omnigraph

Unified Python interface for multiple graph databases

PyPi
GitHub
Code
Docs
Cite

Cite as

If you use pyomnigraph in your research, please cite it via its Zenodo concept DOI (which always resolves to the latest version):

Fahl, W. pyomnigraph — Unified Python interface for multiple graph databases. Zenodo. https://doi.org/10.5281/zenodo.21721772

Machine-readable metadata is available in CITATION.cff; GitHub's "Cite this repository" button and Zenodo pick it up automatically. For a specific release, use the version DOI shown on the corresponding Zenodo record.

Docs and Tutorials

Wiki

Motivation

The graph database landscape is fragmented, with each triple store having its own APIs, deployment methods, and operational quirks. Developers and researchers working with RDF data often need to:

• Switch between different triple stores for performance comparisons
• Migrate data from one system to another
• Test the same queries across multiple backends
• Deploy applications that work with various graph databases

This leads to:

• ❌ Duplicated effort writing database-specific code
• ❌ Vendor lock-in making migrations difficult
• ❌ Inconsistent interfaces slowing development
• ❌ Manual deployment processes for each database

pyomnigraph solves this by providing:

• ✅ Unified API - Same Python interface for all supported databases
• ✅ Standardized deployment - Consistent Docker-based setup
• ✅ Easy switching - Change backends with a single parameter
• ✅ Comparative testing - Run identical operations across multiple stores
• ✅ Simplified management - Start, stop, load data with simple commands

Supported Triple Stores

Whether you're building a semantic web application, conducting research, or evaluating different triple stores, pyomnigraph eliminates the complexity of working with multiple graph database systems.

Examples

omnigraph --list --include-inactive --doc-format github

Server Management

# Start specific servers
omnigraph -s jena --cmd start

# Restart sequence - stop remove and start
omnigraph -s jena --cmd stop rm start

# Start all configured servers
omnigraph -s all --cmd start

# Check server status
omnigraph -s blazegraph --cmd status

# Open web ui
omnigraph -s jena --cmd webui

Data Operations

# Load datasets
omnigraph -s blazegraph --cmd load

# Get triple count
omnigraph -s blazegraph --cmd count

# Use test environment
omnigraph --test -s blazegraph --cmd start load

Usage

omnigraph command line

omnigraph -h
usage: omnigraph [-h] [-a] [-d] [-ds DATASETS [DATASETS ...]] [-dc DATASETS_CONFIG] [-f] [-r {turtle,rdf-xml,n3,json-ld}] [-q] [-V] [--apache APACHE]
                 [-c CONFIG] [--cmd CMD [CMD ...]] [-df DOC_FORMAT] [-l] [--test] [-s SERVERS [SERVERS ...]] [-v]

Unified Python interface for multiple graph databases

options:
  -h, --help            show this help message and exit
  -a, --about           show about info [default: False]
  -d, --debug           show debug info [default: False]
  -ds DATASETS [DATASETS ...], --datasets DATASETS [DATASETS ...]
                        datasets to work with - all is an alias for all datasets [default: ['wikidata_triplestores']]
  -dc DATASETS_CONFIG, --datasets-config DATASETS_CONFIG
                        Path to datasets configuration YAML file [default: /Users/wf/Library/Python/3.12/lib/python/site-
                        packages/omnigraph/resources/examples/datasets.yaml]
  -f, --force           force actions that would modify existing data [default: False]
  -r {turtle,rdf-xml,n3,json-ld}, --rdf_format {turtle,rdf-xml,n3,json-ld}
                        RDF format to use [default: turtle]
  -q, --quiet           avoid any output [default: False]
  -V, --version         show program's version number and exit
  --apache APACHE       create apache configuration file for the given server(s)
  -c CONFIG, --config CONFIG
                        Path to server configuration YAML file [default: /Users/wf/Library/Python/3.12/lib/python/site-
                        packages/omnigraph/resources/examples/servers.yaml]
  --cmd CMD [CMD ...]   commands to execute on servers: bash, clear, count, info, load, logs, needed, rm, start, status, stop, webui
  -df DOC_FORMAT, --doc-format DOC_FORMAT
                        The document format to use [default: plain]
  -l, --list-servers    List available servers [default: False]
  --test                use test environment [default: False]
  -s SERVERS [SERVERS ...], --servers SERVERS [SERVERS ...]
                        servers to work with - 'all' selects all configured servers [default: ['blazegraph']]
  -v, --verbose         show verbose output [default: False]

rdfdump command line

rdfdump -h
usage: rdfdump [-h] [-a] [-d] [-ds DATASETS [DATASETS ...]] [-dc DATASETS_CONFIG] [-f] [-r {turtle,rdf-xml,n3,json-ld}] [-q] [-V] [--limit LIMIT] [-l]
               [--count] [--dump] [-4o] [--max-count MAX_COUNT] [--no-progress] [--output-path OUTPUT_PATH] [--tryit]

Unified Python interface for multiple graph databases

options:
  -h, --help            show this help message and exit
  -a, --about           show about info [default: False]
  -d, --debug           show debug info [default: False]
  -ds DATASETS [DATASETS ...], --datasets DATASETS [DATASETS ...]
                        datasets to work with - all is an alias for all datasets [default: ['wikidata_triplestores']]
  -dc DATASETS_CONFIG, --datasets-config DATASETS_CONFIG
                        Path to datasets configuration YAML file [default: /Users/wf/Library/Python/3.12/lib/python/site-
                        packages/omnigraph/resources/examples/datasets.yaml]
  -f, --force           force actions that would modify existing data [default: False]
  -r {turtle,rdf-xml,n3,json-ld}, --rdf_format {turtle,rdf-xml,n3,json-ld}
                        RDF format to use [default: turtle]
  -q, --quiet           avoid any output [default: False]
  -V, --version         show program's version number and exit
  --limit LIMIT         Number of triples per request [default: 10000]
  -l, --list            List available datasets [default: False]
  --count               List available datasets with triple counts[default: False]
  --dump                perform the dump [default: False]
  -4o, --for-omnigraph  store dump at default omnigraph location [default: False]
  --max-count MAX_COUNT
                        Maximum number of solutions/triples to download (uses dataset expected_solutions if not specified)
  --no-progress         Disable progress bar
  --output-path OUTPUT_PATH
                        Path for dump files
  --tryit               open the try it! URL [default: False]