Welcome to pytest-wdl’s documentation!¶

User manual¶

pytest-wdl is a plugin for the pytest unit testing framework that enables testing of workflows written in Workflow Description Language. Test workflow inputs and expected outputs are configured in a test_data.json file. Workflows are run by one or more executors. By default, actual and expected outputs are compared by MD5 hash, but data type-specific comparisons are provided. Data types and executors are pluggable and can be provided via third-party packages.

Dependencies¶

Python 3.6+
At least one of the supported workflow engines (see Limitations for known limitations of these workflow engines):
- Miniwdl (v0.6.4 is automatically installed as a dependency of pytest-wdl)
- Cromwell JAR file (pytest-wdl is currently tested with Cromwell v45)
- dxWDL JAR file (pytest-wdl is currently tested with dxWDL v1.42)
Java-based workflow engines (e.g. Cromwell and dxWDL) require a Java runtime (typically 1.8+)
If your WDL tasks depend on Docker images, make sure to have the Docker daemon running

Other python dependencies are installed when you install the library.

Installation¶

Install from PyPI¶

$ pip install pytest-wdl

Install from source¶

You can to clone the repository and install:

$ make install

Or use pip to install from github:

$ pip install git+https://github.com/elilillyco/pytest-wdl.git

Install optional dependencies¶

Some optional features of pytest-wdl have additional dependencies that are loaded on-demand.

The plugins that have extra dependencies are:

dx: Support for DNAnexus file storage, and for the dxWDL executor.
bam: More intelligent comparison of expected and actual BAM file outputs of a workflow than just comparing MD5 checksums.
progress: Show progress bars when downloading remote files.

To install a plugin’s dependencies:

$ pip install pytest-wdl[<plugin>]

To do this locally, you can clone the repo and run:

$ pip install -e .[<data_type>]

To install pytest-wdl and all extras dependencies:

$ pip install pytest-wdl[all]

Quick Start¶

The full code for this quick-start project is available in the examples directory.

To demonstrate how to use pytest-wdl, we will create a simple variant calling workflow and tests with the following project structure:

quickstart
|_variant_caller.wdl
|_tests
  |_data
  | |_NA12878.chr22.tiny.vcf
  |_test_data.json
  |_test_variant_caller.py

Our workflow is shown below. It requires a BAM file and a reference sequence, it calls a task to perform variant calling using Freebayes, and it returns a VCF file with the called variants.

version 1.0

struct Reference {
  File fasta
  String organism
}

workflow call_variants {
  input {
    File bam
    Reference ref
  }

  call freebayes {
    input:
      bam=bam,
      index=ref
  }
  
  output {
    File vcf = freebayes.vcf
  }
}

task freebayes {...}

Now you want to test that your workflow runs successfully. You also want to test that your workflow always produces the same output with a given input. You can accomplish both of these objectives with the following test function, which is in the tests/test_variant_caller.py file:

def test_variant_caller(workflow_data, workflow_runner):
    inputs = {
        "bam": workflow_data["bam"],
        "ref": {
            "fasta": workflow_data["reference_fa"],
            "organism": "human"
        }
    }
    expected = workflow_data.get_dict("vcf")
    workflow_runner(
        "variant_caller.wdl",
        inputs,
        expected
    )

This test executes a workflow with the specified inputs, and will compare the outputs to the specified expected outputs. The workflow_data and workflow_runner parameters are fixtures that are injected by the pytest framework at runtime. The workflow_data fixture provides the test data files based on the following configuration in the tests/test_data.json file. Note that some of the data files are stored remotely (in this case, in the GitHub repository for FreeBayes) and are downloaded automatically when needed. Later in the mnaual you’ll find descriptions of these fixtures and details on how to configure your tests’ inputs, exected outputs, and other parameters.

