Matchmaker Exchange Reference Server

A simple server that stores patient records and implements the Matchmaker Exchange API.

This is an example implementation, written by the Matchmaker Exchange technical team. The server uses a single elasticsearch instance to index the patient records, the Human Phenotype Ontology, and Ensembl-Entrez-HGNC gene symbol mappings. By default, you can load the MME API benchmark dataset of 50 rare disease patient records compiled from the literature (see the publication for more details).

This code is intended to be illustrative and is not guaranteed to perform well in a production setting.

Dependencies

• Python 2.7 or 3.3+
• ElasticSearch 2.x

Quickstart

1. Clone the repository:

   git clone https://github.com/MatchmakerExchange/reference-server.git
   cd reference-server

2. Install the Python package dependencies (it's recommended that you do this inside a Python virtual environment):

   pip install -e .

3. Start up your elasticsearch server in another shell (see the ElasticSearch instructions for more information).

   ./path/to/elasticsearch

4. Download and index vocabularies and sample data:

   mme-server quickstart

5. Run tests (must run quickstart first):

   mme-server test

6. Authorize an incoming server:

   mme-server clients add myclient --label "My Client" --key "<CLIENT_AUTH_TOKEN>"

   Leave off the --key option to have a secure key randomly generated for you.

7. Start up MME reference server:

   mme-server start

   By default, the server listens globally (--host 0.0.0.0) on port 8000 (--port 8000).

8. Try it out:

   curl -XPOST \
     -H 'X-Auth-Token: <CLIENT_AUTH_TOKEN>' \
     -H 'Content-Type: application/vnd.ga4gh.matchmaker.v1.0+json' \
     -H 'Accept: application/vnd.ga4gh.matchmaker.v1.0+json' \
     -d '{"patient":{
       "id":"1",
       "contact": {"name":"Jane Doe", "href":"mailto:[email protected]"},
       "features":[{"id":"HP:0000522"}],
       "genomicFeatures":[{"gene":{"id":"NGLY1"}}],
       "test": true
     }}' localhost:8000/v1/match

Installation

Your Python environment

It's recommended that you run the server within a Python virtual environment so dependencies are isolated from your system-wide Python installation.

To set up your Python virtual environment:

# Set up virtual environment within a folder '.virtualenv' (add `-p python3` to force python 3)
virtualenv .virtualenv

You can then activate this environment within a particular shell with:

source .virtualenv/bin/activate

ElasticSearch

First, download elasticsearch:

wget https://download.elastic.co/elasticsearch/elasticsearch/elasticsearch-2.4.2.zip
unzip elasticsearch-2.4.2.zip

Then, start up a local elasticsearch cluster to serve as our database (-Des.path.data=data puts the elasticsearch indices in a subdirectory called data):

./elasticsearch-1.7.6/bin/elasticsearch -Des.path.data=data

Loading custom patient data

Custom patient data can be indexed by the server in two ways (if a patient 'id' matches an existing patient, the existing patient is updated):

1. Batch index from the command line:

   mme-server index patients --filename patients.json

2. Batch index from the Python interface:

   >>> from mme_server.backend import get_backend
   >>> db = get_backend()
   >>> patients = db.get_manager('patients')
   >>> patients.index('/path/to/patients.json')

3. Single patient index the Python interface:

   >>> from mme_server.backend import get_backend
   >>> db = get_backend()
   >>> patients = db.get_manager('patients')
   >>> from mme_server.models import Patient
   >>> patient = Patient.from_api({...})
   >>> patients.index_patient(patient)

Questions

If you have any questions, feel free to post an issue on GitHub.

Contributing

This repository is managed by the Matchmaker Exchange technical team. You can reach us via GitHub or by email.

Contributions are most welcome! Post an issue, submit a bugfix, or just try it out. We hope you find it useful.