Skip to content

A more compact charindex #1331

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
eggrobin merged 3 commits into from
Apr 20, 2026
Merged

A more compact charindex #1331

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
ada735e
A more compact charindex
eggrobin Apr 14, 2026
78f0907
Save another 433 kiB
eggrobin Apr 15, 2026
07ad585
helper class
eggrobin Apr 20, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
167 changes: 118 additions & 49 deletions unicodetools/src/main/java/org/unicode/text/tools/Indexer.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -9,12 +9,14 @@
import com.ibm.icu.text.Transliterator;
import com.ibm.icu.text.UnicodeSet;
import java.io.BufferedReader;
import java.io.ByteArrayOutputStream;
import java.io.File;
import java.io.FileReader;
import java.io.IOException;
import java.io.PrintStream;
import java.util.ArrayList;
import java.util.Arrays;
import java.util.Base64;
import java.util.Comparator;
import java.util.HashMap;
import java.util.HashSet;
Expand All @@ -26,6 +28,7 @@
import java.util.TreeMap;
import java.util.stream.Collectors;
import java.util.stream.StreamSupport;
import java.util.zip.DeflaterOutputStream;
import org.unicode.props.IndexUnicodeProperties;
import org.unicode.props.UcdProperty;
import org.unicode.props.UcdPropertyValues;
Expand All @@ -38,6 +41,8 @@

public class Indexer {

private static final char RECORD_SEPARATOR = 0x001E;

private static Transliterator toHTML;
private static String htmlRulesControls;

Expand Down Expand Up @@ -82,6 +87,27 @@ public class Indexer {

private static int maxRSEntryCharacters = 0;

private static class StringIndexer {
public StringIndexer() {}

public int getStringIndex(String s) {
int result = stringIndices.getOrDefault(s, allTheStrings.length());
if (result == allTheStrings.length()) {
allTheStrings.append(s).append(RECORD_SEPARATOR);
stringIndices.put(s, result);
}
return result;
}

@Override
public String toString() {
return allTheStrings.toString();
}

private final HashMap<String, Integer> stringIndices = new HashMap<>();
private final StringBuilder allTheStrings = new StringBuilder();
}

static {
String baseRules =
"'<' > '&lt;' ;"
Expand Down Expand Up @@ -113,8 +139,8 @@ public class Indexer {
}

private static class IndexEntry {
IndexEntry(String snippet, UnicodeProperty property) {
this.snippet = snippet;
IndexEntry(int snippetIndex, UnicodeProperty property) {
this.snippetIndex = snippetIndex;
this.property = property;
characters = new UnicodeSet();
}
Expand All @@ -127,12 +153,12 @@ List<IndexSubEntry> subEntries() {
/* showName= */ property != NAME,
characters);
} catch (Exception e) {
System.err.println("In entry for " + property.getName() + ": " + snippet);
System.err.println("In entry for " + property.getName() + ": " + snippetIndex);
throw e;
}
}

String snippet;
int snippetIndex;
UnicodeProperty property;
UnicodeSet characters;
Map<String, UnicodeSet> relatedCharacters = new TreeMap<>();
Expand Down Expand Up @@ -258,11 +284,13 @@ public int compare(UnicodeProperty left, UnicodeProperty right) {
return left.getName().compareTo(right.getName());
}
}
// Property to property value to index entry.
Map<UnicodeProperty, Map<String, IndexEntry>> indexEntries =
final var allTheStrings = new StringIndexer();
// Property to snippet based on property value (as an index in allTheStrings) to index
// entry.
Map<UnicodeProperty, Map<Integer, IndexEntry>> indexEntries =
new TreeMap<>(new PropertyComparator());
// Lemma to snippet to position of the word in the snippet.
Map<String, Map<String, Integer>> wordIndex = new TreeMap<>();
// Lemma to snippet (as an index in allTheStrings) to position of the word in the snippet.
Map<String, Map<Integer, Integer>> wordIndex = new TreeMap<>();
final var properties =
List.of(
BLOCK,
Expand Down Expand Up @@ -292,10 +320,9 @@ public int compare(UnicodeProperty left, UnicodeProperty right) {
} else if (prop == NAME) {
snippet = snippet.replace(Utility.hex(cp), "#");
}
// Copy the snippet to a final variable for use in the λ.
final String indexSnippet = snippet;
final int snippetIndex = allTheStrings.getStringIndex(snippet);
propertyIndex
.computeIfAbsent(snippet, k -> new IndexEntry(indexSnippet, prop))
.computeIfAbsent(snippetIndex, k -> new IndexEntry(k, prop))
.characters
.add(cp);
// Override word breaking of ' and - in appropriate contexts so that
Expand All @@ -313,11 +340,11 @@ public int compare(UnicodeProperty left, UnicodeProperty right) {
String lemma = lemmatize(word);
wordIndex
.computeIfAbsent(fold(word), k -> new TreeMap<>())
.putIfAbsent(snippet, start);
.putIfAbsent(snippetIndex, start);
if (!lemma.equals(fold(word))) {
wordIndex
.computeIfAbsent(lemma, k -> new TreeMap<>())
.putIfAbsent(snippet, start);
.putIfAbsent(snippetIndex, start);
}
}
}
Expand All @@ -327,18 +354,20 @@ public int compare(UnicodeProperty left, UnicodeProperty right) {
System.out.println("Indexed plane " + cp / 0x10000);
}
}
final int bettyIndex = allTheStrings.getStringIndex("Betty");
final int theIndex = allTheStrings.getStringIndex("the");
indexEntries
.get(BLOCK)
.computeIfAbsent("Betty", k -> new IndexEntry(k, BLOCK))
.computeIfAbsent(bettyIndex, k -> new IndexEntry(k, BLOCK))
.characters
.add(BOOP);
indexEntries
.get(BLOCK)
.computeIfAbsent("the", k -> new IndexEntry(k, BLOCK))
.computeIfAbsent(theIndex, k -> new IndexEntry(k, BLOCK))
.characters
.add(DOOD);
wordIndex.computeIfAbsent("betty", k -> new TreeMap<>()).putIfAbsent("Betty", 0);
wordIndex.computeIfAbsent("the", k -> new TreeMap<>()).putIfAbsent("the", 0);
wordIndex.computeIfAbsent("betty", k -> new TreeMap<>()).putIfAbsent(bettyIndex, 0);
wordIndex.computeIfAbsent("the", k -> new TreeMap<>()).putIfAbsent(theIndex, 0);