{
  "bam": {
    "url": "https://.../NA12878.chr22.tiny.bam"
  },
  "reference_fa": {
    "url": "https://.../q.fa"
  },
  "vcf": {
    "name": "NA12878.chr22.tiny.vcf",
    "type": "vcf"
  }
}

To run this test, make sure you have pytest-wdl installed and run the following command from within the project folder. The default executor (miniwdl) will be used to execute the workflow. You can configure any of the other supported executors by creating a configuration file.

$ pytest -s -vv --show-capture=all

This test should run successfully, and you should see output that looks like this:

_images/output.png

Project setup¶

pytest-wdl should support most project set-ups, including:

# simple
myproject
|_workflow.wdl
|_subworkflow1.wdl
|_subworkflow2.wdl
|_tests
  |_test_workflow.py
  |_test_data.json

# multi-module with single test directory
myproject
|_main_workflow.wdl
|_module1
| |_module1.wdl
|_module2
| |_module2.wdl
|_tests
 |_main
 | |_test_main.py
 | |_test_data.json
 |_module1
   |_test_module1.py
   |_test_data.json
 ...

# multi-module with separate test directories
myproject
|_main.wdl
|_module1
| |_module1.wdl
| |_tests
|   |_test_module1.py
|   |_test_data.json
|_module2
| |_...
|_tests
  |_test_main.py
  |_test_data.json

By default, pytest-wdl tries to find the files it is expecting relative to one of two directories:

Project root: the base directory of the project. In the above examples, myproject is the project root directory. By default, the project root is discovered by looking for key files (e.g. setup.py), starting from the directory in which pytest is executing the current test. In most cases, the project root will be the same for all tests executed within a project.
Test context directory: starting from the directory in which pytest is executing the current test, the test context directory is the first directory up in the directory hierarchy that contains a “tests” subdirectory. The test context directory may differ between test modules, depending on the setup of your project:
- In the “simple” and “multi-module with single test directory” examples, myproject would be the test context directory
- In the “multi-module with separate test directories” example, the test context directory would be myproject when executing myproject/tests/test_main.py and module1 when executing myproject/module1/tests/test_module1.py.

Fixtures¶

All functionality of pytest-wdl is provided via fixtures. As long as pytest-wdl is in your PYTHONPATH, its fixtures will be discovered and made available when you run pytest.

The two most important fixtures are:

workflow_data:+ Provides access to data files for use as inputs to a workflow, and for comparing to workflow output.
workflow_runner: Given a WDL workflow, inputs, and expected outputs, runs the workflow using one or more executors and compares actual and expected outputs.

There are also several additional fixtures used for configuration of the two main fixtures. In most cases, the default values returned by these fixtures “just work.” However, if you need to override the defaults, you may do so either directly within your test modules, or in a conftest.py file.

Test data¶

Typically, workflows require inputs and generate outputs. Beyond simply ensuring that a workflow runs successfully, we often want to additionally test that it reproducibly generates the same results given the same inputs.

Test inputs and outputs are configured in a test_data.json file that is stored in the same directory as the test module. This file has one entry for each input/output. Primitive types map as expected from JSON to Python to WDL. Object types (e.g. structs) have a special syntax. For example, the following test_data.json file defines an integer input that is loaded as a Python int and then maps to the WDL Integer type when passed as an input parameter to a workflow, and an object tyep that is loaded as a Python dict and then maps to a user-defined type (struct) in WDL:

{
  "input_int": 42,
  "input_obj": {
    "class": "Person",
    "value": {
      "name": "Joe",
      "age": 42 
    }
  }
}

Files¶

For file inputs and outputs, pytest-wdl offers several different options. Test data files may be located remotely (identified by a URL), located within the test directory (using the folder hierarchy established by the datadir-ng plugin), located at an arbitrary local path, or defined by specifying the file contents directly within the JSON file. Files that do not already exist locally are localized on-demand and stored in the cache directory.

Some additional options are available only for expected outputs, in order to specify how they should be compared to the actual outputs.

