Skip to content

Commit 169a45e

Browse files
jycrjycr
committed
refactor: Remove JShell for custom resources post-processing
1 parent 73db5d0 commit 169a45e

File tree

7 files changed

+494
-312
lines changed

7 files changed

+494
-312
lines changed

‎build/main/java/org/greencodeinitiative/tools/exporter/Main.java‎

Lines changed: 66 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,66 @@
1+
package org.greencodeinitiative.tools.exporter;
2+
3+
import org.greencodeinitiative.tools.exporter.infra.MetadataWriter;
4+
5+
import java.nio.file.Path;
6+
import java.util.List;
7+
import java.util.Map;
8+
import java.util.Optional;
9+
10+
import static java.util.Collections.emptyList;
11+
import static java.util.Optional.empty;
12+
import static java.util.Optional.ofNullable;
13+
14+
public class Main implements Runnable {
15+
private final List<String> args;
16+
17+
public Main(String[] args) {
18+
this.args = ofNullable(args).map(List::of).orElse(emptyList());
19+
}
20+
21+
public static void main(String... args) {
22+
new Main(args).run();
23+
}
24+
25+
@Override
26+
public void run() {
27+
Path sourceDir = argAsPath(0, "sourceDir");
28+
Path targetDir = argAsPath(1, "targetDir");
29+
new MetadataWriter(
30+
sourceDir,
31+
targetDir,
32+
Map.of(
33+
"title", arg(2, "specificationTitle"),
34+
"version", arg(3, "specificationVersion"),
35+
"scmRevisionNumber", arg(4, "specificationSCMRevisionNumber"),
36+
"scmRevisionDate", arg(5, "specificationSCMRevisionDate")
37+
),
38+
// indexFile
39+
optionalArg(6)
40+
.map(Path::of)
41+
.orElseGet(() -> targetDir.resolve("index.json")),
42+
// minTermLength
43+
optionalArg(7)
44+
.map(Integer::parseInt)
45+
.orElse(4)
46+
).run();
47+
}
48+
49+
private Optional<String> optionalArg(int index) {
50+
if (args.size() <= index) {
51+
return empty();
52+
}
53+
return Optional.of(args.get(index));
54+
}
55+
56+
private String arg(int index, String description) {
57+
if (args.size() <= index) {
58+
throw new IllegalArgumentException("argument " + (index + 1) + " is required: " + description);
59+
}
60+
return optionalArg(index).orElseThrow(() -> new IllegalArgumentException("argument " + (index + 1) + " is required: " + description));
61+
}
62+
63+
private Path argAsPath(int index, String description) {
64+
return Path.of(arg(index, description));
65+
}
66+
}

‎build/main/java/org/greencodeinitiative/tools/exporter/domain/Rule.java‎

Lines changed: 87 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,87 @@
1+
package org.greencodeinitiative.tools.exporter.domain;
2+
3+
import java.nio.file.Files;
4+
import java.nio.file.Path;
5+
import java.util.Optional;
6+
import java.util.regex.Matcher;
7+
import java.util.regex.Pattern;
8+
9+
import static java.util.Optional.empty;
10+
import static java.util.Optional.of;
11+
12+
public class Rule {
13+
/**
14+
* Resources to include
15+
*/
16+
private static final Pattern TARGET_RESOURCES = Pattern.compile("[^/]{1,1000}/(?<ruleKey>GCI\\d{1,20})/(?<language>[^/]{1,100})/[^/]{3,20}\\.html$");
17+
18+
public static Optional<Rule> createFromHtmlDescription(Path htmlDescription) {
19+
final Matcher matcher = TARGET_RESOURCES.matcher(htmlDescription.toString().replace('\\', '/'));
20+
if (!matcher.find()) {
21+
return empty();
22+
}
23+
final String ruleKey = matcher.group("ruleKey");
24+
final Path metadata = htmlDescription.getParent().getParent().resolve(ruleKey + ".json");
25+
final Path specificMetadata = htmlDescription.getParent().resolve(ruleKey + ".json");
26+
27+
if (!Files.isRegularFile(htmlDescription) || !Files.isRegularFile(metadata)) {
28+
return empty();
29+
}
30+
31+
return of(new Rule(
32+
ruleKey,
33+
matcher.group("language"),
34+
htmlDescription,
35+
metadata,
36+
specificMetadata
37+
));
38+
}
39+
40+
private final String ruleKey;
41+
private final String language;
42+
private final Path htmlDescription;
43+
private final Path metadata;
44+
private final Path specificMetadata;
45+
46+
Rule(
47+
String ruleKey,
48+
String language,
49+
Path htmlDescription,
50+
Path metadata,
51+
Path specificMetadata
52+
) {
53+
this.ruleKey = ruleKey;
54+
this.language = language;
55+
this.htmlDescription = htmlDescription;
56+
this.metadata = metadata;
57+
this.specificMetadata = specificMetadata;
58+
}
59+
60+
public Path getHtmlDescriptionTargetPath(Path targetDir) {
61+
return targetDir.resolve(language).resolve(htmlDescription.getFileName());
62+
}
63+
64+
public Path getMetadataTargetPath(Path targetDir) {
65+
return targetDir.resolve(language).resolve(metadata.getFileName());
66+
}
67+
68+
public String ruleKey() {
69+
return ruleKey;
70+
}
71+
72+
public String language() {
73+
return language;
74+
}
75+
76+
public Path htmlDescription() {
77+
return htmlDescription;
78+
}
79+
80+
public Path metadata() {
81+
return metadata;
82+
}
83+
84+
public Path specificMetadata() {
85+
return specificMetadata;
86+
}
87+
}

