y
                                thisrecord.add(rsst.getDouble(7)); // z
                                
                                resultset.add(thisrecord);
                        }
                        if (count==0){
                                throw new PdbaseInconsistencyError("atom data query returned no data at all for entry_key="+entrykey+", asym_id="+asymid+", entity_key="+entitykey+", model_num="+model+", alt_locs_sql_str='"+alt_locs_sql_str+"'");
                        }
                        rsst.close();
                        stmt.close();
                } catch (SQLException e) {
                        e.printStackTrace();
                }
                return resultset;
        }
        
        public String read_seq() throws PdbaseInconsistencyError{
                String sequence="";
                String pdbstrandid=chaincode;
                if (chaincode.equals("NULL")){
                        pdbstrandid="A";
                }
        // we use seq_id+0 (implicitly converts to int) in ORDER BY because seq_id is varchar!!
        String sql="SELECT mon_id" +
                        " FROM pdbx_poly_seq_scheme " +
                        " WHERE entry_key=" + entrykey +
                        " AND asym_id='"+asymid+"' " +
                        " AND pdb_strand_id='"+pdbstrandid+"' " +
                        " ORDER BY seq_id+0";
                try {
                        Statement stmt = conn.createStatement();
                        ResultSet rsst = stmt.executeQuery(sql);
                        ArrayList<String> aalist=AA.aas();
                        int count=0;
                        while (rsst.next()) {
                                count++;
                                String res_type = rsst.getString(1);
                                if (aalist.contains(res_type)){
                                        sequence+=AA.threeletter2oneletter(res_type);
                                } else {
                                        sequence+="X";
                                }
                        } 
                        if (count==0) {
                                System.err.println("No sequence data match for entry_key="+entrykey+", asym_id="+asymid+", pdb_strand_id="+pdbstrandid);
                                throw new PdbaseInconsistencyError("No sequence data match for entry_key="+entrykey+", asym_id="+asymid+", pdb_strand_id="+pdbstrandid);
                        }
                        rsst.close();
                        stmt.close();
                } catch (SQLException e) {
                        e.printStackTrace();
                }

                return sequence;
        }
}
Revision:	135
Committed:	Mon May 14 09:49:04 2007 UTC (17 years, 4 months ago) by duarte
File size:	9005 byte(s)
Log Message:	NEW FUNCTIONALITY: reading of graph from db is fully implemented for all cases New chain member variable in the Info classes, read in get_asym_id (Pdbase) and in get_chain_id (Msdsd) Reading chain also in Pdb in read_pdb_data_from_file Not reading chain anymore in read_atomData of PdbaseInfo Added oneletter2threeletter and getoneletter2threeletter to AA class Changes in Graph: - added db static vars and getUserName method - new member variables graphid and sm_id - new method read_graph_from_db to read contacts, nodes (and sequence from nodes) from db - new method getgraphid New Exception class GraphIdNotFoundError
Line	User	Rev	File contents
1	duarte	123	package proteinstructure;
2
3			import tools.MySQLConnection;
4	duarte	124
5	duarte	123	import java.sql.Statement;
6			import java.sql.ResultSet;
7			import java.sql.SQLException;
8			import java.util.ArrayList;
9			import java.util.HashMap;
10			import java.util.Collections;
11
12			public class PdbaseInfo {
13			public final static String MYSQLSERVER="white";
14	duarte	126	public final static String MYSQLUSER=getUserName();
15	duarte	123	public final static String MYSQLPWD="nieve";
16			public final static String mymsdsdDB="my_msdsd_00_07_a";
17			public final static String msdsdDB="msdsd_00_07_a";
18			public final static String pdbaseDB="pdbase";
19
20			MySQLConnection conn;
21			String accode="";
22			String chaincode="";
23	duarte	133	int model=DEFAULT_MODEL;
24	duarte	135	String chain=""; // the internal pdbase chain identifier (asym_id)
25	duarte	123	int entrykey;
26			String asymid;
27			int entitykey;
28			String alt_locs_sql_str;
29
30	duarte	133	static int DEFAULT_MODEL=1;
31
