bioblend

NAME

bioblend - BioBlend Documentation

ABOUT

BioBlend is a Python library for interacting with the Galaxy API. BioBlend is supported and tested on:
BioBlend's goal is to make it easier to script and automate the running of Galaxy analyses and administering of a Galaxy server. In practice, it makes it possible to do things like this:

About the library name

The library was originally called just Blend but we renamed it to reflect more of its domain and a make it bit more unique so it can be easier to find. The name was intended to be short and easily pronounceable. In its original implementation, the goal was to provide a lot more support for CloudMan and other integration capabilities, allowing them to be blended together via code. BioBlend fitted the bill.

INSTALLATION

Stable releases of BioBlend are best installed via pip from PyPI:
Alternatively, the most current source code from our Git repository can be installed with:
After installing the library, you will be able to simply import it into your Python environment with import bioblend. For details on the available functionality, see the API documentation. BioBlend requires a number of Python libraries. These libraries are installed automatically when BioBlend itself is installed, regardless whether it is installed via PyPi or by running python3 setup.py install command. The current list of required libraries is always available from setup.py in the source code repository. If you also want to run tests locally, some extra libraries are required. To install them, run:

USAGE

To get started using BioBlend, install the library as described above. Once the library becomes available on the given system, it can be developed against. The developed scripts do not need to reside in any particular location on the system. It is probably best to take a look at the example scripts in docs/examples source directory and browse the API documentation. Beyond that, it's up to your creativity :).

DEVELOPMENT

Anyone interested in contributing or tweaking the library is more then welcome to do so. To start, simply fork the Git repository on Github and start playing with it. Then, issue pull requests.

API DOCUMENTATION

BioBlend's API focuses around and matches the services it wraps. Thus, there are two top-level sets of APIs, each corresponding to a separate service and a corresponding step in the automation process. Note that each of the service APIs can be used completely independently of one another. Effort has been made to keep the structure and naming of those API's consistent across the library but because they do bridge different services, some discrepancies may exist. Feel free to point those out and/or provide fixes. For Galaxy, an alternative object-oriented API is also available. This API provides an explicit modeling of server-side Galaxy instances and their relationships, providing higher-level methods to perform operations such as retrieving all datasets for a given history, etc. Note that, at the moment, the oo API is still incomplete, providing access to a more restricted set of Galaxy modules with respect to the standard one.

Galaxy API

API used to manipulate genomic analyses within Galaxy, including data management and workflow execution.

API documentation for interacting with Galaxy

GalaxyInstance

----

Config

Contains possible interaction dealing with Galaxy configuration.
----

Datasets

----

Dataset collections

----

Datatypes

----

Folders

----

Forms

----

FTP files

----

Genomes

Groups

----

Histories

----

Invocations

----

Jobs

----

Libraries

----

Quotas

----

Roles

----

Tools

----

Tool data tables

----

Tool dependencies

----

ToolShed

----

Users

----

Visual

----

Workflows

Object-oriented Galaxy API

Client

Wrappers

Usage documentation

This page describes some sample use cases for the Galaxy API and provides examples for these API calls. In addition to this page, there are functional examples of complete scripts in the docs/examples directory of the BioBlend source code repository.

Connect to a Galaxy server

To connect to a running Galaxy server, you will need an account on that Galaxy instance and an API key for the account. Instructions on getting an API key can be found at https://galaxyproject.org/develop/api/ . To open a connection call:
We now have a GalaxyInstance object which allows us to interact with the Galaxy server under our account, and access our data. If the account is a Galaxy admin account we also will be able to use this connection to carry out admin actions.

