Note: This file is written in Markdown and is best viewed with a Markdown viewer (e.g., GitHub, GitLab, VS Code, or a dedicated Markdown reader). Viewing it in a plain text editor may not render the formatting as intended.
Copyright (c) 2025 Software Tree
Basic example demonstrating RESTful CRUD operations for JSON objects with Gilhari ORM
Gilhari is a Docker-compatible microservice framework that provides RESTful Object-Relational Mapping (ORM) functionality for JSON objects with any relational database.
Remarkably, Gilhari automates REST APIs (POST, GET, PUT, DELETE, etc.) handling, JSON CRUD operations, and database schema setup — no manual coding required.
This repository contains a foundational example showing how to use Gilhari to create a RESTful microservice for persisting JSON objects with basic CRUD operations, filtering, and aggregate queries.
The example uses the base Gilhari docker image (softwaretree/gilhari) to easily create a new docker image (gilhari_example1) that can run as a RESTful microservice (server) to persist app specific JSON objects.
This example can be used standalone as a RESTful microservice or with the ORMCP Server for AI-powered database interactions. This example is particularly useful for getting started with ORMCP.
Related:
- ORMCP Documentation: https://github.com/softwaretree/ormcp-docs - Uses this example as a reference implementation
- ORMCP/Gilhari Examples: https://github.com/SoftwareTree/ormcp-docs/tree/main/examples - Comprehensive list of examples
Note: This example is included in both the Gilhari SDK distribution and the ORMCP Server package.
- If you have the Gilhari SDK installed, you can use it directly from the
examples/gilhari_example1directory - If you have the ORMCP Server package installed, this example is bundled in the distribution
- If accessing from GitHub, you'll need to clone this repository
The example showcases a JSON object model with one type of object: User
Object Model Overview:
- User: Simple user object with basic demographic information
- Attributes: id (int), name (string), age (int), city (string), state (string)
- Database Table: USER
{
"id": 39,
"name": "John39",
"age": 39,
"city": "San Francisco",
"state": "CA"
}Features Demonstrated:
- Basic CRUD operations (Create, Read, Update, Delete)
- Querying with filters (e.g.,
age > 40,state='CA') - Aggregate queries (COUNT, AVG)
- Batch operations (inserting multiple objects at once)
- Advanced projections using
operationDetailsparameter - Object model introspection via
getObjectModelSummaryendpoint
gilhari_example1/
├── src/ # Container domain model classes
│ └── com/softwaretree/... # User.java and base classes
├── config/ # Configuration files
│ ├── gilhari_example1.jdx # ORM specification
│ └── classnames_map_example.json
├── bin/ # Compiled .class files
├── scripts/ # Development scripts
│ └── compile.cmd / .sh # Compiles the container domain model classes
├── gilhari/ # Gilhari microservice (Docker) related files
│ ├── Dockerfile # Docker image definition
│ ├── gilhari_service.config # Service configuration
│ ├── build.cmd / .sh # Builds the Docker image
│ ├── run_docker_app.cmd / .sh # Runs the Docker container
│ ├── curlCommands.cmd / .sh # REST API testing scripts
│ ├── curlCommandsPopulate.cmd / .sh # Populate sample data via the REST API
│ └── connectORMCP.md # Connecting ORMCP Server to this microservice
└── sources.txt # Generated by scripts/compile (not in Git)
All scripts are meant to be run from the project root directory, for example gilhari\build.cmd (Windows) or ./gilhari/build.sh (Linux/Mac). They also work when started from any other directory.
The src directory contains the declarations of the underlying shell (container) classes (e.g., User) that are used to define the object-relational mapping (ORM) specification for the corresponding conceptual domain-specific JSON object model classes:
- User class: Simple shell (container) class (.java file) corresponding to the domain-specific JSON object model class (Container domain model class)
- JDX_JSONObject: Base class of the container domain model classes for handling persistence of domain-specific JSON objects
- Container domain model classes: Only need to define two constructors, with most processing handled by the JDX_JSONObject superclass
Note: Gilhari does not require any explicit programmatic definitions (e.g., ES6 style JavaScript classes) for domain-specific JSON object model classes. It handles the data of domain-specific JSON objects using instances of the container domain model classes and the ORM specification.
A declarative ORM specification for the domain-specific JSON object model classes and their attributes is defined in config/gilhari_example1.jdx using the container domain model classes. This file defines the mappings between JSON objects and database tables.
