owl/trunk/genGraph.java

import gnu.getopt.Getopt;

import java.io.File;
import java.io.IOException;
import java.sql.SQLException;

import proteinstructure.Pdb;
import proteinstructure.PdbCodeNotFoundError;
import proteinstructure.PdbLoadError;
import proteinstructure.PdbasePdb;
import proteinstructure.PdbfilePdb;
import proteinstructure.RIGraph;
import proteinstructure.TemplateList;
import tools.MySQLConnection;


public class genGraph {
        /*------------------------------ constants ------------------------------*/
        
        public static final String                      PDB_DB = "pdbase";
        public static final String                      DB_HOST = "white";                                                              
        public static final String                      DB_USER = getUserName();
        public static final String                      DB_PWD = "nieve";
        
        /*---------------------------- private methods --------------------------*/
        /** 
         * 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.");
                }
                return user;
        }
        
        public static void main(String[] args) throws IOException {
                
                String progName = "genGraph";
                
                String help = "Usage, 3 options:\n" +
                                "1)  "+progName+" -i <listfile> -d <distance_cutoff> -t <contact_type> [-o <output_dir>] [-D <pdbase_db>] [-P] \n" +
                                "2)  "+progName+" -p <pdb_code> -d <distance_cutoff> -t <contact_type> [-o <output_dir>] [-D <pdbase_db>] [-P] \n" +
                                "3)  "+progName+" -f <pdbfile> [-c <chain_pdb_code>] -d <distance_cutoff> -t <contact_type> [-o <output_dir>] [-P] \n\n" +
                                "In case 2) also a list of comma separated pdb codes+chain codes can be specified, e.g. -p 1bxyA,1josA \n" +
                                "In case 3) if -c not specified then the first chain code in the pdb file will be taken\n" +
                                "If pdbase_db not specified, the default pdbase will be used\n" +
                                "If output dir not specified default is current\n" +
                                "In all cases option -P can be specified to get the graph in paul format instead of aglappe format\n"; 

                String listfile = "";
                String[] pdbIds = null;
                String pdbChainCode4file = null;
                String pdbfile = "";
                String pdbaseDb = PDB_DB;
                String edgeType = "";
                double cutoff = 0.0;
                String outputDir = "."; // we set default to current directory
                boolean paul = false;
                
                Getopt g = new Getopt(progName, args, "i:p:c:f:d:t:o:D:Ph?");
                int c;
                while ((c = g.getopt()) != -1) {
                        switch(c){
                        case 'i':
                                listfile = g.getOptarg();
                                break;
                        case 'p':
                                pdbIds = g.getOptarg().split(",");
                                break;
                        case 'c':
                                pdbChainCode4file = g.getOptarg();
                                break;
                        case 'f':
                                pdbfile = g.getOptarg();
                                break;
                        case 'd':
                                cutoff = Double.valueOf(g.getOptarg());
                                break;
                        case 't':
                                edgeType = g.getOptarg();
                                break;
                        case 'o':
                                outputDir = g.getOptarg();
                                break;
                        case 'D':
                                pdbaseDb = g.getOptarg();
                                break;
                        case 'P':
                                paul = true;
                                break;                          
                        case 'h':
                        case '?':
                                System.out.println(help);
                                System.exit(0);
                                break; // getopt() already printed an error
                        }
                }

                if (edgeType.equals("") || cutoff==0.0) {
                        System.err.println("Some missing option");
                        System.err.println(help);
                        System.exit(1);
                }
                if (listfile.equals("") && pdbIds==null && pdbfile.equals("")){
                        System.err.println("Either a listfile, some pdb codes/chain codes or a pdbfile must be given");
                        System.err.println(help);
                        System.exit(1);
                }
                if ((!listfile.equals("") && pdbIds!=null) || (!listfile.equals("") && !pdbfile.equals("")) || (pdbIds!=null && !pdbfile.equals(""))) {
                        System.err.println("Options -p, -i and -f/-c are exclusive. Use only one of them");
                        System.err.println(help);
                        System.exit(1);                 
                }

                
                MySQLConnection conn = null;            

                try{
                        conn = new MySQLConnection(DB_HOST, DB_USER, DB_PWD);
                } catch (Exception e) {
                        System.err.println("Error opening database connection. Exiting");
                        System.exit(1);
                }
                