System.out.println("Radicals…");
final var radicalSets = getRadicalSets();
Expand Down Expand Up @@ -379,61 +408,101 @@ public int compare(UnicodeProperty left, UnicodeProperty right) {
}
css.close();
} else if (htmlLine.contains("JS HERE")) {
file.println("let wordIndex = new Map([");
// No pretty-printing in the loops that print these two maps; each space or newline
// here enlarges charindex.html by hundreds of kilobytes. These are not suitable
// for human consumption anyway, since anything readable is turned into indices in
// allTheStrings.
file.print("let wordIndex = new Map([");
System.out.println("wordIndex...");
{
int i = 0;
for (var wordAndSnippets : wordIndex.entrySet()) {
if (++i % 1000 == 0) {
System.out.println(i + "/" + wordIndex.size() + "...");
}
file.println(
" ['"
file.print(
"['"
+ wordAndSnippets.getKey().replace("'", "\\'")
+ "', new Map([");
for (var snippetAndPosition : wordAndSnippets.getValue().entrySet()) {
file.println(
" ['"
+ snippetAndPosition.getKey().replace("'", "\\'")
+ "', "
+ snippetAndPosition.getValue()
+ "],");
}
file.println("])],");
+ "',new Map([");
// Stream and collect for the innermost map to avoid trailing commas, for
// size.
file.print(
wordAndSnippets.getValue().entrySet().stream()
.map(
snippetAndPosition ->
"["
+ snippetAndPosition.getKey()
+ ","
+ snippetAndPosition.getValue()
+ "]")
.collect(Collectors.joining(",")));
file.print("])],");
}
}
file.println("]);");
System.out.println("indexEntries...");
file.println("let indexEntries = new Map([");
file.print("let indexEntries = new Map([");
for (var property : properties) {
System.out.println(property.getName() + "...");
final var propertyIndex = indexEntries.get(property);
file.println(" ['" + property.getName() + "', new Map([");
file.print("['" + property.getName() + "',new Map([");
int i = 0;
for (var indexEntry : propertyIndex.values()) {
if (++i % 1000 == 0) {
System.out.println(i + "/" + propertyIndex.size() + "...");
}
file.println(" ['" + indexEntry.snippet.replace("'", "\\'") + "', {");
file.println(
" html: \""
+ indexEntry.toHTML().replace("\"", "\\\"")
+ "\",");
file.println(" characters: [");
for (var range : indexEntry.coveredCharacters().ranges()) {
file.println(
" [0x"
+ Utility.hex(range.codepoint)
+ ", 0x"
+ Utility.hex(range.codepointEnd)
+ "],");
}
file.println(" ],");
file.println(" }],");
final int htmlIndex = allTheStrings.getStringIndex(indexEntry.toHTML());
file.print("[" + indexEntry.snippetIndex + ",{");
file.print("html:" + htmlIndex + ",");
file.print("characters:[");
// Stream and collect for the innermost array to avoid trailing commas, for
// size.
file.print(
indexEntry
.coveredCharacters()
.rangeStream()
.map(
range ->
// Code points in decimal without
// zero-padding for size.
"["
+ range.codepoint
+ (range.codepointEnd
!= range.codepoint
? "," + range.codepointEnd
: "")
+ "]")
.collect(Collectors.joining(",")));
file.print("]}],");
}
file.println(" ])],");
file.print("])],");
}
file.println("]);");
file.println("let bettyIndex = " + bettyIndex + ";");
file.println("let theIndex = " + theIndex + ";");
final var compressed = new ByteArrayOutputStream();
final var compressor = new DeflaterOutputStream(compressed);
final var uncompressed = allTheStrings.toString().getBytes("UTF-8");
compressor.write(uncompressed);
compressor.close();
final var compressedBytes = compressed.toByteArray();
System.out.println(
"Strings compressed from "
+ (uncompressed.length >> 20)
+ " MiB to "
+ (compressedBytes.length >> 10)
+ " kiB ("
+ 100 * compressedBytes.length / uncompressed.length
+ "%)");
System.out.println(
"Compressed payload is "
+ compressedBytes.length
+ " bytes, first byte is "
+ Byte.toUnsignedInt(compressedBytes[0]));
file.println(
"let allTheStringsCompressed = '"
+ Base64.getEncoder().encodeToString(compressedBytes)
+ "'");
final var js =
new BufferedReader(new FileReader(new File(resources + "charindex.js")));
for (String jsLine = js.readLine(); jsLine != null; jsLine = js.readLine()) {
Expand Down
Loading
Loading