Skip to content

CLAUDE.md

Latest commit

 

History

History
118 lines (79 loc) · 4.65 KB

CLAUDE.md

File metadata and controls

118 lines (79 loc) · 4.65 KB

CLAUDE.md — Java Module

This file provides guidance to Claude Code (claude.ai/code) when working with the Java module.

Build Commands

Maven with wrapper (./mvnw from repo root). All commands require -P with-java.

# Build (skip tests)
./mvnw clean package -P with-java -DskipTests

# Install to local Maven repo
./mvnw install -P with-java -DskipTests

# Run all tests (unit + integration)
./mvnw clean verify -P with-java

# Run unit tests only
./mvnw test -P with-java

# Run integration tests only
./mvnw verify -P with-java -DskipUTs=true

# Run a single test class (use -pl to target submodule)
./mvnw test -P with-java -pl java/tsfile -Dtest=TsFileWriterTest

# Run a single test method
./mvnw test -P with-java -pl java/tsfile -Dtest=TsFileWriterTest#testWrite

# Format code (requires Java 17+ runtime)
./mvnw spotless:apply

# Check formatting
./mvnw spotless:check

Module Structure

Four submodules under java/:

  • common — Shared types: TSDataType, ColumnCategory, Binary, BitMap, constants
  • tsfile — Core implementation: read/write APIs, encoding, compression, file format
  • examples — Reference implementations for read/write operations
  • tools — CLI utilities for TsFile inspection

Architecture

Key packages in java/tsfile/src/main/java/org/apache/tsfile/:

  • write/ — Write path: TsFileWriter, Tablet (batch API), TSRecord (row API)
  • read/ — Read path: TsFileReader, TsFileSequenceReader, QueryExpression
  • encoding/ — Encoding algorithms (RLE, TS_2DIFF, GORILLA, DICTIONARY)
  • compress/ — Compression codecs (Snappy, LZ4, ZSTD, XZ)
  • file/ — File format structures (metadata, headers, footers)
  • compatibility/ — Version compatibility handling
  • parser/ — ANTLR4-generated path parser (grammar in src/main/antlr4/)

Code generation: FreeMarker templates in tsfile/src/main/codegen/ generate type-specific implementations (e.g., per-datatype readers/writers) into target/generated-sources/.

Code Style

  • Formatter: Spotless with Google Java Format 1.28.0 (Spotless requires Java 17+ to run but the project targets Java 8 bytecode)
  • Checkstyle: Google style variant in root checkstyle.xml, 100 char line limit
  • Import order: org.apache.tsfile, javax, java, static imports

Testing Conventions

  • Framework: JUnit 4
  • Unit tests: *Test.java — run via mvn test
  • Integration tests: *IT.java — run via mvn verify
  • Tests run in random order with reuseForks=false

Internationalization (i18n)

All user-facing log messages and exception messages in java/common, java/tsfile, and java/tools go through org.apache.tsfile.i18n.Messages. Hardcoded English strings in LOGGER.* / LOG.* / logger.* or throw new XxxException(...) are not allowed for new code — add an entry to java/common/src/main/resources/org/apache/tsfile/i18n/messages.properties and reference it via:

  • Messages.get(key) for SLF4J patterns (containing {} placeholders)
  • Messages.format(key, args) for exception patterns (containing %1$s positional placeholders)

Locale at runtime is chosen in this order:

  1. -Dtsfile.locale=<tag> system property (e.g. zh, zh-CN)
  2. Locale.getDefault()
  3. English fallback (root bundle)

When adding a key, mirror it in messages_zh.properties with a Simplified Chinese translation. MessagesTest enforces key-set alignment between the two bundles — CI fails on drift.

Technical terms (class names, method names, type names, codec names like TsFile, Chunk, Page, Schema, RLE, GORILLA) are kept in English in Chinese translations.

License Header

Every new file must include the Apache License 2.0 header at the top. For Java files, use the /* */ block comment style. See any existing .java file for the exact wording.

Git Commit

  • Do NOT add Co-Authored-By trailer to commit messages.