File data can be defined the same as object data (i.e. “file” is a special class of object type):

{
  "config": {
    "class": "file",
    "value": {
      "path": "config.json"
    }
  }
}

As a short-cut, the “class” attribute can be omitted and the map describing the file provided directly as the value. Below is an example test_data.json file that demonstrates different ways to define input and output data files:

{
  "bam": {
    "url": "http://example.com/my.bam",
    "http_headers": {
      "auth_token": "TOKEN"
    },
    "digests": {
      "md5": "8db3048a86e16a08d2d8341d1c72fecb"
    }
  },
  "reference": {
    "path": "${REFERENCE_DIR}/chr22.fa"
  },
  "sample": {
    "path": "samples.vcf",
    "contents": "sample1\nsample2"
  },
  "output_vcf": {
    "name": "output.vcf",
    "type": {
      "name": "vcf",
      "allowed_diff_lines": 2
    }
  }
}

The available keys for configuring file inputs/outputs are:

name: Filename to use when localizing the file; when none of url, path, or contents are defined, name is also used to search for the data file within the tests directory, using the same directory structure defined by the datadir-ng fixture.
path: The local path to the file. If the path does not already exist, the file will be localized to this path. Typically, this is defined as a relative path that will be prefixed with the cache directory path. Environment variables can be used to enable the user to configure an environment-specific path.
env: The name of an environment variable in which to look up the local path of the file.
url: A URL that can be resolved by urllib.
- http_headers: Optional dict mapping header names to values. These headers are used for file download requests. Keys are header names and values are either strings (environment variable name) or mappings with the following keys:
  - env: The name of an environment variable in which to look up the header value.
  - value: The header value; only used if an environment variable is not specified or is unset.
contents: The contents of the file, specified as a string. The file is written to path the first time it is requested.
digests: Optional mapping of hash algorithm name to digest. These are digests that have been computed on the remote file and are used to validate the downloaded file. Currently only used for files resolved from URLs.

In addition, the following keys are recognized for output files only:

type: The file type. This is optional and only needs to be provided for certain types of files that are handled specially for the sake of comparison. The value can also be a hash with required key “name” and any other comparison-related attributes (including data type-specific attributes).
allowed_diff_lines: Optional and only used for outputs comparison. If ‘0’ or not specified, it is assumed that the expected and actual outputs are identical.

URL Schemes¶

pytest_wdl uses urllib, which by default supports http, https, and ftp. If you need to support alternate URL schemes, you can do so via a plugin. Currently, the following plugins are avaiable:

dx (DNAnexus): requires the dxpy module

Data Types¶

When comparing actual and expected outputs, the “type” of the expected output is used to determine how the files are compared. If no type is specified, then the type is assumed to be “default.”

default¶

The default type if one is not specified.

It can handle raw text files, as well as gzip compressed files.
If allowed_diff_lines is 0 or not specified, then the files are compared by their MD5 hashes.
If allowed_diff_lines is > 0, the files are converted to text and compared using the linux diff tool.

vcf¶

During comparison, headers are ignored, as are the QUAL, INFO, and FORMAT columns; for sample columns, only the first sample column is compared between files, and only the genotype values for that sample.
Optional attributes:
- compare_phase: Whether to compare genotype phase; defaults to False.

bam*:¶

BAM is converted to SAM.
Replaces random UNSET-\w*\b type IDs that samtools often adds.
One comparison is performed using all rows and a subset of columns that are expected to be invariate. Rows are sorted by name and then by flag.
A second comparison is performed using all columns and a subset of rows based on filtering criteria. Rows are sorted by coordinate and then by name.
Optional attributes:
- min_mapq: The minimum MAPQ when filtering rows for the second comparison.
- compare_tag_columns: Whether to include tag columns (12+) when comparing all columns in the second comparison.

