|
a |
|
b/src/utils/utils.java |
|
|
1 |
/* |
|
|
2 |
* To change this license header, choose License Headers in Project Properties. |
|
|
3 |
* To change this template file, choose Tools | Templates |
|
|
4 |
* and open the template in the editor. |
|
|
5 |
*/ |
|
|
6 |
package utils; |
|
|
7 |
|
|
|
8 |
import biodiscml.Main; |
|
|
9 |
import java.io.BufferedReader; |
|
|
10 |
import java.io.File; |
|
|
11 |
import java.io.FileInputStream; |
|
|
12 |
import java.io.FileOutputStream; |
|
|
13 |
import java.io.FileReader; |
|
|
14 |
import java.io.IOException; |
|
|
15 |
import java.io.InputStream; |
|
|
16 |
import java.io.OutputStream; |
|
|
17 |
import java.nio.channels.FileChannel; |
|
|
18 |
import java.text.DecimalFormat; |
|
|
19 |
import java.text.DecimalFormatSymbols; |
|
|
20 |
import java.util.ArrayList; |
|
|
21 |
import java.util.HashMap; |
|
|
22 |
import java.util.Iterator; |
|
|
23 |
import java.util.TreeMap; |
|
|
24 |
import org.apache.commons.math3.stat.descriptive.moment.Mean; |
|
|
25 |
import org.apache.commons.math3.stat.descriptive.moment.StandardDeviation; |
|
|
26 |
|
|
|
27 |
/** |
|
|
28 |
* |
|
|
29 |
* @author Administrator |
|
|
30 |
*/ |
|
|
31 |
public class utils { |
|
|
32 |
|
|
|
33 |
public static HashMap<String, ArrayList<String>> convertDataColumnsToHashMap(ArrayList<String[]> altable) { |
|
|
34 |
HashMap<String, ArrayList<String>> hmData = new HashMap<>(); |
|
|
35 |
for (int i = 1; i < altable.get(0).length; i++) { //for column |
|
|
36 |
String header = altable.get(0)[i]; //get column header |
|
|
37 |
ArrayList<String> al = new ArrayList<>(); //data list of the column |
|
|
38 |
for (int j = 1; j < altable.size(); j++) { //for each value in the column |
|
|
39 |
try { |
|
|
40 |
al.add(altable.get(j)[i]); |
|
|
41 |
} catch (Exception e) { |
|
|
42 |
al.add(""); |
|
|
43 |
} |
|
|
44 |
} |
|
|
45 |
hmData.put(header, al); |
|
|
46 |
} |
|
|
47 |
return hmData; |
|
|
48 |
} |
|
|
49 |
|
|
|
50 |
public static ArrayList<String[]> readTable(String file, String separator) { |
|
|
51 |
if (separator.isEmpty()) { |
|
|
52 |
separator = detectSeparator(file); |
|
|
53 |
} |
|
|
54 |
ArrayList<String[]> altable = new ArrayList<>(); |
|
|
55 |
int cpt = 1; |
|
|
56 |
try { |
|
|
57 |
BufferedReader br = new BufferedReader(new FileReader(file)); |
|
|
58 |
while (br.ready()) { |
|
|
59 |
if (Main.debug2) { |
|
|
60 |
System.out.println("Reading line " + cpt++); |
|
|
61 |
} |
|
|
62 |
altable.add(br.readLine().split(separator)); |
|
|
63 |
} |
|
|
64 |
br.close(); |
|
|
65 |
} catch (Exception e) { |
|
|
66 |
e.printStackTrace(); |
|
|
67 |
} |
|
|
68 |
return altable; |
|
|
69 |
} |
|
|
70 |
|
|
|
71 |
public static ArrayList<String[]> readTable(String file) { |
|
|
72 |
String separator = detectSeparator(file); |
|
|
73 |
if (Main.debug) { |
|
|
74 |
System.out.println("reading table " + file); |
|
|
75 |
} |
|
|
76 |