‎build/main/java/org/greencodeinitiative/tools/exporter/infra/MetadataWriter.java‎

Lines changed: 184 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,184 @@
1+
package org.greencodeinitiative.tools.exporter.infra;
2+
3+
import jakarta.json.Json;
4+
import jakarta.json.JsonMergePatch;
5+
import jakarta.json.JsonObject;
6+
import jakarta.json.JsonObjectBuilder;
7+
import jakarta.json.JsonReader;
8+
import jakarta.json.JsonValue;
9+
import jakarta.json.JsonWriter;
10+
import org.greencodeinitiative.tools.exporter.domain.Rule;
11+
import org.jsoup.Jsoup;
12+
13+
import java.io.IOException;
14+
import java.nio.file.Files;
15+
import java.nio.file.Path;
16+
import java.nio.file.StandardCopyOption;
17+
import java.text.Normalizer;
18+
import java.util.List;
19+
import java.util.Locale;
20+
import java.util.Map;
21+
import java.util.Optional;
22+
import java.util.TreeMap;
23+
import java.util.stream.Collectors;
24+
import java.util.stream.Stream;
25+
26+
import static java.lang.System.Logger.Level.DEBUG;
27+
28+
public class MetadataWriter implements Runnable {
29+
private static final System.Logger LOGGER = System.getLogger("MetadataWriter");
30+
31+
private final Path sourceDir;
32+
private final Path targetDir;
33+
private final Path indexFilepath;
34+
private final Map<String, String> specificationInfo;
35+
private final int minTermLength;
36+
37+
public MetadataWriter(
38+
Path sourceDir,
39+
Path targetDir,
40+
Map<String, String> specificationInfo,
41+
Path indexFilepath,
42+
Integer minTermLength
43+
) {
44+
this.sourceDir = sourceDir;
45+
this.targetDir = targetDir;
46+
this.indexFilepath = indexFilepath;
47+
this.specificationInfo = specificationInfo;
48+
this.minTermLength = minTermLength;
49+
}
50+
51+
@Override
52+
public void run() {
53+
var rulesMap = new TreeMap<String, JsonObjectBuilder>();
54+
55+
getResourcesToCopy().forEach(rule -> {
56+
var rulesByLanguage = rulesMap.computeIfAbsent(rule.ruleKey(), k -> Json.createObjectBuilder());
57+
var resultMetadata = mergeOrCopyJsonMetadata(rule.metadata(), rule.specificMetadata(), rule.getMetadataTargetPath(targetDir));
58+
59+
var htmlDescriptionRelativePath = this.indexFilepath.getParent()
60+
.relativize(rule.getHtmlDescriptionTargetPath(targetDir))
61+
.toString();
62+
var resultMetadataBuilder = Json.createObjectBuilder(resultMetadata);
63+
resultMetadataBuilder.add("key", rule.ruleKey());
64+
resultMetadataBuilder.add("language", rule.language());
65+
resultMetadataBuilder.add("severity", resultMetadata.getString("defaultSeverity").toUpperCase(Locale.ENGLISH));
66+
resultMetadataBuilder.add("htmlDescription", htmlDescriptionRelativePath);
67+
resultMetadataBuilder.add("terms", extractTermsFromHtmlFile(rule.htmlDescription()));
68+
69+
rulesByLanguage.add(rule.language(), resultMetadataBuilder);
70+
copyFile(rule.htmlDescription(), rule.getHtmlDescriptionTargetPath(targetDir));
71+
});
72+
73+
writeIndexFile(rulesMap);
74+
}
75+
76+
private String extractTermsFromHtmlFile(Path htmlFile) {
77+
try {
78+
var textContent = Jsoup.parse(htmlFile).select("body").text();
79+
return Stream
80+
.of(
81+
removeDiacritics(textContent)
82+
.toLowerCase(Locale.ENGLISH)
83+
.replaceAll("[^a-zA-Z0-9]", " ")
84+
.trim()
85+
.split("[\\s\\n\\r]+")
86+
)
87+
.filter(term -> term.length() >= minTermLength)
88+
.distinct()
89+
.sorted()
90+
.collect(Collectors.joining(" "));
91+
} catch (IOException e) {
92+
throw new ProcessException("Unable to parse HTML file: " + htmlFile, e);
93+
}
94+
}
95+
96+
private String removeDiacritics(String text) {
97+
Normalizer.normalize(text, Normalizer.Form.NFKD);
