postgres-omini

PostgreSQL 18 image with these extensions preinstalled and enabled on first database initialization:

• pgvector for vector similarity search
• ParadeDB pg_search for BM25 full-text search
• ParadeDB Jieba/Lindera/CJK tokenizers for Chinese BM25 search
• timescaledb for time-series hypertables, compression, and continuous aggregates
• pgmq for Postgres-native message queues
• zhparser with SCWS for PostgreSQL native Chinese full-text search

Quick start

docker compose up -d --build
psql postgres://postgres:[email protected]:5432/postgres

The init script creates all bundled extensions in the default database:

SELECT extname, extversion
FROM pg_extension
WHERE extname IN ('vector', 'pg_search', 'timescaledb', 'pgmq', 'zhparser')
ORDER BY extname;

Chinese search options:

-- Native PostgreSQL FTS with zhparser.
SELECT to_tsvector('chinese', '中文分词用于全文检索');

-- ParadeDB BM25 with Jieba tokenizer.
CREATE TABLE cn_docs (id bigserial PRIMARY KEY, body text NOT NULL);
INSERT INTO cn_docs (body) VALUES ('PostgreSQL 中文分词和向量搜索开箱即用');
CREATE INDEX cn_docs_search_idx ON cn_docs
USING bm25 (id, (body::pdb.jieba))
WITH (key_field='id');
SELECT id, pdb.score(id) FROM cn_docs WHERE body @@@ '中文搜索';

TimescaleDB example:

CREATE TABLE metrics (
  time timestamptz NOT NULL,
  device text NOT NULL,
  value double precision NOT NULL
);
SELECT create_hypertable('metrics', 'time');
INSERT INTO metrics (time, device, value)
SELECT now() - (i * interval '1 minute'), 'dev-' || (i % 10), random() * 100
FROM generate_series(1, 1000) AS i;
SELECT time_bucket('5 minutes', time) AS bucket, avg(value)
FROM metrics
GROUP BY bucket
ORDER BY bucket DESC;

Build

docker build -t postgres-omini:local .

The default build uses the widely deployed Debian-based postgres:18 image. It compiles pgvector, PGMQ, SCWS, zhparser, and TimescaleDB from source, and installs the official ParadeDB pg_search Debian package for PostgreSQL 18.

If your network needs a proxy or a local Debian mirror:

docker build --add-host win230:100.119.20.46 \
  --build-arg APT_MIRROR=http://mirrors.aliyun.com/debian \
  --build-arg FETCH_HTTP_PROXY=http://win230:10808 \
  --build-arg FETCH_HTTPS_PROXY=http://win230:10808 \
  -t postgres-omini:local .

Benchmark

./scripts/benchmark.sh

Options:

ROWS=100000 IMAGE=postgres-omini:bench ./scripts/benchmark.sh

When the build needs the same proxy arguments:

DOCKER_BUILD_ARGS="--add-host win230:100.119.20.46 --build-arg APT_MIRROR=http://mirrors.aliyun.com/debian --build-arg FETCH_HTTP_PROXY=http://win230:10808 --build-arg FETCH_HTTPS_PROXY=http://win230:10808" \
  ROWS=100000 ./scripts/benchmark.sh

The script builds the image, starts a temporary container, tests pgvector, pg_search, pgmq, TimescaleDB, and Chinese search paths, prints a timing report, then removes the container.

Published image

GitHub Actions publishes to:

ghcr.io/ider-zh/postgres-omini:latest