Python Synapse Client

Branch | Build Status --------|--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- develop | master |

A Python client for Sage Bionetworks' Synapse, a collaborative, open-source research platform that allows teams to share data, track analyses, and collaborate. The Python client can be used as a library for development of software that communicates with Synapse or as a command-line utility.

There is also a Synapse client for R.

Documentation

For more information about the Python client, see:

• Python client API docs

For more information about interacting with Synapse, see:

• Synapse API docs
• Use cases
• Getting Started Guide to Synapse

For release information, see:

• Release notes

Installation

The Python Synapse client has been tested on versions 3.9, 3.10, 3.11, 3.12 and 3.13 on Mac OS X, Ubuntu Linux and Windows.

Starting from Synapse Python client version 3.0, Synapse Python client requires Python >= 3.9

Install using pip

The Python Synapse Client is on PyPI and can be installed with pip:

# Here are a few ways to install the client. Choose the one that fits your use-case
# sudo may optionally be needed depending on your setup

pip install --upgrade synapseclient
pip install --upgrade "synapseclient[pandas]"
pip install --upgrade "synapseclient[pandas, pysftp, boto3]"

...or to upgrade an existing installation of the Synapse client:

# sudo may optionally be needed depending on your setup
pip install --upgrade synapseclient

The dependencies on pandas, pysftp, and boto3 are optional. Synapse Tables integrate with Pandas. The library pysftp is required for users of SFTP file storage. All libraries require native code to be compiled or installed separately from prebuilt binaries.

Install from source

Clone the source code repository.

git clone git://github.com/Sage-Bionetworks/synapsePythonClient.git
cd synapsePythonClient
pip install .

Alternatively, you can use pip to install a particular branch, commit, or other git reference:

pip install git+https://github.com/Sage-Bionetworks/synapsePythonClient@master

or

pip install git+https://github.com/Sage-Bionetworks/synapsePythonClient@my-commit-hash

Command line usage

The Synapse client can be used from the shell command prompt. Valid commands include: query, get, cat, add, update, delete, and onweb. A few examples are shown.

downloading test data from Synapse

synapse -p auth_token get syn1528299

getting help

synapse -h

Note that a Synapse account is required.

Usage as a library

The Synapse client can be used to write software that interacts with the Sage Bionetworks Synapse repository. More examples can be found in the Tutorial section found here

Examples

Log-in and create a Synapse object

``` import synapseclient

syn = synapseclient.Synapse()

You may optionally specify the debug flag to True to print out debug level messages.

A debug level may help point to issues in your own code, or uncover a bug within ours.

syn = synapseclient.Synapse(debug=True)

log in using auth token

syn.login(authToken='auth_token') ```

Sync a local directory to synapse

This is the recommended way of synchronizing more than one file or directory to a synapse project through the use of synapseutils. Using this library allows us to handle scheduling everything required to sync an entire directory tree. Read more about the manifest file format in synapseutils.syncToSynapse ``` import synapseclient import synapseutils import os

syn = synapseclient.Synapse()

log in using auth token

syn.login(authToken='auth_token')

path = os.path.expanduser("~/synapseproject") manifestpath = f"{path}/myprojectmanifest.tsv" project_id = "syn1234"

Create the manifest file on disk

with open(manifest_path, "w", encoding="utf-8") as f: pass

Walk the specified directory tree and create a TSV manifest file

synapseutils.generatesyncmanifest( syn, directorypath=path, parentid=projectid, manifestpath=manifest_path, )

Using the generated manifest file, sync the files to Synapse

synapseutils.syncToSynapse( syn, manifestFile=manifest_path, sendMessages=False, ) ```

Store a Project to Synapse

``` import synapseclient from synapseclient.entity import Project

syn = synapseclient.Synapse()

log in using auth token

syn.login(authToken='auth_token')

project = Project('My uniquely named project') project = syn.store(project)

print(project.id) print(project) ```

Store a Folder to Synapse (Does not upload files within the folder)

``` import synapseclient

syn = synapseclient.Synapse()

log in using auth token

syn.login(authToken='auth_token')

folder = Folder(name='my_folder', parent="syn123") folder = syn.store(folder)

print(folder.id) print(folder)

```

Store a File to Synapse

``` import synapseclient

syn = synapseclient.Synapse()

log in using auth token

syn.login(authToken='auth_token')

file = File( path=filepath, parent="syn123", ) file = syn.store(file)

print(file.id) print(file) ```

Get a data matrix

``` import synapseclient

syn = synapseclient.Synapse()

log in using auth token

syn.login(authToken='auth_token')

retrieve a 100 by 4 matrix

matrix = syn.get('syn1901033')

inspect its properties

print(matrix.name) print(matrix.description) print(matrix.path)

load the data matrix into a dictionary with an entry for each column

with open(matrix.path, 'r') as f: labels = f.readline().strip().split('\t') data = {label: [] for label in labels} for line in f: values = [float(x) for x in line.strip().split('\t')] for i in range(len(labels)): data[labels[i]].append(values[i])

load the data matrix into a numpy array

import numpy as np np.loadtxt(fname=matrix.path, skiprows=1) ```

Authentication

Authentication toward Synapse can be accomplished with the clients using personal access tokens. Learn more about Synapse personal access tokens

Learn about the multiple ways one can login to Synapse.