32	duarte	124	PdbaseInfo (String accode, String chaincode, int model_serial, String db) throws PdbaseInconsistencyError, PdbaseAcCodeNotFoundError{
33	duarte	123	this.accode=accode;
34			this.chaincode=chaincode;
35			this.model=model_serial;
36			this.conn = new MySQLConnection(MYSQLSERVER,MYSQLUSER,MYSQLPWD,db);
37			this.entrykey=get_entry_key();
38			this.asymid=get_asym_id();
39			this.entitykey=get_entity_key();
40			this.alt_locs_sql_str=get_atom_alt_locs();
41
42			}
43
44	duarte	124	PdbaseInfo (String accode, String chaincode, String db) throws PdbaseInconsistencyError, PdbaseAcCodeNotFoundError {
45	duarte	133	this(accode,chaincode,DEFAULT_MODEL,db);
46	duarte	123	}
47	duarte	124	PdbaseInfo (String accode, String chaincode, int model_serial) throws PdbaseInconsistencyError, PdbaseAcCodeNotFoundError {
48	duarte	123	this(accode,chaincode,model_serial,pdbaseDB);
49			}
50
51	duarte	124	PdbaseInfo (String accode, String chaincode) throws PdbaseInconsistencyError, PdbaseAcCodeNotFoundError {
52	duarte	133	this(accode,chaincode,DEFAULT_MODEL,pdbaseDB);
53	duarte	123	}
54	duarte	126
55			/** get user name from operating system (for use as database username) */
56			private static String getUserName() {
57			String user = null;
58			user = System.getProperty("user.name");
59			if(user == null) {
60			System.err.println("Could not get user name from operating system. Exiting");
61			System.exit(1);
62			}
63			return user;
64			}
65	duarte	123
66			public void close() {
67			conn.close();
68			}
69
70	duarte	124	public int get_entry_key() throws PdbaseAcCodeNotFoundError {
71	duarte	123	String sql="SELECT entry_key FROM struct WHERE entry_id='"+accode.toUpperCase()+"'";
72			try {
73			Statement stmt = conn.createStatement();
74			ResultSet rsst = stmt.executeQuery(sql);
75			if (rsst.next()) {
76			entrykey = rsst.getInt(1);
77			if (! rsst.isLast()) {
78			System.err.println("More than 1 entry_key match for accession_code="+accode+", chain_pdb_code="+chaincode);
79	duarte	124	throw new PdbaseAcCodeNotFoundError();
80	duarte	123	}
81			} else {
82			System.err.println("No entry_key match for accession_code="+accode+", chain_pdb_code="+chaincode);
83	duarte	124	throw new PdbaseAcCodeNotFoundError();
84	duarte	123	}
85			rsst.close();
86			stmt.close();
87			} catch (SQLException e) {
88			e.printStackTrace();
89			}
90			return entrykey;
91			}
92
93	duarte	124	public String get_asym_id() throws PdbaseInconsistencyError {
94	duarte	123	String pdbstrandid=chaincode;
95			if (chaincode.equals("NULL")){
96			pdbstrandid="A";
97			}
98			String sql="SELECT asym_id " +
99			" FROM pdbx_poly_seq_scheme " +
100			" WHERE entry_key=" + entrykey +
101			" AND pdb_strand_id='"+pdbstrandid+"' " +
102			" LIMIT 1";
103			try {
104			Statement stmt = conn.createStatement();
105			ResultSet rsst = stmt.executeQuery(sql);
106			if (rsst.next()) {
107			asymid = rsst.getString(1);
108			} else {
109			System.err.println("No asym_id match for entry_key="+entrykey+", pdb_strand_id="+chaincode);
110	duarte	124	throw new PdbaseInconsistencyError("No asym_id match for entry_key="+entrykey+", pdb_strand_id="+chaincode);
111	duarte	123	}
112			rsst.close();
113			stmt.close();
114			} catch (SQLException e) {
115			e.printStackTrace();
116			}
117	duarte	135	// we set the internal chain identifier self.chain from asymid
118			chain = asymid;
119	duarte	123	return asymid;
120			}
121
122	duarte	124	public int get_entity_key() throws PdbaseInconsistencyError {
123	duarte	123	String sql="SELECT entity_key " +
124			" FROM struct_asym " +
125			" WHERE entry_key="+ entrykey +
126			" AND id='"+asymid+"'";
127			try {
128			Statement stmt = conn.createStatement();
129			ResultSet rsst = stmt.executeQuery(sql);
130			if (rsst.next()) {
131			entitykey = rsst.getInt(1);
132			if (! rsst.isLast()) {
133			System.err.println("More than 1 entity_key match for entry_key="+entrykey+", asym_id="+asymid);
134	duarte	124	throw new PdbaseInconsistencyError("More than 1 entity_key match for entry_key="+entrykey+", asym_id="+asymid);
135	duarte	123	}
136			} else {
137			System.err.println("No entity_key match for entry_key="+entrykey+", asym_id="+asymid);
138	duarte	124	throw new PdbaseInconsistencyError("No entity_key match for entry_key="+entrykey+", asym_id="+asymid);
139	duarte	123	}
140			rsst.close();
141			stmt.close();
142			} catch (SQLException e) {
143			e.printStackTrace();
144			}
145			return entitykey;
146			}
147
148	duarte	124	public String get_atom_alt_locs() throws PdbaseInconsistencyError{
149	duarte	123	ArrayList<String> alt_ids = new ArrayList<String>();