* requires extra dependencies to be installed, see Install Optional Dependencies

Executors¶

An Executor is a wrapper around a WDL workflow execution engine that prepares inputs, runs the tool, captures outputs, and handles errors. Currently, the following executors are supported (but aternative executors can be implemented as plugins):

The workflow_runner fixture is a callable that runs the workflow using the executor.

The first argument to workflow_runner is the path to the WDL file. It is required. It may be an absolute or a relative path; if the latter, it is first searched relative to the current tests directory (i.e. test_context_dir/tests), and then the project root.

The remaining arguments to workflow_runner are optional:

inputs: Dict that will be serialized to JSON and provided to the executor as the workflow inputs. If not specified, the workflow must not have any required inputs.
expected: Dict mapping output parameter names to expected values. Any workflow outputs that are not specified are ignored. This parameter can be omitted if, for example, you only want to test that the workflow completes successfully.
workflow_name: The name of the workflow to execute in the WDL script. If not specified, the name of the workflow is extracted from the WDL file.
inputs_file: Specify the inputs.json file to use, or the path to the inputs.json file to write, instead of a temp file.
imports_file: Specify the imports file to use. By default, all WDL files under the test context directory are imported if an import_paths.txt file is not provided.

You can also pass executor-specific keyword arguments.

Executor-specific `workflow_runner` arguments¶

Cromwell¶

imports_file: Instead of specifying the imports file, this can also be the path to the imports zip file to write, instead of a temp file.
java_args: Override the default Java arguments.
cromwell_args: Override the default Cromwell arguments.

Miniwdl¶

task_name: Name of the task to run, e.g. for a WDL file that does not have a workflow. This takes precedence over workflow_name.

dxWDL¶

Requires the dxpy package to be installed.

project_id: ID of the project where the workflow will be built. Defaults to the currently selected project. You can also specify different projects for workflows and data using workflow_project_id and data_project_id varaibles.
folder: The folder within the project where the workflow will be built. Defaults to ‘/’. You can also specify different folders for workflows and data using workflow_folder_id and data_folder_id varaibles.
stage_id: Stage ID to use when inputs don’t come prefixed with the stage. Defaults to “stage-common”.
force: Boolean; whether to force the workflow to be built even if the WDL file has not changed since the last build.
archive: Boolean; whether to archive existing applets/workflows (True) or overwrite them (False) when building the workflow. Defaults to True.
extras: Extras file to use when building the workflow.

Known limitations¶

Cromwell¶

Cromwell issues are tracked in the Broad’s Jira.

miniwdl¶

pytest-wdl is currently pinned to miniwdl version 0.6.4, which has the following known limitations:
- Task input files are mounted read-only by default; commands to rename or remove them can succeed only with –copy-input-files
See the miniwdl open issues for other potential limitations

dxWDL¶

DNAnexus (and thus the dxWDL executor) does not support optional collection types (e.g. Array[String]?, Map[String, File]?).
See the dxWDL open issues for other potential limitations

Configuration¶

pytest-wdl has two levels of configuration:

Project-specific configuration, which generally deals with the structure of the project, and may require customization if the structure of your project differs substantially from what is expected, but also encompases executor-specific configuration.
Environment-specific configuration, which generally deals with idiosyncrasies of the local environment.

Project-specific configuration¶

Configuration at the project level is handled by overriding fixtures, either in the test module or in a top-level conftest.py file. The following fixtures may be overridden:

fixture name	scope	description	default
`project_root_files`	module	List of filenames that are found in the project root directory.	`["setup.py", "pyproject.toml", ".git"]`
`project_root`	module	The root directory of the project. Relative paths are relative to this directory, unless specified otherwise.	Starting in the current test directory/module, scan up the directory hierarchy until one of the `project_root_files` are located.
`workflow_data_descriptor_file`	module	Path to the JSON file that describes the test data.	`tests/test_data.json`
`workflow_data_descriptors`	module	Mapping of workflow input/output names to values (as described in the Files section).	Loaded from the `workflow_data_descriptor_file`
`workflow_data_resolver`	module	Provides the `DataResolver` object that resolves test data; this should only need to be overridden for testing/debugging purposes	`DataResolver` created from `workflow_data_descriptors`
`import_paths`	module	Provides the path to the file that lists the directories from which to import WDL dependencies	"import_paths.txt"
`import_dirs`	module	Provides the directories from which to import WDL dependencies	Loaded from `import_paths` file, if any, otherwise all WDL files under the current test context directory are imported
`default_executors`	session	Specify the default set of executors to use when running tests	`user_config.default_executors`

