Mercurial > dive4elements > river

annotate flys-backend/src/main/java/de/intevation/flys/importer/StaFileParser.java @ 192:f1fce41347ea

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Added missing ImportGauge.getPeer() method flys-backend/trunk@1530 c6561f87-3c4e-4783-a992-168aeb5c3f6f
author Sascha L. Teichmann <sascha.teichmann@intevation.de>
date Mon, 21 Mar 2011 10:27:14 +0000
parents bc3747a371cc
children 8d3533a03e10
rev   line source
189
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
1 package de.intevation.flys.importer;
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
2
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
3 import java.io.File;
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
4 import java.io.IOException;
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
5 import java.io.LineNumberReader;
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
6 import java.io.FileInputStream;
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
7 import java.io.InputStreamReader;
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
8
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
9 import java.math.BigDecimal;
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
10
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
11 import java.util.regex.Pattern;
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
12 import java.util.regex.Matcher;
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
13
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
14 import java.util.HashMap;
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
15
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
16 import org.apache.log4j.Logger;
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
17
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
18 public class StaFileParser
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
19 {
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
20 private static Logger log = Logger.getLogger(StaFileParser.class);
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
21
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
22 public static final String ENCODING = "ISO-8859-1";
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
23
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
24 public static final Pattern QWTD_ =
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
25 Pattern.compile("\\s*([^\\s]+)\\s+([^\\s]+)\\s+([QWTD-]).*");
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
26
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
27 public StaFileParser() {
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
28 }
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
29
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
30 public boolean parse(ImportGauge gauge) throws IOException {
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
31
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
32 File file = gauge.getStaFile();
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
33
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
34 log.info("parsing STA file: " + file);
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
35 LineNumberReader in = null;
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
36 try {
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
37 in =
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
38 new LineNumberReader(
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
39 new InputStreamReader(
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
40 new FileInputStream(file), ENCODING));
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
41
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
42 String line = in.readLine();
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
43
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
44 if (line == null) {
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
45 log.warn("STA file is empty.");
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
46 return false;
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
47 }
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
48
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
49 if (line.length() < 37) {
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
50 log.warn("first line in STA file is too short.");
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
51 return false;
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
52 }
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
53
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
54 gauge.setName(line.substring(16, 37).trim());
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
55
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
56 String [] values = line.substring(38).trim().split("\\s+", 2);
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
57
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
58 if (values.length < 2) {
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
59 log.warn("Not enough columns for aeo and datum");
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
60 }
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
61 try {
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
62 gauge.setAeo(new BigDecimal(values[0].replace(",", ".")));
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
63 gauge.setDatum(new BigDecimal(values[1].replace(",", ".")));
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
64 }
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
65 catch (NumberFormatException nfe) {
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
66 log.warn("cannot parse aeo or datum");
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
67 return false;
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
68 }
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
69
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
70 line = in.readLine();
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
71
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
72 if (line == null) {
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
73 log.warn("STA file has not enough lines");
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
74 return false;
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
75 }
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
76
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
77 if (line.length() < 36) {
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
78 log.warn("second line is too short");
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
79 return false;
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
80 }
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
81
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
82 try {
192
f1fce41347ea Added missing ImportGauge.getPeer() method
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents: 189
diff changeset
83 gauge.setStation(
189
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
84 new BigDecimal(line.substring(29, 36).trim()));
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
85 }
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
86 catch (NumberFormatException nfe) {
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
87 log.warn("parsing of the datum of the gauge failed");
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
88 return false;
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
89 }
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
90
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
91 // overread the next six lines
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
92 for (int i = 0; i < 6; ++i) {
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
93 if ((line = in.readLine()) == null) {
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
94 log.warn("STA file is too short");
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
95 return false;
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
96 }
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
97 }
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
98
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
99 HashMap<String, ImportMainValueType> types =
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
100 new HashMap<String, ImportMainValueType>();
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
101
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
102 while ((line = in.readLine()) != null) {
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
103 Matcher m = QWTD_.matcher(line);
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
104 if (m.matches()) {
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
105 BigDecimal value;
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
106 try {
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
107 value = new BigDecimal(m.group(2).replace(",", "."));
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
108 }
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
109 catch (NumberFormatException nfe) {
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
110 log.warn("value not parseable in line "
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
111 + in.getLineNumber());
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
112 continue;
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
113 }
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
114 String typeString = m.group(3);
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
115 log.debug("\t type: " + typeString);
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
116 ImportMainValueType type = types.get(typeString);
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
117 if (type == null) {
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
118 type = new ImportMainValueType(typeString);
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
119 types.put(typeString, type);
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
120 }
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
121 }
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
122 else {
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
123 // TODO: treat as a comment
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
124 }
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
125 }
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
126 }
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
127 finally {
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
128 if (in != null) {
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
129 in.close();
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
130 }
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
131 }
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
132 log.info("finished parsing STA file: " + file);
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
133 return true;
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
134 }
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
135 }
bc3747a371cc First part of parsing main values.
Sascha L. Teichmann <sascha.teichmann@intevation.de>
parents:
diff changeset
136 // vim:set ts=4 sw=4 si et sta sts=4 fenc=utf8 :
http://dive4elements.wald.intevation.org