                
                if (pdbfile.equals("")){
                        
                        if (!listfile.equals("")) {             
                                pdbIds = TemplateList.readIdsListFile(new File(listfile));
                        }

                        int numPdbs = 0;

                        for (int i=0;i<pdbIds.length;i++) {
                                String pdbCode = pdbIds[i].substring(0,4);
                                String pdbChainCode = pdbIds[i].substring(4);

                                try {
                                        
                                        //long start = System.currentTimeMillis();
                                        
                                        Pdb pdb = new PdbasePdb(pdbCode, pdbaseDb, conn);
                                        pdb.load(pdbChainCode);

                                        // get graph
                                        RIGraph graph = pdb.get_graph(edgeType, cutoff);
                                        
                                        String edgeTypeStr = edgeType.replaceAll("/", ":");
                                        
                                        File outputFile = new File(outputDir,pdbCode+pdbChainCode+"_"+edgeTypeStr+"_"+cutoff+".cm");
                                        if (!paul) {
                                                graph.write_graph_to_file(outputFile.getAbsolutePath());
                                        } else {
                                                graph.writeToPaulFile(outputFile.getAbsolutePath());
                                        }

                                        //long end = System.currentTimeMillis();
                                        //double time = (double) (end -start)/1000;

                                        System.out.println("Wrote "+outputFile.getAbsolutePath());
                                        //System.out.printf("%5.3f s\n",time);
                                        
                                        numPdbs++;

                                } catch (PdbLoadError e) {
                                        System.err.println("Error loading pdb data for " + pdbCode + pdbChainCode+", specific error: "+e.getMessage());
                                } catch (PdbCodeNotFoundError e) {
                                        System.err.println("Couldn't find pdb code "+pdbCode);
                                } catch (SQLException e) {
                                        System.err.println("SQL error for structure "+pdbCode+pdbChainCode+", error: "+e.getMessage());
                                }

                        }

                        // output results
                        System.out.println("Number of structures done successfully: " + numPdbs);


                } else {
                        File pdbFile = new File(pdbfile);
                        try {
                                Pdb pdb = new PdbfilePdb(pdbfile);
                                if (pdbChainCode4file==null) {
                                        pdbChainCode4file = pdb.getChains()[0];
                                }

                                pdb.load(pdbChainCode4file);
                                RIGraph graph = pdb.get_graph(edgeType, cutoff);

                                String edgeTypeStr = edgeType.replaceAll("/", ":");
                                
                                String filename = pdbFile.getName().substring(0, pdbFile.getName().lastIndexOf("."));
                                File outputFile = new File(outputDir,filename+"_"+edgeTypeStr+"_"+cutoff+".cm");
                                if (!paul) {
                                        graph.write_graph_to_file(outputFile.getAbsolutePath());
                                } else {
                                        graph.writeToPaulFile(outputFile.getAbsolutePath());
                                }
                                System.out.println("Wrote graph file "+outputFile.getAbsolutePath()+" from pdb file "+pdbFile);
                                
                        } catch (PdbLoadError e) {
                                System.err.println("Error loading from pdb file "+pdbFile+", specific error: "+e.getMessage());
                        }
                }
        }