Environment-specific configuration¶

There are several aspects of pytest-wdl that can be configured to the local environment, for example to enable the same tests to run both on a user’s development machine and in a continuous integration environment.

Environment-specific configuration is specified either or both of two places: a JSON configuration file and environment variables. Environment variables always take precendence over values in the configuration file. Keep in mind that (on a *nix system) environment variables can be set (semi-)permanently (using export) or temporarily (using env):

# Set environment variable durably
$ export FOO=bar

# Set environment variable only in the context of a single command
$ env FOO=bar echo "foo is $FOO"

Configuration file¶

The pytest-wdl configuration file is a JSON-format file. Its default location is $HOME/.pytest_wdl_config.json. To get started, you can copy one of the following example config files and modify as necessary:

simple: Uses only the miniwdl executor
more complex: Uses both miniwdl and Cromwell; shows how to configure proxies and headers for accessing remote data files in a private repository

The available configuration options are listed in the following table:

configuration file key	environment variable	description	default	recommendation
`cache_dir`	`PYTEST_WDL_CACHE_DIR`	Directory to use for localizing test data files.	Temporary directory; a separate directory is used for each test module	pro: saves time when multiple tests rely on the same test data files; con: can cause conflicts, if tests use different files with the same name
`execution_dir`	`PYTEST_WDL_EXECUTION_DIR`	Directory in which tests are executed	Temporary directory; a separate directory is used for each test function	Only use for debugging; use an absolute path
`proxies`	Configurable	Proxy server information; see details below	None	Use environment variable(s) to configure your proxy server(s), if any
`http_headers`	Configurable	HTTP header configuration that applies to all URLs matching a given pattern; see details below	None	Configure headers by URL pattern; configure headers for specific URLs in the test_data.json file
`show_progress`	N/A	Whether to show progress bars when downloading files	False
`default_executors`	PYTEST_WDL_EXECUTORS	Comma-delimited list of executor names to run by default	["cromwell"]
`executors`	Executor-dependent	Configuration options specific to each executor; see below	None
`providers`	Provider-dependent	Configuration options specific to each provider; see below	None
N/A	`LOGLEVEL`	Level of detail to log; can set to 'DEBUG', 'INFO', 'WARNING', or 'ERROR'	'WARNING'	Use 'DEBUG' when developing plugins/fixtures/etc., otherwise 'WARNING'

Proxies¶

In the proxies section of the configuration file, you can define the proxy servers for schemes used in data file URLs. The keys are scheme names and the values are either strings - environment variable names - or mappings with the following keys:

env: The name of an environment variable in which to look for the proxy server address.
value: The value to use for the proxy server address, if the environment variable is not defined or is unset.

{
  "proxies": {
    "http": {
      "env": "HTTP_PROXY"
    },
    "https": {
      "value": "https://foo.com/proxy",
      "env": "HTTPS_PROXY"
    }
  }
}

HTTP(S) Headers¶

In the http_headers section of the configuration file, you can define a list of headers to use when downloading data files. In addition to env and value keys (which are interpreted the same as for proxies, two additional keys are allowed:

name: Required; the header name
pattern: A regular expression used to match the URL; if not specified, the header is used with all URLs.

