Tabularis — Elasticsearch Plugin

A Tabularis driver plugin that lets Tabularis users inspect and query Elasticsearch clusters.

Current features

• Index browsing: list indices and view basic index stats
• Mapping viewer: inspect index mappings (fields, types, nested structures)
• Document samples: preview documents from an index (sampling/scroll)
• Query execution: run Elasticsearch queries and show results in the Tabularis UI

Usage

The plugin supports four query modes. The mode is determined by an optional shebang (#!) at the beginning of the query.

Mode	Prefix	Description
SQL (default)	(none) or #!sql	Execute Elasticsearch SQL queries.
ES|QL	#!esql	Execute ES|QL queries.
REST	#!rest	Send raw Elasticsearch REST requests.

For examples

SELECT * FROM user_index_000000004;

-- Or explicitly:

#!sql
SELECT * FROM user_index_000000004;


#!esql
FROM user_index_000000004

#!rest
POST /post_index/_search
{"query":{"match_all":{}},"fields":[{"field":"id"},{"field":"content"}],"sort":[{"_doc":{"order":"asc"}}],"track_total_hits":-1,"_source":true}

The first line must contain the HTTP method and endpoint. The remaining content is sent as the request body.

Installation

Install from the registry

This plugin is published to the Tabularium registry as tabularis-elasticsearch-plugin. Search for Elasticsearch in the Tabularis plugin browser and install it from there.

If you point Tabularis at a different Tabularium instance via customRegistryUrl in config.json, make sure that registry has ingested this plugin's releases.

Installation (developer workflow)

Build and install the plugin locally:

just dev-install    # builds and installs to $HOME/Library/Application Support/com.debba.tabularis/plugins/elasticsearch

Open Tabularis and choose the "Elasticsearch" driver in the connection picker. Configure a connection http://<username>:<password>@<host>:<port> to start exploring indices.

Explain

Copy the binary and manifest.json into the Tabularis plugins folder under a elasticsearch/ subdirectory:

OS	Path
Linux	~/.local/share/tabularis/plugins/elasticsearch/
macOS	~/Library/Application Support/tabularis/plugins/elasticsearch/
Windows	%APPDATA%\debba\tabularis\data\plugins\elasticsearch\

Restart Tabularis (or install via Settings) and pick Elasticsearch in the connection form.

Build

cargo build --release
# binary: target/release/tabularis-elasticsearch-plugin

Development

For setup test environment (run Elasticsearch and seed-data)

# Run Elasticsearch 
just run-es

# Seed data for Elasticsearch
just seed-es

Run the local REPL to test handlers without Tabularis:

just repl
# use REPL commands to exercise rpc handlers

Project layout (high level):

• src/main.rs — stdio loop for plugin RPC transport
• src/rpc.rs — method dispatch and helper responses
• src/es/* — Elasticsearch client wrapper and connection config
• src/handlers/ — metadata, query, sample, and mapping handlers
• src/models.rs — connection params and shared types
• bin/test_plugin.rs — REPL for exercising RPC handlers locally

Releasing

This repo ships two manifests that must stay in sync with each other and with the release tag:

• manifest.json — the runtime manifest, packaged inside each platform .zip.
• .tabularium — the registry manifest read by Tabularium when it ingests a release.

To cut a release:

1. Bump version in Cargo.toml, manifest.json, and .tabularium to the same value.
2. Commit, then push a tag vX.Y.Z matching that version.
3. .github/workflows/release.yml validates the tag against both manifests, builds binaries for all 5 platforms, and publishes a GitHub Release with the zips attached.
4. Submit (or re-submit) the plugin at registry.tabularis.dev/submit so the new version is ingested. Check /requests for approval status.

Maintainers

• @erwin-lovecraft

License

Apache-2.0.