View Histories and Datasets

Methods for accessing histories and datasets are grouped under GalaxyInstance.histories.* and GalaxyInstance.datasets.* respectively. To get information on the Histories currently in your account, call:
This returns a list of dictionaries containing basic metadata, including the id and name of each History. In this case, we have two existing Histories in our account, 'RNAseq_DGE_BASIC_Prep' and 'June demo'. To get more detailed information about a History we can pass its id to the show_history method:
This gives us a dictionary containing the History's metadata. With contents=False (the default), we only get a list of ids of the datasets contained within the History; with contents=True we would get metadata on each dataset. We can also directly access more detailed information on a particular dataset by passing its id to the show_dataset method:

Uploading Datasets to a History

To upload a local file to a Galaxy server, you can run the upload_file method, supplying the path to a local file:
If files are greater than 2GB in size, they will need to be uploaded via FTP. Importing files from the user's FTP folder can be done via running the upload tool again:

View Data Libraries

Methods for accessing Data Libraries are grouped under GalaxyInstance.libraries.*. Most Data Library methods are available to all users, but as only administrators can create new Data Libraries within Galaxy, the create_folder and create_library methods can only be called using an API key belonging to an admin account. We can view the Data Libraries available to our account using:
This gives a list of metadata dictionaries with basic information on each library. We can get more information on a particular Data Library by passing its id to the show_library method:

Upload files to a Data Library

We can get files into Data Libraries in several ways: by uploading from our local machine, by retrieving from a URL, by passing the new file content directly into the method, or by importing a file from the filesystem on the Galaxy server. For instance, to upload a file from our machine we might call:

>>> gi.libraries.upload_file_from_local_path('8e6f930d00d123ea', '/local/path/to/mydata.fastq', file_type='fastqsanger')

Note that we have provided the id of the destination Data Library, and in this case we have specified the type that Galaxy should assign to the new dataset. The default value for file_type is 'auto', in which case Galaxy will attempt to guess the dataset type.

View Workflows

Methods for accessing workflows are grouped under GalaxyInstance.workflows.*. To get information on the Workflows currently in your account, use:
This returns a list of metadata dictionaries. We can get the details of a particular Workflow, including its steps, by passing its id to the show_workflow method:

Export or import a workflow

Workflows can be exported from or imported into Galaxy. This makes it possible to archive workflows, or to move them between Galaxy instances. To export a workflow, we can call:
This gives us a complex dictionary representing the workflow. We can import this dictionary as a new workflow with:
This call returns a dictionary containing basic metadata on the new workflow. Since in this case we have imported the dictionary into the original Galaxy instance, we now have a duplicate of the original workflow in our account:

>>> gi.workflows.get_workflows()
[{'id': 'c0bacafdfe211f9a',
  'name': 'TopHat + cufflinks part 1 (imported from API)',
  'url': '/api/workflows/c0bacafdfe211f9a'},
 {'id': 'e8b85ad72aefca86',
  'name': 'TopHat + cufflinks part 1',
  'url': '/api/workflows/e8b85ad72aefca86'},
 {'id': 'b0631c44aa74526d',
  'name': 'CuffDiff',
  'url': '/api/workflows/b0631c44aa74526d'}]

Instead of using dictionaries directly, workflows can be exported to or imported from files on the local disk using the export_workflow_to_local_path and import_workflow_from_local_path methods. See the API reference for details. NOTE:

Invoke a workflow

To invoke a workflow, we need to tell Galaxy which datasets to use for which workflow inputs. We can use datasets from histories or data libraries. Examine the workflow above. We can see that it takes only one input file. That is:

>>> wf = gi.workflows.show_workflow('e8b85ad72aefca86')
>>> wf['inputs']
{'252': {'label': 'Input RNA-seq fastq', 'value': ''}}

There is one input, labelled 'Input RNA-seq fastq'. This input is passed to the Tophat tool and should be a fastq file. We will use the dataset we examined above, under View Histories and Datasets, which had name 'C1_R2_1.chr4.fq' and id '10a4b652da44e82a'. To specify the inputs, we build a data map and pass this to the invoke_workflow method. This data map is a nested dictionary object which maps inputs to datasets. We call:
In this case the only input id is '252' and the corresponding dataset id is '10a4b652da44e82a'. We have specified the dataset source to be 'hda' (HistoryDatasetAssociation) since the dataset is stored in a History. See the API reference for allowed dataset specifications. We have also requested that a new History be created and used to store the results of the run, by setting history_name='New output history'. The invoke_workflow call submits all the jobs which need to be run to the Galaxy workflow engine, with the appropriate dependencies so that they will run in order. The call returns immediately, so we can continue to submit new jobs while waiting for this workflow to execute. invoke_workflow returns the a dictionary describing the workflow invocation. If we view the output History immediately after calling invoke_workflow, we will see something like:
In this case, because the submitted jobs have not had time to run, the output History contains 8 datasets in the 'queued' state and has a total size of 0 bytes. If we make this call again later we should instead see completed output files.

