Skip to content

Commit

Permalink
CLDR-17153 Investigate discrepancy between growth and coverage charts (…
Browse files Browse the repository at this point in the history 
…#3324)

* CLDR-17153 Investigate discrepancy between growth and coverage charts

* CLDR-17153 Cleanup

(cherry picked from commit 3bea383)
  • Loading branch information
@macchiati @pedberg-icu
macchiati authored and pedberg-icu committed Nov 1, 2023
1 parent fb9a9d2 commit 45d4cf7
Show file tree
Hide file tree
Showing 4 changed files with 325 additions and 42 deletions.
147 changes: 131 additions & 16 deletions tools/cldr-code/src/main/java/org/unicode/cldr/tool/ChartLocaleGrowth.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -7,7 +7,6 @@
import com.ibm.icu.util.VersionInfo;
import java.io.File;
import java.io.IOException;
import java.io.PrintWriter;
import java.util.ArrayList;
import java.util.Arrays;
import java.util.Collections;
Expand All @@ -21,7 +20,6 @@
import java.util.Set;
import java.util.TreeMap;
import java.util.regex.Matcher;
import org.unicode.cldr.draft.FileUtilities;
import org.unicode.cldr.tool.Option.Options;
import org.unicode.cldr.util.CLDRConfig;
import org.unicode.cldr.util.CLDRFile;
Expand All @@ -32,13 +30,16 @@
import org.unicode.cldr.util.Counter2;
import org.unicode.cldr.util.Level;
import org.unicode.cldr.util.LocaleNames;
import org.unicode.cldr.util.Organization;
import org.unicode.cldr.util.PathHeader;
import org.unicode.cldr.util.PathHeader.Factory;
import org.unicode.cldr.util.PatternCache;
import org.unicode.cldr.util.RegexLookup;
import org.unicode.cldr.util.RegexLookup.LookupType;
import org.unicode.cldr.util.SimpleFactory;
import org.unicode.cldr.util.StandardCodes;
import org.unicode.cldr.util.SupplementalDataInfo;
import org.unicode.cldr.util.TempPrintWriter;
import org.unicode.cldr.util.VettingViewer;
import org.unicode.cldr.util.VettingViewer.MissingStatus;

Expand All @@ -51,6 +52,11 @@ public class ChartLocaleGrowth {
private static CLDRConfig testInfo = ToolConfig.getToolInstance();
private static final SupplementalDataInfo SUPPLEMENTAL_DATA_INFO =
testInfo.getSupplementalDataInfo();
static final Set<String> CldrModernLocales =
StandardCodes.make().getLocaleCoverageLocales(Organization.cldr, Set.of(Level.MODERN));
static final Set<String> SpecialLocales =
StandardCodes.make()
.getLocaleCoverageLocales(Organization.special, Set.of(Level.MODERN));

private static org.unicode.cldr.util.Factory factory =
testInfo.getCommonAndSeedAndMainAndAnnotationsFactory();
Expand Down Expand Up @@ -79,7 +85,11 @@ public class ChartLocaleGrowth {
static final Options myOptions = new Options();

private enum MyOptions {
filter(".+", ".*", "Filter the information based on id, using a regex argument."),
filter(".+", ".*", "Filter the information based on locale, using a regex argument."),
Versions(
".+",
".*",
"Filter the information based on cldr version, using a regex argument."),
// draftStatus(".+", "unconfirmed", "Filter the information to a minimum draft status."),
;

Expand All @@ -95,25 +105,46 @@ private enum MyOptions {
public static void main(String[] args) throws IOException {
myOptions.parse(MyOptions.filter, args, true);

Matcher matcher = PatternCache.get(MyOptions.filter.option.getValue()).matcher("");

try (PrintWriter out =
FileUtilities.openUTF8Writer(
CLDRPaths.CHART_DIRECTORY + "tsv/", "locale-growth.tsv")) {
doGrowth(matcher, out);
Matcher localeMatcher = PatternCache.get(MyOptions.filter.option.getValue()).matcher("");
Matcher versionMatcher = PatternCache.get(MyOptions.Versions.option.getValue()).matcher("");

try (TempPrintWriter out =
new TempPrintWriter(
CLDRPaths.CHART_DIRECTORY + "tsv/", "locale-growth.tsv");
TempPrintWriter log =
new TempPrintWriter(
CLDRPaths.CHART_DIRECTORY + "tsv/", "locale-growth-log.tsv");
TempPrintWriter logPaths =
new TempPrintWriter(
CLDRPaths.CHART_DIRECTORY + "tsv/", "locale-growth-paths.tsv"); ) {
doGrowth(localeMatcher, versionMatcher, out, log, logPaths);
return;
}
}