Key points:
- Update the database URL and JDBC driver in this file according to your setup
- See the JDX_DATABASE and JDBC_DRIVER Specification Guide for configuring different databases
- The container domain model class (User) corresponding to the conceptual domain-specific JSON object model class is defined as a subclass of the JDX_JSONObject class
- Appropriate mappings for the domain-specific JSON object model class are defined in the ORM specification file using the corresponding container domain model class
- The .jdx file includes commented examples showing how to enable auto-increment IDs if desired
For comprehensive details on defining and using container classes and the ORM specification for JSON object models, refer to the "Persisting JSON Objects" section in the JDX User Manual.
The gilhari/Dockerfile builds a RESTful Gilhari microservice using:
- Base Gilhari image (softwaretree/gilhari)
- Compiled domain model (.class) files
- Configuration files including the ORM specification and a JDBC driver
The Docker build context is the project root, so the ADD paths in the Dockerfile (bin, config, gilhari/gilhari_service.config) are relative to the project root.
The gilhari/gilhari_service.config file specifies runtime parameters for the RESTful Gilhari microservice:
{
"gilhari_microservice_name": "gilhari_example1",
"jdx_orm_spec_file": "./config/gilhari_example1.jdx",
"jdbc_driver_path": "/node/node_modules/jdxnode/external_libs/sqlite-jdbc-3.50.3.0.jar",
"jdx_debug_level": 3,
"jdx_force_create_schema": "true",
"jdx_persistent_classes_location": "./bin",
"classnames_map_file": "config/classnames_map_example.json",
"gilhari_rest_server_port": 8081
}The paths in this file (e.g., ./config/..., ./bin) are resolved inside the container, relative to the working directory where the Dockerfile copies bin/ and config/.
| Parameter | Description | Default |
|---|---|---|
gilhari_microservice_name |
Optional name to identify this Gilhari microservice. The name is logged on console during start up | - |
jdx_orm_spec_file |
Location of the ORM specification file containing mapping for persistent classes | - |
jdbc_driver_path |
Path to the JDBC driver (.jar) file. SQLite driver included by default | - |
jdx_debug_level |
Debug output level (0-5). 0 = most verbose, 5 = minimal. Level 3 outputs all SQL statements | 5 |
jdx_force_create_schema |
Whether to recreate database schema on each run. true = useful for development, false = create only once |
false |
jdx_persistent_classes_location |
Root location for compiled persistent (Container domain model) classes. Can be a directory (e.g., ./bin) or a JAR file path. Used as a Java CLASSPATH | - |
classnames_map_file |
Optional JSON file that can map names of container domain model classes to (simpler) object class (type) names (e.g., by omitting a package name) to simplify REST URL | - |
gilhari_rest_server_port |
Port number for the RESTful service. This port number may be mapped to different port number (e.g., 80) by a docker run command. | 8081 |
scripts/compile.cmd/scripts/compile.sh: Compiles the container domain model classessources.txt: Lists the container domain model class source (.java) files for compilation. The compile script generates it from the.javafiles undersrc/, so new classes are picked up automaticallygilhari/build.cmd/gilhari/build.sh: Creates the Gilhari Docker image (gilhari_example1) usinggilhari/Dockerfile
Note: Compilation targets JDK version 1.8, which is compatible with the current Gilhari version.
IMPORTANT: Docker is required for building and running a Gilhari microservice — Get Docker if not already installed on your machine
If you just want to see this example in action without modifications:
- Clone this repository (pre-compiled classes included)
- Install Docker (skip, if already installed)
- Build and run (skip compilation step)
If you want to modify the object model or create your own Gilhari microservices:
- Gilhari SDK: Download and install from https://softwaretree.com
- JX_HOME environment variable: Set to the root directory of your Gilhari SDK installation
- Java Development Kit (JDK 1.8+) for compilation
- Docker installed on your system
Note: The Gilhari SDK contains necessary libraries (JARs) and base classes required for compiling container domain model classes. While pre-compiled .class files are included in this repository for immediate use, you'll need the SDK to make any modifications to the object model or to create your own Gilhari microservices.