View Users

Methods for managing users are grouped under GalaxyInstance.users.*. User management is only available to Galaxy administrators, that is, the API key used to connect to Galaxy must be that of an admin account. To get a list of users, call:

>>> gi.users.get_users()
[{'email': '[email protected]',
  'id': '975a9ce09b49502a',
  'quota_percent': None,
  'url': '/api/users/975a9ce09b49502a'},
 {'email': '[email protected]',
  'id': '0193a95acf427d2c',
  'quota_percent': None,
  'url': '/api/users/0193a95acf427d2c'}]

Using BioBlend for raw API calls

BioBlend can be used to make HTTP requests to the Galaxy API in a more convenient way than using e.g. the requests Python library. There are 5 available methods corresponding to the most common HTTP methods: make_get_request, make_post_request, make_put_request, make_delete_request and make_patch_request. One advantage of using these methods is that the API keys stored in the GalaxyInstance object is automatically added to the request. To make a GET request to the Galaxy API with BioBlend, call:

>>> gi.make_get_request(gi.base_url + "/api/version").json()
{'version_major': '19.05',
 'extra': {}}

To make a POST request to the Galaxy API with BioBlend, call:

>>> gi.make_post_request(gi.base_url + "/api/histories", payload={"name": "test history"})
{'importable': False,
 'create_time': '2019-07-05T20:10:04.823716',
 'contents_url': '/api/histories/a77b3f95070d689a/contents',
 'id': 'a77b3f95070d689a',
 'size': 0, 'user_id': '5b732999121d4593',
 'username_and_slug': None,
 'annotation': None,
 'state_details': {'discarded': 0,
                   'ok': 0,
                   'failed_metadata': 0,
                   'upload': 0,
                   'paused': 0,
                   'running': 0,
                   'setting_metadata': 0,
                   'error': 0,
                   'new': 0,
                   'queued': 0,
                   'empty': 0},
 'state': 'new',
 'empty': True,
 'update_time': '2019-07-05T20:10:04.823742',
 'tags': [],
 'deleted': False,
 'genome_build': None,
 'slug': None,
 'name': 'test history',
 'url': '/api/histories/a77b3f95070d689a',
 'state_ids': {'discarded': [],
               'ok': [],
               'failed_metadata': [],
               'upload': [],
               'paused': [],
               'running': [],
               'setting_metadata': [],
               'error': [],
               'new': [],
               'queued': [],
               'empty': []},
 'published': False,
 'model_class': 'History',
 'purged': False}

Toolshed API

API used to interact with the Galaxy Toolshed, including repository management.

API documentation for interacting with the Galaxy Toolshed

ToolShedInstance

Categories

Repositories

Tools

CONFIGURATION

BioBlend allows library-wide configuration to be set in external files. These configuration files can be used to specify access keys, for example.

Configuration documents for BioBlend

BioBlend

Config

TESTING

If you would like to do more than just a mock test, you need to point BioBlend to an instance of Galaxy. Do so by exporting the following two variables:
The unit tests, stored in the tests folder, can be run using pytest. From the project root:

GETTING HELP

If you have run into issues, found a bug, or can't seem to find an answer to your question regarding the use and functionality of BioBlend, please use the Github Issues page to ask your question.

RELATED DOCUMENTATION

Links to other documentation and libraries relevant to this library:

INDICES AND TABLES

AUTHOR

Enis Afgan

COPYRIGHT

2012-2022, Enis Afgan