ArrayList<String[]> altable = new ArrayList<>(); |
|
|
77 |
try { |
|
|
78 |
BufferedReader br = new BufferedReader(new FileReader(file)); |
|
|
79 |
while (br.ready()) { |
|
|
80 |
altable.add(br.readLine().split(separator)); |
|
|
81 |
} |
|
|
82 |
} catch (Exception e) { |
|
|
83 |
e.printStackTrace(); |
|
|
84 |
} |
|
|
85 |
return altable; |
|
|
86 |
} |
|
|
87 |
|
|
|
88 |
public static String[][] transposeMatrix(String[][] m) { |
|
|
89 |
String[][] temp = new String[m[0].length][m.length]; |
|
|
90 |
for (int i = 0; i < m.length; i++) { |
|
|
91 |
for (int j = 0; j < m[0].length; j++) { |
|
|
92 |
temp[j][i] = m[i][j]; |
|
|
93 |
} |
|
|
94 |
} |
|
|
95 |
return temp; |
|
|
96 |
} |
|
|
97 |
|
|
|
98 |
/** |
|
|
99 |
* |
|
|
100 |
* @param doubles |
|
|
101 |
* @return |
|
|
102 |
*/ |
|
|
103 |
public static double[] convertStringListToDoubles(ArrayList<String> doubles) { |
|
|
104 |
double[] ret = new double[doubles.size()]; |
|
|
105 |
Iterator<String> iterator = doubles.iterator(); |
|
|
106 |
int i = 0; |
|
|
107 |
while (iterator.hasNext()) { |
|
|
108 |
ret[i] = Double.valueOf(iterator.next()); |
|
|
109 |
i++; |
|
|
110 |
} |
|
|
111 |
return ret; |
|
|
112 |
} |
|
|
113 |
|
|
|
114 |
/** |
|
|
115 |
* auto detection of the delimiter |
|
|
116 |
* |
|
|
117 |
* @param infile |
|
|
118 |
* @return |
|
|
119 |
*/ |
|
|
120 |
public static String detectSeparator(String infile) { |
|
|
121 |
if (Main.debug) { |
|
|
122 |
//System.out.println("Delimiter not specified, auto-detection 10 first lines among these delimiters: \"\\t\" \" \" \";\" \",\" \"~\" \":\" \"/\" \"\\|\""); |
|
|
123 |
System.out.print("Delimiter not specified, auto-detection in the 10 first lines..."); |
|
|
124 |
} |
|
|
125 |
String delimiter = ""; |
|
|
126 |
String potentialDelimiters[] = {"\\t", " ", ";", ",", "~", ":", "/", "\\|"}; |
|
|
127 |
for (String potentialDelimiter : potentialDelimiters) { |
|
|
128 |
try { |
|
|
129 |
BufferedReader br = new BufferedReader(new FileReader(infile)); |
|
|
130 |
boolean sameNumberAsPreviousLine = true; |
|
|
131 |
String line = br.readLine(); |
|
|
132 |
int init = line.split(potentialDelimiter).length; |
|
|
133 |
if (init > 1) { |
|
|
134 |
int cpt = 0; |
|
|
135 |
while (br.ready() & sameNumberAsPreviousLine && cpt < 10) { |
|
|
136 |
cpt++; |
|
|
137 |
line = br.readLine(); |
|
|
138 |
if (!line.trim().isEmpty()) { |
|
|
139 |
int split = line.split(potentialDelimiter).length; |
|
|
140 |
if (split == init) { |
|
|
141 |
init = split; |
|
|
142 |
} else { |
|
|
143 |
sameNumberAsPreviousLine = false; |
|
|
144 |
} |
|
|
145 |
} |
|
|
146 |
} |
|
|
147 |
if (sameNumberAsPreviousLine) { |
|
|
148 |
delimiter = potentialDelimiter; |
|
|
149 |
break; |
|
|
150 |
} |
|
|
151 |
} |
|
|
152 |
} catch (Exception e) { |
|
|
153 |
e.printStackTrace(); |
|
|
154 |
} |
|
|
155 |
} |
|
|
156 |
if (delimiter.isEmpty()) { |
|
|
157 |
System.err.print("[error] CSV separator not detected. Guessing there is just one column. " |