Skip compilation and go straight to Docker:
# Windows
gilhari\build.cmd
gilhari\run_docker_app.cmd
# Linux/Mac
./gilhari/build.sh
./gilhari/run_docker_app.shIf you've made changes to the source code:
-
Ensure JX_HOME is set to your Gilhari SDK installation directory (if JX_HOME is not set, the compile script uses
../.., which is the SDK root when this example is in the SDK'sexamplesdirectory) -
Compile the classes:
# Windows scripts\compile.cmd # Linux/Mac ./scripts/compile.sh
-
Build and run the Docker container:
# Windows gilhari\build.cmd gilhari\run_docker_app.cmd # Linux/Mac ./gilhari/build.sh ./gilhari/run_docker_app.sh
Once running, access the Gilhari microservice at:
http://localhost:<port>/gilhari/v1/:className
Example endpoints:
http://localhost:80/gilhari/v1/User
http://localhost:80/gilhari/v1/getObjectModelSummary/now
| Method | Purpose | Example |
|---|---|---|
| GET | Retrieve objects | GET /gilhari/v1/User |
| POST | Create objects | POST /gilhari/v1/User |
| PUT | Update objects | PUT /gilhari/v1/User |
| PATCH | Partial update | PATCH /gilhari/v1/User |
| DELETE | Delete objects | DELETE /gilhari/v1/User |
Get Object Model Summary:
curl -X GET "http://localhost:80/gilhari/v1/getObjectModelSummary/now"Create a User:
curl -X POST http://localhost:80/gilhari/v1/User \
-H "Content-Type: application/json" \
-d '{
"entity": {
"id": 39,
"name": "John39",
"age": 39,
"city": "San Francisco",
"state": "CA"
}
}'Create Multiple Users:
curl -X POST http://localhost:80/gilhari/v1/User \
-H "Content-Type: application/json" \
-d '{
"entity": [
{
"id": 40,
"name": "Mike40",
"age": 40,
"city": "New York",
"state": "NY"
},
{
"id": 41,
"name": "Mary41",
"age": 41,
"city": "Austin",
"state": "TX"
}
]
}'Query with Filter:
curl -X GET "http://localhost:80/gilhari/v1/User?filter=age>40" \
-H "Content-Type: application/json"Count Users in California:
curl -X GET "http://localhost:80/gilhari/v1/User/getAggregate?attribute=id&aggregateType=COUNT&filter=state='CA'" \
-H "Content-Type: application/json"Average Age (California Users):
curl -X GET "http://localhost:80/gilhari/v1/User/getAggregate?attribute=age&aggregateType=AVG&filter=state='CA'" \
-H "Content-Type: application/json"Query with Projections (Selected Attributes Only):
curl -G "http://localhost:80/gilhari/v1/User" \
--data-urlencode "filter=state='CA'" \
--data-urlencode 'operationDetails=[{"opType": "projections", "projectionsDetails": [{"type": "User", "attribs": ["id", "name", "city"]}]}]' \
-H "Content-Type: application/json"Delete Users with Filter:
curl -X DELETE "http://localhost:80/gilhari/v1/User?filter=age>40"Comprehensive test scripts:
-
gilhari/curlCommands.cmd / .sh - Complete demonstration of CRUD operations
Demonstrates:
- Getting object model summary
- Creating single and multiple User objects
- Querying with filters
- Aggregate operations (COUNT, AVG)
- Filtered deletion
- Complete cleanup
-
gilhari/curlCommandsPopulate.cmd / .sh - Data population with advanced queries
Demonstrates:
- Populating sample User data across multiple states
- Advanced filtering (
age>=55) - Aggregate queries for specific groups
- Projections using
operationDetailsparameter for selecting specific attributes
Run the scripts to generate a curl.log file with all responses:
# Windows
gilhari\curlCommands.cmd
gilhari\curlCommandsPopulate.cmd
# Linux/Mac
chmod +x gilhari/*.sh
./gilhari/curlCommands.sh
./gilhari/curlCommandsPopulate.sh
# Custom port
gilhari\curlCommands.cmd 8899
./gilhari/curlCommandsPopulate.sh 8899Each script first calls the health/check endpoint and stops with a message if the microservice is not running.
The scripts will create a curl.log file (in the directory they are run from) with all the API responses, demonstrating User object management with the Gilhari microservice. The populate script also demonstrates how queries for only a selected set of attributes can easily be done using the projections operation type with the operationDetails parameter.
Note: The operationDetails parameter in a query allows you to fine-tune query operations with operational directives similar to GraphQL capabilities. It accepts a JSON array containing one or more operation directives that refine the shape and scope of returned objects. For more details, see the operationDetails Parameter reference.
Other options:
- Postman: Import the endpoints for interactive testing
- Browser: Access GET endpoints directly
- Any REST Client: Standard HTTP methods work with any REST client
- ORMCP Server: Use ORMCP Server tools for AI-powered interactions
This Gilhari microservice can be used with ORMCP Server for AI-powered database interactions. ORMCP Server discovers the object model automatically via the getObjectModelSummary endpoint, so you can query and manipulate User objects in natural language.