{
  "http_headers": [
    {
      "name": "X-JFrog-Art-Api",
      "pattern": "http://my.company.com/artifactory/*",
      "env": "TOKEN"
    }
  ]
}

Executor-specific configuration¶

Java-based Executors¶

These options apply to all Java-based executors (currently Cromwell and dxWDL):

configuration file key	environment variable	description	default
`java_bin`	`JAVA_HOME`	Path to java executable; If not specified, then Java executable is expected to be in $JAVA_HOME/bin/java	None
`java_args`	`JAVA_ARGS`	Arguments to add to the `java` command	`-Dconfig.file=<cromwell_config_file>` (for Cromwell executor, if `cromwell_config_file` is specified)
N/A	`CLASSPATH`	Java classpath; searched for a file matching "cromwell*.jar" if `cromwell_jar` is not specified	None

Cromwell¶

configuration file key	environment variable	description	default
`cromwell_jar_file`	`CROMWELL_JAR`	Path to Cromwell JAR file	None
`cromwell_config_file`	`CROMWELL_CONFIG`	Path to Cromwell configuration file	None
`cromwell_args`	`CROMWELL_ARGS`	Arguments to add to the `cromwell run` command	None;

Note that if you are doing your development locally and using Docker images you’ve built yourself, it is recommended to add -Ddocker.hash-lookup.enabled=false to java_args to disable Docker lookup by hash. Otherwise, you must push your Docker image(s) to a remote repository (e.g. DockerHub) before running your tests.

dxWDL¶

The dxWDL executor (as well as URLs using the dx:// scheme) require you to be logged into DNAnexus. You can configure either a username and password or an auth token in the config file to log in automatically (see provider configuration), otherwise you will be asked to log in interactively.

configuration file key	environment variable	description	default
`dxwdl_jar_file`	`DXWDL_JAR`	Path to dxWDL JAR file	None
`dxwdl_cache_dir`	`DXWDL_CACHE_DIR`	Directory that should be used to cache downloaded results	A temporary directory is used and deleted after each test

Provider-specific configuration¶

A “provider” is a remote (generally cloud-based) service that provides both an execution engine and data storage.

DNAnexus¶

configuration file key	environment variable	description	default
`dx_username`	None	Username to use for logging into DNAnexus if the user is not already logged in	None
`dx_password`	None	Password to use for logging into DNAnexus if the user is not already logged in	None
`dx_token`	None	Token to use for logging into DNAnexus if the user is not already logged in (mutually exclusive with username/password	None

Fixtures¶

There are two fixtures that control the loading of the user configuration:

fixture name	scope	description	default
`user_config_file`	session	The location of the user configuration file	The value of the `PYTEST_WDL_CONFIG` environment variable if set, otherwise `$HOME/.pytest_wdl_config.json`
`user_config`	session	Provides a `UserConfiguration` object that is used by other fixtures to access configuration values	Default values are loaded from `user_config_file`, but most values can be overridden via environment variables (see Configuration)

Plugins¶

pytest-wdl provides the ability to implement plugins for data types, executors, and url schemes. When two plugins with the same name are present, the third-party plugin takes precedence over the built-in plugin (however, if there are two conflicting third-party plugins, an exception is raised).

Creating new data types¶

To create a new data type plugin, add a module in the data_types package of pytest-wdl, or create it in your own 3rd party package.

Your plugin should subclass the pytest_wdl.data_types.DataFile class and override its methods for _assert_contents_equal() and/or _diff() to define the behavior for this file type.

Next, add an entry point in setup.py. If the data type requires more dependencies to be installed, make sure to use a try/except ImportError and raise a PluginError with an informative message, and to add the extra dependencies under the setup.py’s extras_require. For example:

# plugin.py
from pytest_wdl.plugins import PluginError

try:
    import mylib
except ImportError as err:
    raise PluginError(
        "mytype is not available because the mylib library is not "
        "installed"
    ) from err

from setuptools import setup

setup(
    ...,
    entry_points={
        "pytest_wdl.data_types": [
            "mydata = pytest_wdl.data_types.mytype:MyDataFile"
        ]
    },
    extras_require={
        "mydata": ["mylib"]
    }
)

In this example, the extra dependencies can be installed with pip install pytest-wdl[mydata].