}
Revision:	609
Committed:	Wed Apr 30 14:11:50 2008 UTC (16 years, 5 months ago) by duarte
File size:	6429 byte(s)
Log Message:	Extracted constant NULL_CHAIN_CODE that was still hard-coded in MANY places. Now should be safe to replace the value of the constant from "NULL" to something else. Using TemplateList.readIdsListFile() in averageGraph, dumpseq and genGraph so that is now a standard way of reading list files with pdbcodes/chaincodes. NOTE: this introduces an incompatibility with legacy list files: we can't read a NULL pdb chain code with the "NULL" string e.g. '1i1b NULL'. This is fine because since PDB-REMEDIATED all NULL chains are now "A". Fixed a few things and made more standard the scripts dumpseq and genGraph
Line	User	Rev	File contents
1	duarte	277	import gnu.getopt.Getopt;
2
3			import java.io.File;
4			import java.io.IOException;
5			import java.sql.SQLException;
6
7			import proteinstructure.Pdb;
8			import proteinstructure.PdbCodeNotFoundError;
9	duarte	441	import proteinstructure.PdbLoadError;
10	duarte	277	import proteinstructure.PdbasePdb;
11			import proteinstructure.PdbfilePdb;
12	duarte	420	import proteinstructure.RIGraph;
13	duarte	609	import proteinstructure.TemplateList;
14	duarte	277	import tools.MySQLConnection;
15
16
17			public class genGraph {
18			/------------------------------ constants ------------------------------/
19
20			public static final String PDB_DB = "pdbase";
21			public static final String DB_HOST = "white";
22			public static final String DB_USER = getUserName();
23			public static final String DB_PWD = "nieve";
24
25			/---------------------------- private methods --------------------------/
26			/**
27			* Get user name from operating system (for use as database username).
28			* */
29			private static String getUserName() {
30			String user = null;
31			user = System.getProperty("user.name");
32			if(user == null) {
33			System.err.println("Could not get user name from operating system.");
34			}
35			return user;
36			}
37
38			public static void main(String[] args) throws IOException {
39
40	duarte	599	String progName = "genGraph";
41	duarte	277
42			String help = "Usage, 3 options:\n" +
43	duarte	609	"1) "+progName+" -i <listfile> -d <distance_cutoff> -t <contact_type> [-o <output_dir>] [-D <pdbase_db>] [-P] \n" +
44			"2) "+progName+" -p <pdb_code> -d <distance_cutoff> -t <contact_type> [-o <output_dir>] [-D <pdbase_db>] [-P] \n" +
45			"3) "+progName+" -f <pdbfile> [-c <chain_pdb_code>] -d <distance_cutoff> -t <contact_type> [-o <output_dir>] [-P] \n\n" +
46			"In case 2) also a list of comma separated pdb codes+chain codes can be specified, e.g. -p 1bxyA,1josA \n" +
47			"In case 3) if -c not specified then the first chain code in the pdb file will be taken\n" +
48	duarte	599	"If pdbase_db not specified, the default pdbase will be used\n" +
49	duarte	609	"If output dir not specified default is current\n" +
50	duarte	600	"In all cases option -P can be specified to get the graph in paul format instead of aglappe format\n";
51	duarte	277
52			String listfile = "";
53	duarte	609	String[] pdbIds = null;
54			String pdbChainCode4file = null;
55	duarte	277	String pdbfile = "";
56			String pdbaseDb = PDB_DB;
57			String edgeType = "";
58			double cutoff = 0.0;
59	duarte	609	String outputDir = "."; // we set default to current directory
60	duarte	599	boolean paul = false;
61	duarte	277
62	duarte	599	Getopt g = new Getopt(progName, args, "i:p:c:f:d:t:o:D:Ph?");
63	duarte	277	int c;
64			while ((c = g.getopt()) != -1) {
65			switch(c){
66			case 'i':
67			listfile = g.getOptarg();
68			break;
69			case 'p':
70	duarte	609	pdbIds = g.getOptarg().split(",");
71	duarte	277	break;
72			case 'c':
73	duarte	609	pdbChainCode4file = g.getOptarg();
74	duarte	277	break;
75			case 'f':
76			pdbfile = g.getOptarg();
77			break;
78			case 'd':
79			cutoff = Double.valueOf(g.getOptarg());
80			break;
81			case 't':
82			edgeType = g.getOptarg();
83			break;
84			case 'o':
85			outputDir = g.getOptarg();
86			break;
87			case 'D':
88			pdbaseDb = g.getOptarg();
89			break;
90	duarte	599	case 'P':
91			paul = true;
92			break;
93	duarte	277	case 'h':