Synapse Utilities (synapseutils)

The purpose of synapseutils is to create a space filled with convenience functions that includes traversing through large projects, copying entities, recursively downloading files and many more.

Example

import synapseutils
import synapseclient
syn = synapseclient.login()

# copies all Synapse entities to a destination location
synapseutils.copy(syn, "syn1234", destinationId = "syn2345")

# copies the wiki from the entity to a destination entity. Only a project can have sub wiki pages.
synapseutils.copyWiki(syn, "syn1234", destinationId = "syn2345")


# Traverses through Synapse directories, behaves exactly like os.walk()
walkedPath = synapseutils.walk(syn, "syn1234")

for dirpath, dirname, filename in walkedPath:
    print(dirpath)
    print(dirname)
    print(filename)

OpenTelemetry (OTEL)

OpenTelemetry helps support the analysis of traces and spans which can provide insights into latency, errors, and other performance metrics. The synapseclient is ready to provide traces should you want them. The Synapse Python client supports OTLP Exports and can be configured via environment variables as defined here.

Read more about OpenTelemetry in Python here

Quick-start

The following shows an example of setting up jaegertracing via docker and executing a simple python script that implements the Synapse Python client.

Running the jaeger docker container

Start a docker container with the following options: docker run --name jaeger \ -e COLLECTOR_OTLP_ENABLED=true \ -p 16686:16686 \ -p 4318:4318 \ jaegertracing/all-in-one:latest Explanation of ports: * 4318 HTTP port for OTLP data collection * 16686 Jaeger UI for visualizing traces

Once the docker container is running you can access the Jaeger UI via: http://localhost:16686

Environment Variable Configuration

By default, the OTEL exporter sends trace data to http://localhost:4318/v1/traces. You can customize the behavior through environment variables:

• OTEL_SERVICE_NAME: Defines a unique identifier for your application or service in telemetry data (defaults to 'synapseclient'). Set this to a descriptive name that represents your specific implementation, making it easier to filter and analyze traces in your monitoring system.
• OTEL_EXPORTER_OTLP_ENDPOINT: Specifies the destination URL for sending telemetry data (defaults to 'http://localhost:4318'). Configure this to direct data to your preferred OpenTelemetry collector or monitoring service.
• OTEL_DEBUG_CONSOLE: Controls local visibility of telemetry data. Set to 'true' to output trace information to the console, which is useful for development and troubleshooting without an external collector.
• OTEL_SERVICE_INSTANCE_ID: Distinguishes between multiple instances of the same service (e.g., 'prod', 'development', 'local'). This helps identify which specific deployment or environment generated particular traces.
• OTEL_EXPORTER_OTLP_HEADERS: Configures authentication and metadata for telemetry exports. Use this to add API keys, authentication tokens, or custom metadata when sending traces to secured collectors or third-party monitoring services.

Enabling OpenTelemetry in your code

To enable OpenTelemetry with the Synapse Python client, simply call the enable_open_telemetry() method on the Synapse class. Additionally you can access an instance of the OpenTelemetry tracer via the get_tracer() call. This will allow you to create new spans for your code.

```python import synapseclient

Enable OpenTelemetry with default settings

synapseclient.Synapse.enableopentelemetry() tracer = synapseclient.Synapse.get_tracer()

Then create and use the Synapse client as usual

with tracer.startascurrentspan("myfunctionspan"): syn = synapseclient.Synapse() syn.login(authToken='authtoken') ```

Advanced Configuration

You can pass additional resource attributes to enable_open_telemetry():

```python import synapseclient

Enable with custom resource attributes

synapseclient.Synapse.enableopentelemetry( resourceattributes={ "deployment.environment": "development", "service.version": "1.2.3", # Overrides the `OTELSERVICENAMEenvironment variable "service.instance.id": "4.5.6", # Overrides theOTELSERVICEINSTANCEIDenvironment variable "custom.attribute": "value" } ) ``

When OpenTelemetry is enabled in the Synapse client, the following happens automatically:

1. Instrumentation is set up for:

   • Threading (via ThreadingInstrumentor): Ensures proper context propagation across threads, which is essential for maintaining trace continuity in multi-threaded applications
   • HTTP libraries:
     • requests (via RequestsInstrumentor): Captures all HTTP requests made using the requests library, including methods, URLs, status codes, and timing information
     • httpx (via HTTPXClientInstrumentor): Tracks both synchronous and asynchronous HTTP requests made with the httpx library
     • urllib (via URLLibInstrumentor): Monitors lower-level HTTP operations made directly with Python's standard library
   • Each instrumented HTTP library includes custom hooks that extract Synapse entity IDs from URLs when possible and add them as span attributes

2. Traces are configured to collect spans across your application:

   • Spans automatically capture operation duration, status, and errors.
   • An attribute propagation mechanism ensures that certain attributes (like synapse.transfer.direction and synapse.operation.category) are properly passed to child spans for uploads/downloads.
   • Trace data is exported via OTLP (OpenTelemetry Protocol).

3. Resource information is automatically added to your traces, including:

   • Python version
   • OS type
   • Synapse client version
   • Service name (defaults to "synapseclient" but can be customized via environment variables)
   • Service instance ID

Note that once enabled, OpenTelemetry cannot be disabled in the same process - you would need to restart your Python interpreter to disable it.

License and Copyright

© Copyright 2013-25 Sage Bionetworks

This software is licensed under the Apache License, Version 2.0.