Creating new executors¶

To create a new executor, add a module in the executors package, or in your own 3rd party package.

Your plugin should subclass pytest_wdl.executors.Executor and implement the run_workflow() method.

Next, add an entry point in setup.py. If the executor requires more dependencies to be installed, make sure to use a try/except ImportError to warn about this and add the extra dependencies under the setup.py’s extras_require (see example under Creating new data types). For example:

from setuptools import setup

setup(
    ...,
    entry_points={
        "pytest_wdl.executors": [
            "myexec = pytest_wdl.executors.myexec:MyExecutor"
        ]
    },
    extras_require={
        "myexec": ["mylib"]
    }
)

Supporting alternative URL schemes¶

If you want to use test data files that are available via a service that does not support http/https/ftp downloads, you can implement a custom URL scheme.

Your plugin should subclass pytest_wdl.url_schemes.UrlScheme and implement the scheme, handles, and any of the urlopen, request, and response methods that are required.

Next, add an entry point in setup.py. If the schem requires more dependencies to be installed, make sure to use a try/except ImportError to warn about this and add the extra dependencies under the setup.py’s extras_require (see example under Creating new data types). For example:

from setuptools import setup

setup(
    ...,
    entry_points={
        "pytest_wdl.url_schemes": [
            "myexec = pytest_wdl.url_schemes.myscheme:MyUrlScheme"
        ]
    },
    extras_require={
        "myexec": ["mylib"]
    }
)

pytest_wdl¶

pytest_wdl package¶

Subpackages¶

pytest_wdl.data_types package¶

Submodules¶

pytest_wdl.data_types.bam module¶

Convert BAM to SAM for diff.

class pytest_wdl.data_types.bam.BamDataFile(local_path: pathlib.Path, localizer: Optional[pytest_wdl.localizers.Localizer] = None, **compare_opts)[source]¶

Bases: pytest_wdl.data_types.DataFile

Supports comparing output of BAM file. This uses pysam to convert BAM to SAM, so that DataFile can carry out a regular diff on the SAM files.

class pytest_wdl.data_types.bam.Sorting[source]¶

Bases: enum.Enum

An enumeration.

COORDINATE = 1¶

NAME = 2¶

NONE = 0¶

pytest_wdl.data_types.bam.assert_bam_files_equal(file1: pathlib.Path, file2: pathlib.Path, allowed_diff_lines: int = 0, min_mapq: int = 0, compare_tag_columns: bool = False)[source]¶

Compare two BAM files: * Convert them to SAM format * Optionally re-sort the files by chromosome, position, and flag * First compare all lines using only a subset of columns that should be deterministic * Next, filter the files by MAPQ and compare the remaining rows using all columns

Parameters

file1 – First BAM to compare
file2 – Second BAM to compare
allowed_diff_lines – Number of lines by which the BAMs are allowed to differ (after being convert to SAM)
min_mapq – Minimum mapq used to filter reads when comparing all columns
compare_tag_columns – Whether to include tag columns (12+) when comparing all columns

pytest_wdl.data_types.bam.bam_to_sam(input_bam: pathlib.Path, output_sam: pathlib.Path, headers: Optional[Iterable[str]] = ('HD', 'SQ', 'RG'), min_mapq: Optional[int] = None, sorting: pytest_wdl.data_types.bam.Sorting = <Sorting.NONE: 0>)[source]¶: Use PySAM to convert bam to sam.

pytest_wdl.data_types.bam.diff_bam_columns(file1: pathlib.Path, file2: pathlib.Path, columns: str) → int[source]¶

pytest_wdl.data_types.json module¶

class pytest_wdl.data_types.json.JsonDataFile(local_path: pathlib.Path, localizer: Optional[pytest_wdl.localizers.Localizer] = None, **compare_opts)[source]¶: Bases: pytest_wdl.data_types.DataFile

pytest_wdl.data_types.vcf module¶

Some tools that generate VCF (callers) will result in very slightly different qual scores and other floating-point-valued fields when run on different hardware. This handler ignores the QUAL and INFO columns and only compares the genotype (GT) field of sample columns. Only works for single-sample VCFs.