See gilhari/connectORMCP.md for step-by-step setup instructions, including a ready-to-use Claude Desktop configuration and sample prompts.
Shell into a running container:
# Find container ID
docker ps
# Access container
docker exec -it <container-id> bashdocker logs <container-id>docker stop <container-id>- JDX User Manual: "Persisting JSON Objects" section for detailed ORM specification documentation
- Gilhari SDK Documentation: The SDK available for download at https://softwaretree.com
- ORMCP Documentation: https://github.com/softwaretree/ormcp-docs
- Database Configuration Guide: JDX_DATABASE and JDBC_DRIVER Specification Guide
- operationDetails Documentation: See the operationDetails Parameter reference in gilhari-docs for GraphQL-like query capabilities
- Connecting ORMCP Server: See gilhari/connectORMCP.md
Script files are provided for both Windows (.cmd) and Linux/Mac (.sh).
Linux/Mac users: Make scripts executable before running:
chmod +x scripts/*.sh gilhari/*.shProblem: Docker image build fails
- Solution: Ensure the base Gilhari image is pulled:
docker pull softwaretree/gilhari
Problem: Compilation errors
- Solution: Verify JDK 1.8+ is installed and JX_HOME environment variable is set correctly
Problem: Port 80 already in use
- Solution: Either stop the service that is already listening on port 80, or change the host port in the
gilhari/run_docker_appscript (e.g.,-p 8080:8081). In the latter case, pass the same port to the curl scripts, for examplegilhari\curlCommands.cmd 8080(Windows) or./gilhari/curlCommands.sh 8080(Linux/Mac)
Problem: Database connection errors
- Solution: Check
config/gilhari_example1.jdxfor the correct database URL and JDBC driver. The URL is used from inside the Docker container, wherelocalhostrefers to the container itself. If your database runs outside the container (on your machine or on another server), specify its host and port in the URL in one of these formats:host.docker.internal:<PortNumber> <DatabaseServer_IP_Address>:<PortNumber>host.docker.internalrefers to the machine running Docker (on Linux, add--add-host=host.docker.internal:host-gatewayto thedocker runcommand ingilhari/run_docker_app). See the JDX_DATABASE and JDBC_DRIVER Specification Guide for examples for different databases.
Problem: Changes to the SQLite database are lost when the container is removed, or the microservice should use (and update) the database file in your project
- Solution:
gilhari/Dockerfilecopies theconfigdirectory, including the SQLite.dbfile, into the image, so each container works on its own copy of the database. Changes survivedocker stop/docker startbut are lost when the container is removed, and they never reach the.dbfile in your project. To use the original file instead, mount the project'sconfigdirectory over the container's copy. From the project root, run:The database URL in the# Windows (Command Prompt) docker run --platform linux/amd64 -p 80:8081 -v "%CD%\config:/opt/gilhari_example1/config" gilhari_example1:1.0 # Linux/Mac docker run --platform linux/amd64 -p 80:8081 -v "$(pwd)/config:/opt/gilhari_example1/config" gilhari_example1:1.0
.jdxfile (jdbc:sqlite:./config/...) is relative to the container's working directory (/opt/gilhari_example1), so it points to the mounted file without any change. Also make sure thatjdx_force_create_schemais"false"ingilhari/gilhari_service.config(if you change it, rebuild the image); otherwise the tables are recreated, and their data erased, each time the microservice starts. Keep a backup copy of the.dbfile, and stop the container before you open or copy the file with other tools.
Problem: ORMCP Server cannot connect to the microservice
- Solution: Ensure the microservice is running and accessible at the configured endpoint. Verify the
getObjectModelSummaryendpoint is reachable
Problem: Projections not working in queries
- Solution: Ensure the
operationDetailsparameter is properly URL-encoded. Use--data-urlencodewith curl or proper encoding in your HTTP client
For issues or questions:
- ORMCP Documentation & Issues: https://github.com/softwaretree/ormcp-docs/issues
- This example: https://github.com/SoftwareTree/gilhari_example1/issues
- Gilhari SDK: Contact support at gilhari_support@softwaretree.com
This example code is licensed under the MIT License - see the LICENSE file for details.
Important: This license applies ONLY to the example code in this repository. The Gilhari software (including the softwaretree/gilhari Docker image and Gilhari SDK) and the embedded JDX ORM software are proprietary products owned by Software Tree.
The Gilhari Docker image includes an evaluation license for testing purposes. For production use or licensing beyond the evaluation period, please visit https://www.softwaretree.com or contact gilhari_support@softwaretree.com.
Ready to try it? Start with the Quick Start section above!