|
|
158 |
+ "If not, change your CSV separator to a standard one (ex: tabulation or comma), or check the file consistency"); |
|
|
159 |
delimiter = " "; |
|
|
160 |
} |
|
|
161 |
if (Main.debug) { |
|
|
162 |
System.out.println("Delimiter found:" + delimiter); |
|
|
163 |
} |
|
|
164 |
return delimiter; |
|
|
165 |
} |
|
|
166 |
|
|
|
167 |
public static void copyFileUsingStream(File source, File dest) throws IOException { |
|
|
168 |
InputStream is = null; |
|
|
169 |
OutputStream os = null; |
|
|
170 |
try { |
|
|
171 |
is = new FileInputStream(source); |
|
|
172 |
os = new FileOutputStream(dest); |
|
|
173 |
byte[] buffer = new byte[1024]; |
|
|
174 |
int length; |
|
|
175 |
while ((length = is.read(buffer)) > 0) { |
|
|
176 |
os.write(buffer, 0, length); |
|
|
177 |
} |
|
|
178 |
} finally { |
|
|
179 |
is.close(); |
|
|
180 |
os.close(); |
|
|
181 |
} |
|
|
182 |
} |
|
|
183 |
|
|
|
184 |
public static void copyFileUsingChannel(File source, File dest) throws IOException { |
|
|
185 |
FileChannel sourceChannel = null; |
|
|
186 |
FileChannel destChannel = null; |
|
|
187 |
try { |
|
|
188 |
sourceChannel = new FileInputStream(source).getChannel(); |
|
|
189 |
destChannel = new FileOutputStream(dest).getChannel(); |
|
|
190 |
destChannel.transferFrom(sourceChannel, 0, sourceChannel.size()); |
|
|
191 |
} finally { |
|
|
192 |
sourceChannel.close(); |
|
|
193 |
destChannel.close(); |
|
|
194 |
} |
|
|
195 |
} |
|
|
196 |
|
|
|
197 |
public static String arrayToString(String[] array, String separator) { |
|
|
198 |
String s = ""; |
|
|
199 |
for (int i = 0; i < array.length; i++) { |
|
|
200 |
if (array[i] != null) { |
|
|
201 |
s += array[i] + separator; |
|
|
202 |
} |
|
|
203 |
} |
|
|
204 |
return s.substring(0, s.length() - 1); |
|
|
205 |
} |
|
|
206 |
|
|
|
207 |
public static String arrayToString(ArrayList<Integer> array, String separator) { |
|
|
208 |
String s = ""; |
|
|
209 |
for (int i = 0; i < array.size(); i++) { |
|
|
210 |
s += array.get(i) + separator; |
|
|
211 |
} |
|
|
212 |
return s.substring(0, s.length() - 1); |
|
|
213 |
} |
|
|
214 |
|
|
|
215 |
public static double[] arrayToDouble(ArrayList<Double> al) { |
|
|
216 |
return al.stream().mapToDouble(Double::doubleValue).toArray(); |
|
|
217 |
} |
|
|
218 |
|
|
|
219 |
public static String getStandardDeviation(ArrayList<Double> al) { |
|
|
220 |
DecimalFormat df = new DecimalFormat(); |
|
|
221 |
df.setMaximumFractionDigits(3); |
|
|
222 |
DecimalFormatSymbols dfs = new DecimalFormatSymbols(); |
|
|
223 |
dfs.setDecimalSeparator('.'); |
|
|
224 |
df.setDecimalFormatSymbols(dfs); |
|
|
225 |
|
|
|
226 |
StandardDeviation s = new StandardDeviation(); |
|
|
227 |
s.setData(utils.arrayToDouble(al)); |
|
|
228 |
|
|
|
229 |
return df.format(s.evaluate()); |
|
|
230 |
} |
|
|
231 |
|
|
|
232 |
/** |
|
|
233 |
* Table Object to manipulate table |
|
|
234 |
*/ |