150			HashMap<String,Integer> alt_ids2keys = new HashMap<String,Integer>();
151			String alt_loc_field="label_alt_key";
152			String sql="SELECT id, atom_sites_alt_key FROM atom_sites_alt WHERE entry_key="+entrykey;
153			try {
154			Statement stmt = conn.createStatement();
155			ResultSet rsst = stmt.executeQuery(sql);
156			int count=0;
157			while (rsst.next()) {
158			count++;
159			alt_ids.add(rsst.getString(1));
160			alt_ids2keys.put(rsst.getString(1), rsst.getInt(2));
161			}
162			if (count!=0){
163			if ((! alt_ids.contains(".")) \|\| alt_ids.indexOf(".")!=alt_ids.lastIndexOf(".")){ // second term is a way of finding out if there is more than 1 ocurrence of "." in the ArrayList
164			System.err.println("alt_codes exist for entry_key "+entrykey+" but there is either no default value '.' or more than 1 '.'. Something wrong with this entry_key or with "+pdbaseDB+" db!");
165	duarte	124	throw new PdbaseInconsistencyError("alt_codes exist for entry_key "+entrykey+" but there is either no default value '.' or more than 1 '.'. Something wrong with this entry_key or with "+pdbaseDB+" db!");
166	duarte	123	}
167			alt_ids.remove(".");
168			Collections.sort(alt_ids);
169			String lowest_alt_id = alt_ids.get(0);
170			alt_locs_sql_str = "("+alt_loc_field+"="+alt_ids2keys.get(".")+" OR "+alt_loc_field+"="+alt_ids2keys.get(lowest_alt_id)+")";
171			} else {
172			alt_locs_sql_str=alt_loc_field+" IS NULL";
173			}
174
175			rsst.close();
176			stmt.close();
177			} catch (SQLException e) {
178			e.printStackTrace();
179			}
180
181			return alt_locs_sql_str;
182			}
183
184	duarte	124	public ArrayList<ArrayList> read_atomData() throws PdbaseInconsistencyError{
185	duarte	123	ArrayList<ArrayList> resultset = new ArrayList<ArrayList>();
186	duarte	135	String sql = "SELECT id, label_atom_id, label_comp_id, label_seq_id, Cartn_x, Cartn_y, Cartn_z " +
187	duarte	123	" FROM atom_site " +
188			" WHERE entry_key="+entrykey +
189			" AND label_asym_id='"+asymid+"' " +
190			" AND label_entity_key="+ entitykey +
191			" AND model_num="+ model +
192			" AND "+alt_locs_sql_str;
193			try {
194			Statement stmt = conn.createStatement();
195			ResultSet rsst = stmt.executeQuery(sql);
196	duarte	124	int count=0;
197	duarte	123	while (rsst.next()){
198	duarte	124	count++;
199	duarte	123	ArrayList thisrecord = new ArrayList();
200			thisrecord.add(rsst.getInt(1)); //atomserial
201			thisrecord.add(rsst.getString(2).trim()); // atom
202			thisrecord.add(rsst.getString(3).trim()); // res_type
203	duarte	135	thisrecord.add(rsst.getInt(4)); //res_serial
204			thisrecord.add(rsst.getDouble(5)); // x
205			thisrecord.add(rsst.getDouble(6)); // y
206			thisrecord.add(rsst.getDouble(7)); // z
207	duarte	123
208			resultset.add(thisrecord);
209			}
210	duarte	124	if (count==0){
211			throw new PdbaseInconsistencyError("atom data query returned no data at all for entry_key="+entrykey+", asym_id="+asymid+", entity_key="+entitykey+", model_num="+model+", alt_locs_sql_str='"+alt_locs_sql_str+"'");
212			}
213	duarte	123	rsst.close();
214			stmt.close();
215			} catch (SQLException e) {
216			e.printStackTrace();
217			}
218			return resultset;
219			}
220
221	duarte	124	public String read_seq() throws PdbaseInconsistencyError{
222	duarte	123	String sequence="";
223			String pdbstrandid=chaincode;
224			if (chaincode.equals("NULL")){
225			pdbstrandid="A";
226			}
227			// we use seq_id+0 (implicitly converts to int) in ORDER BY because seq_id is varchar!!
228			String sql="SELECT mon_id" +
229			" FROM pdbx_poly_seq_scheme " +
230			" WHERE entry_key=" + entrykey +
231			" AND asym_id='"+asymid+"' " +
232			" AND pdb_strand_id='"+pdbstrandid+"' " +
233			" ORDER BY seq_id+0";
234			try {
235			Statement stmt = conn.createStatement();
236			ResultSet rsst = stmt.executeQuery(sql);
237			ArrayList<String> aalist=AA.aas();
238			int count=0;
239			while (rsst.next()) {
240			count++;
241			String res_type = rsst.getString(1);
242			if (aalist.contains(res_type)){
243			sequence+=AA.threeletter2oneletter(res_type);
244			} else {
245			sequence+="X";
246			}
247			}
248			if (count==0) {
249			System.err.println("No sequence data match for entry_key="+entrykey+", asym_id="+asymid+", pdb_strand_id="+pdbstrandid);
250	duarte	124	throw new PdbaseInconsistencyError("No sequence data match for entry_key="+entrykey+", asym_id="+asymid+", pdb_strand_id="+pdbstrandid);
251	duarte	123	}
252			rsst.close();
253			stmt.close();
254			} catch (SQLException e) {
255			e.printStackTrace();
256			}
257
258			return sequence;
259			}
260			}