98+
return text.replaceAll("[^\\p{ASCII}]", "").replaceAll("\\p{M}", "");
99+
}
100+
101+
private void writeIndexFile(TreeMap<String, JsonObjectBuilder> rulesMap) {
102+
if (indexFilepath == null) {
103+
return;
104+
}
105+
var rules = Json.createObjectBuilder();
106+
rulesMap.forEach(rules::add);
107+
var result = Json.createObjectBuilder();
108+
109+
var specification = Json.createObjectBuilder();
110+
this.specificationInfo.forEach(specification::add);
111+
result.add("specification", specification);
112+
113+
result.add("rules", rules);
114+
115+
try (var resultJsonWriter = Json.createWriter(Files.newBufferedWriter(indexFilepath))) {
116+
resultJsonWriter.write(result.build());
117+
} catch (IOException e) {
118+
throw new ProcessException("cannot write file: " + indexFilepath, e);
119+
}
120+
}
121+
122+
private List<Rule> getResourcesToCopy() {
123+
try (Stream<Path> stream = Files.walk(sourceDir)) {
124+
return stream
125+
.filter(Files::isRegularFile)
126+
.map(Rule::createFromHtmlDescription)
127+
.filter(Optional::isPresent)
128+
.map(Optional::get)
129+
.collect(Collectors.toList());
130+
} catch (IOException e) {
131+
throw new IllegalStateException(e);
132+
}
133+
}
134+
135+
private JsonObject mergeOrCopyJsonMetadata(Path source, Path merge, Path target) {
136+
try {
137+
Files.createDirectories(target.getParent());
138+
} catch (IOException e) {
139+
throw new ProcessException("cannot create directory: " + target.getParent(), e);
140+
}
141+
if (Files.isRegularFile(merge)) {
142+
return mergeJsonFile(source, merge, target).asJsonObject();
143+
} else {
144+
copyFile(source, target);
145+
try (JsonReader targetJsonReader = Json.createReader(Files.newBufferedReader(target))) {
146+
return targetJsonReader.readObject();
147+
} catch (IOException e) {
148+
throw new ProcessException("cannot process source: " + source, e);
149+
}
150+
}
151+
}
152+
153+
private JsonValue mergeJsonFile(Path source, Path merge, Path target) {
154+
LOGGER.log(DEBUG, "Merge: {0} and {1} -> {2}", source, merge, target);
155+
156+
try (
157+
JsonReader sourceJsonReader = Json.createReader(Files.newBufferedReader(source));
158+
JsonReader mergeJsonReader = Json.createReader(Files.newBufferedReader(merge));
159+
JsonWriter resultJsonWriter = Json.createWriter(Files.newBufferedWriter(target));
160+
) {
161+
Files.createDirectories(target.getParent());
162+
163+
JsonObject sourceJson = sourceJsonReader.readObject();
164+
JsonObject mergeJson = mergeJsonReader.readObject();
165+
166+
JsonMergePatch mergePatch = Json.createMergePatch(mergeJson);
167+
JsonValue result = mergePatch.apply(sourceJson);
168+
169+
resultJsonWriter.write(result);
170+
return result;
171+
} catch (IOException e) {
172+
throw new ProcessException("cannot process source: " + source, e);
173+
}
174+
}
175+
176+
private void copyFile(Path source, Path target) {
177+
try {
178+
LOGGER.log(DEBUG, "Copy: {0} -> {1}", source, target);
179+
Files.copy(source, target, StandardCopyOption.REPLACE_EXISTING);
180+
} catch (IOException e) {
181+
throw new ProcessException("unable to copy '" + source + "' to '" + target + "'", e);
182+
}
183+
}
184+
}

‎build/main/java/org/greencodeinitiative/tools/exporter/infra/ProcessException.java‎

Lines changed: 7 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,7 @@
1+
package org.greencodeinitiative.tools.exporter.infra;
2+
3+
public class ProcessException extends RuntimeException {
4+
public ProcessException(String message, Throwable cause) {
5+
super(message, cause);
6+
}
7+
}

0 commit comments

Comments
 (0)