94			case '?':
95			System.out.println(help);
96			System.exit(0);
97			break; // getopt() already printed an error
98			}
99			}
100
101	duarte	609	if (edgeType.equals("") \|\| cutoff==0.0) {
102	duarte	277	System.err.println("Some missing option");
103			System.err.println(help);
104			System.exit(1);
105			}
106	duarte	609	if (listfile.equals("") && pdbIds==null && pdbfile.equals("")){
107	duarte	277	System.err.println("Either a listfile, some pdb codes/chain codes or a pdbfile must be given");
108			System.err.println(help);
109			System.exit(1);
110			}
111	duarte	609	if ((!listfile.equals("") && pdbIds!=null) \|\| (!listfile.equals("") && !pdbfile.equals("")) \|\| (pdbIds!=null && !pdbfile.equals(""))) {
112			System.err.println("Options -p, -i and -f/-c are exclusive. Use only one of them");
113	duarte	277	System.err.println(help);
114			System.exit(1);
115			}
116
117
118			MySQLConnection conn = null;
119
120			try{
121			conn = new MySQLConnection(DB_HOST, DB_USER, DB_PWD);
122			} catch (Exception e) {
123			System.err.println("Error opening database connection. Exiting");
124			System.exit(1);
125			}
126
127
128			if (pdbfile.equals("")){
129
130	duarte	609	if (!listfile.equals("")) {
131			pdbIds = TemplateList.readIdsListFile(new File(listfile));
132	duarte	277	}
133
134			int numPdbs = 0;
135
136	duarte	609	for (int i=0;i<pdbIds.length;i++) {
137			String pdbCode = pdbIds[i].substring(0,4);
138			String pdbChainCode = pdbIds[i].substring(4);
139	duarte	277
140			try {
141	duarte	285
142	duarte	600	//long start = System.currentTimeMillis();
143	duarte	285
144	duarte	441	Pdb pdb = new PdbasePdb(pdbCode, pdbaseDb, conn);
145			pdb.load(pdbChainCode);
146	duarte	277
147			// get graph
148	duarte	420	RIGraph graph = pdb.get_graph(edgeType, cutoff);
149	duarte	431
150			String edgeTypeStr = edgeType.replaceAll("/", ":");
151
152	duarte	609	File outputFile = new File(outputDir,pdbCode+pdbChainCode+"_"+edgeTypeStr+"_"+cutoff+".cm");
153	duarte	599	if (!paul) {
154			graph.write_graph_to_file(outputFile.getAbsolutePath());
155			} else {
156			graph.writeToPaulFile(outputFile.getAbsolutePath());
157			}
158	duarte	277
159	duarte	600	//long end = System.currentTimeMillis();
160			//double time = (double) (end -start)/1000;
161	duarte	285
162	duarte	277	System.out.println("Wrote "+outputFile.getAbsolutePath());
163	duarte	600	//System.out.printf("%5.3f s\n",time);
164	duarte	285
165	duarte	277	numPdbs++;
166
167	duarte	441	} catch (PdbLoadError e) {
168			System.err.println("Error loading pdb data for " + pdbCode + pdbChainCode+", specific error: "+e.getMessage());
169	duarte	277	} catch (PdbCodeNotFoundError e) {
170			System.err.println("Couldn't find pdb code "+pdbCode);
171			} catch (SQLException e) {
172	duarte	609	System.err.println("SQL error for structure "+pdbCode+pdbChainCode+", error: "+e.getMessage());
173	duarte	277	}
174
175			}
176
177			// output results
178			System.out.println("Number of structures done successfully: " + numPdbs);
179
180
181			} else {
182	duarte	609	File pdbFile = new File(pdbfile);
183	duarte	277	try {
184	duarte	441	Pdb pdb = new PdbfilePdb(pdbfile);
185	duarte	609	if (pdbChainCode4file==null) {
186			pdbChainCode4file = pdb.getChains()[0];
187	duarte	277	}
188	duarte	609
189			pdb.load(pdbChainCode4file);
190	duarte	420	RIGraph graph = pdb.get_graph(edgeType, cutoff);
191	duarte	277
192	duarte	431	String edgeTypeStr = edgeType.replaceAll("/", ":");
193
194	duarte	609	String filename = pdbFile.getName().substring(0, pdbFile.getName().lastIndexOf("."));
195			File outputFile = new File(outputDir,filename+"_"+edgeTypeStr+"_"+cutoff+".cm");
196	duarte	599	if (!paul) {
197			graph.write_graph_to_file(outputFile.getAbsolutePath());
198			} else {
199			graph.writeToPaulFile(outputFile.getAbsolutePath());
200			}
201	duarte	609	System.out.println("Wrote graph file "+outputFile.getAbsolutePath()+" from pdb file "+pdbFile);
202	duarte	277
203	duarte	441	} catch (PdbLoadError e) {
204	duarte	609	System.err.println("Error loading from pdb file "+pdbFile+", specific error: "+e.getMessage());
205	duarte	277	}
206			}
207			}
208
209			}