private static void doGrowth(Matcher matcher, PrintWriter out) {
private static void doGrowth(
Matcher localeMatcher,
Matcher versionMatcher,
TempPrintWriter out,
TempPrintWriter log,
TempPrintWriter logPaths) {
TreeMap<String, List<Double>> growthData =
new TreeMap<>(Ordering.natural().reverse()); // sort by version, descending
Map<String, FoundAndTotal> latestData = null;
ReleaseInfo last = versionToYear.get(0);
for (ReleaseInfo versionNormalizedVersionAndYear : versionToYear) {
if (versionMatcher != null
&& !versionMatcher
.reset(versionNormalizedVersionAndYear.version.getVersionString(1, 2))
.matches()) {
continue;
}
VersionInfo version = versionNormalizedVersionAndYear.version;
int year = versionNormalizedVersionAndYear.year;
String dir = ToolConstants.getBaseDirectory(version.getVersionString(2, 3));
Map<String, FoundAndTotal> currentData = addGrowth(factory, dir, matcher, false);
boolean showMissing = last == versionNormalizedVersionAndYear;
Map<String, FoundAndTotal> currentData =
addGrowth(factory, dir, localeMatcher, showMissing, log, logPaths);
long found = 0;
long total = 0;
for (Entry<String, FoundAndTotal> entry : currentData.entrySet()) {
Expand Down Expand Up @@ -169,7 +200,7 @@ private ReleaseInfo(VersionInfo versionInfo, int year) {

static {
Object[][] mapping = {
{VersionInfo.getInstance(43), 2023},
{VersionInfo.getInstance(44), 2023},
{VersionInfo.getInstance(42), 2022},
{VersionInfo.getInstance(40), 2021},
{VersionInfo.getInstance(38), 2020},
Expand Down Expand Up @@ -258,8 +289,10 @@ public String toString() {
private static Map<String, FoundAndTotal> addGrowth(
org.unicode.cldr.util.Factory latestFactory,
String dir,
Matcher matcher,
boolean showMissing) {
Matcher localeMatcher,
boolean showMissing,
TempPrintWriter log,
TempPrintWriter logPaths) {
final File mainDir = new File(dir + "/common/main/");
final File annotationDir = new File(dir + "/common/annotations/");
File[] paths =
Expand All @@ -273,8 +306,9 @@ private static Map<String, FoundAndTotal> addGrowth(
Map<String, FoundAndTotal> data = new HashMap<>();
char c = 0;
Set<String> latestAvailable = newFactory.getAvailableLanguages();
boolean firstShowMissing = true;
for (String locale : newFactory.getAvailableLanguages()) {
if (!matcher.reset(locale).matches()) {
if (!localeMatcher.reset(locale).matches()) {
continue;
}
if (!latestAvailable.contains(locale)) {
Expand Down Expand Up @@ -334,6 +368,73 @@ private static Map<String, FoundAndTotal> addGrowth(
missingPaths,
unconfirmedPaths);

if (showMissing) {
if (CldrModernLocales.contains(locale)) {
final boolean isSpecial = SpecialLocales.contains(locale);
if (firstShowMissing) {
firstShowMissing = false;
log.printlnWithTabs(
16,
"Locale\tTC"
+ "\tCore\tUnc\tMiss"
+ "\tBasic\tUnc\tMiss"
+ "\tModer\tUnc\tMiss"
+ "\tModern\tUnc\tMiss"
+ "\tTotal\tUnc\tMiss");
logPaths.printlnWithTabs(3, "Locale\tLevel\tStatus\tPath");
}
log.printlnWithTabs(
16,
locale
+ "\t"
+ (isSpecial ? "" : "TC")
+ show(
Level.CORE,
foundCounter,
unconfirmedCounter,
missingCounter)
+ show(
Level.BASIC,
foundCounter,
unconfirmedCounter,
missingCounter)
+ show(
Level.MODERATE,
foundCounter,
unconfirmedCounter,
missingCounter)
+ show(
Level.MODERN,
foundCounter,
unconfirmedCounter,
missingCounter)
+ show(
null, // total
foundCounter,
unconfirmedCounter,
missingCounter));
if (!isSpecial) {
long count = unconfirmedCounter.getTotal() + missingCounter.getTotal();
for (Entry<MissingStatus, String> statusAndPath : missingPaths.entrySet()) {
logPaths.printlnWithTabs(
3,
locale
+ "\t"
+ count
+ "\t"
+ statusAndPath.getKey()
+ "\t"
+ statusAndPath.getValue());
}
for (String path : unconfirmedPaths) {
logPaths.printlnWithTabs(
3, locale + "\t" + count + "\tunconfirmed\t" + path);
}
}
int line = 0;
}
}

// HACK
Set<Entry<MissingStatus, String>> missingRemovals = new HashSet<>();
for (Entry<MissingStatus, String> e : missingPaths.keyValueSet()) {
Expand All @@ -355,7 +456,7 @@ private static Map<String, FoundAndTotal> addGrowth(
}
// END HACK

if (showMissing) {
if (false && showMissing) {
int count = 0;
for (String s : unconfirmedPaths) {
System.out.println(
Expand All @@ -382,4 +483,18 @@ private static Map<String, FoundAndTotal> addGrowth(
}
return Collections.unmodifiableMap(data);
}

/** "\tCore\tUnc\tMiss" */
private static String show(
Level level,
Counter<Level> foundCounter,
Counter<Level> unconfirmedCounter,
Counter<Level> missingCounter) {
return "\t"
+ (level != null ? foundCounter.get(level) : foundCounter.getTotal())
+ "\t"
+ (level != null ? unconfirmedCounter.get(level) : unconfirmedCounter.getTotal())
+ "\t"
+ (level != null ? missingCounter.get(level) : missingCounter.getTotal());
}
}
Loading

0 comments on commit 45d4cf7

Please sign in to comment.