|
|
235 |
public static class TableObject { |
|
|
236 |
|
|
|
237 |
public HashMap<String, ArrayList<String>> hmData = new HashMap<>();//<ColName,Values> |
|
|
238 |
public HashMap<String, Integer> hmIDsList = new HashMap<>(); //<ID,index> |
|
|
239 |
|
|
|
240 |
public TableObject(ArrayList<String[]> altable) { |
|
|
241 |
if (Main.debug) { |
|
|
242 |
System.out.println("converting table"); |
|
|
243 |
} |
|
|
244 |
hmData = convertDataColumnsToHashMap(altable); |
|
|
245 |
int ID_index = getIdIndex(altable); |
|
|
246 |
for (int i = 1; i < altable.size(); i++) { //for each ID |
|
|
247 |
//hmIDsList.put(altable.get(i)[ID_index].toLowerCase(), i - 1); |
|
|
248 |
Integer j = hmIDsList.put(altable.get(i)[ID_index], i - 1); |
|
|
249 |
if (j != null) { |
|
|
250 |
System.err.println("DUPLICATED ID DETECTED: " + altable.get(i)[ID_index]); |
|
|
251 |
System.err.println("All IDs must be unique"); |
|
|
252 |
System.exit(0); |
|
|
253 |
} |
|
|
254 |
} |
|
|
255 |
} |
|
|
256 |
|
|
|
257 |
public ArrayList<String> getTheClass(String theclass) { |
|
|
258 |
ArrayList<String> toreturn = hmData.get(theclass); |
|
|
259 |
hmData.remove(theclass); |
|
|
260 |
return toreturn; |
|
|
261 |
} |
|
|
262 |
|
|
|
263 |
public boolean containsClass(String theClass) { |
|
|
264 |
return hmData.containsKey(theClass); |
|
|
265 |
} |
|
|
266 |
|
|
|
267 |
public int getIdIndex(ArrayList<String[]> altable) { |
|
|
268 |
int cpt = 0; |
|
|
269 |
for (String s : altable.get(0)) { |
|
|
270 |
if (s.equals(Main.mergingID)) { |
|
|
271 |
return cpt; |
|
|
272 |
} else { |
|
|
273 |
cpt++; |
|
|
274 |
} |
|
|
275 |
} |
|
|
276 |
return 0; |
|
|
277 |
} |
|
|
278 |
|
|
|
279 |
public ArrayList<String> getSortedHmDataKeyset() { |
|
|
280 |
ArrayList<String> al = new ArrayList<>(); |
|
|
281 |
TreeMap<String, ArrayList<String>> tm = new TreeMap<>(); |
|
|
282 |
tm.putAll(hmData); |
|
|
283 |
for (String key : tm.keySet()) { |
|
|
284 |
al.add(key); |
|
|
285 |
} |
|
|
286 |
return al; |
|
|
287 |
} |
|
|
288 |
} |
|
|
289 |
|
|
|
290 |
/** |
|
|
291 |
* calculate mean of an array of doubles |
|
|
292 |
* |
|
|
293 |
* @param al |
|
|
294 |
* @return |
|
|
295 |
*/ |
|
|
296 |
public static String getMean(ArrayList<Double> al) { |
|
|
297 |
|
|
|
298 |
if (!al.isEmpty()) { |
|
|
299 |
double d[] = new double[al.size()]; |
|
|
300 |
for (int i = 0; i < al.size(); i++) { |
|
|
301 |
if (!al.get(i).isNaN()) { |
|
|
302 |
d[i] = (double) al.get(i); |
|
|
303 |
} |
|
|
304 |
} |
|
|
305 |
|
|
|
306 |
Mean m = new Mean(); |
|
|
307 |
DecimalFormat df = new DecimalFormat(); |
|
|
308 |
df.setMaximumFractionDigits(3); |
|
|
309 |
DecimalFormatSymbols dfs = new DecimalFormatSymbols(); |
|
|
310 |
dfs.setDecimalSeparator('.'); |
|
|
311 |
df.setDecimalFormatSymbols(dfs); |
|
|
312 |
return df.format(m.evaluate(d)); |
|
|
313 |
} else { |
|
|
314 |
return ""; |
|
|
315 |
} |
|
|
316 |
} |
|
|
317 |
|
|
|
318 |
} |