class pytest_wdl.data_types.vcf.VcfDataFile(local_path: pathlib.Path, localizer: Optional[pytest_wdl.localizers.Localizer] = None, **compare_opts)[source]¶: Bases: pytest_wdl.data_types.DataFile

pytest_wdl.data_types.vcf.diff_vcf_columns(file1: pathlib.Path, file2: pathlib.Path, compare_phase: bool = False) → int[source]¶

Module contents¶

class pytest_wdl.data_types.DataFile(local_path: pathlib.Path, localizer: Optional[pytest_wdl.localizers.Localizer] = None, **compare_opts)[source]¶

Bases: object

A data file, which may be local, remote, or represented as a string.

Parameters

local_path – Path where the data file should exist after being localized.
localizer – Localizer object, for persisting the file on the local disk.
allowed_diff_lines – Number of lines by which the file is allowed to differ from another and still be considered equal.
compare_opts – Additional type-specific comparison options.

assert_contents_equal(other: Union[str, pathlib.Path, DataFile]) → None[source]¶

Assert the contents of two files are equal.

If allowed_diff_lines == 0, files are compared using MD5 hashes, otherwise their contents are compared using the linux diff command.

Parameters: other – A DataFile or string file path.
Raises: AssertionError if the files are different. –

property path¶

set_compare_opts(**kwargs)[source]¶

Update comparison options.

Parameters: **kwargs – Comparison options to update.

class pytest_wdl.data_types.DefaultDataFile(local_path: pathlib.Path, localizer: Optional[pytest_wdl.localizers.Localizer] = None, **compare_opts)[source]¶: Bases: pytest_wdl.data_types.DataFile

pytest_wdl.data_types.assert_binary_files_equal(file1: pathlib.Path, file2: pathlib.Path, digest: str = 'md5') → None[source]¶

pytest_wdl.data_types.assert_text_files_equal(file1: pathlib.Path, file2: pathlib.Path, allowed_diff_lines: int = 0, diff_fn: Callable[[pathlib.Path, pathlib.Path], int] = <function diff_default>) → None[source]¶

pytest_wdl.data_types.compare_gzip(file1: pathlib.Path, file2: pathlib.Path)[source]¶

pytest_wdl.data_types.diff_default(file1: pathlib.Path, file2: pathlib.Path) → int[source]¶

Default diff command.

Parameters

file1 – First file to compare
file2 – Second file to compare

Returns

Number of different lines.

pytest_wdl.executors package¶

Submodules¶

pytest_wdl.executors.cromwell module¶

class pytest_wdl.executors.cromwell.CromwellExecutor(import_dirs: Optional[Sequence[pathlib.Path]] = None, java_bin: Union[str, pathlib.Path, None] = None, java_args: Optional[str] = None, cromwell_jar_file: Union[str, pathlib.Path, None] = None, cromwell_config_file: Union[str, pathlib.Path, None] = None, cromwell_args: Optional[str] = None)[source]¶

Bases: pytest_wdl.executors.JavaExecutor

Manages the running of WDL workflows using Cromwell.

Parameters

import_dirs – Relative or absolute paths to directories containing WDL scripts that should be available as imports.
java_bin – Path to the java executable.
java_args – Default Java arguments to use; can be overidden by passing java_args=… to run_workflow.
cromwell_jar_file – Path to the Cromwell JAR file.
cromwell_args – Default Cromwell arguments to use; can be overridden by passing cromwell_args=… to run_workflow.

run_workflow(wdl_path: pathlib.Path, inputs: Optional[dict] = None, expected: Optional[dict] = None, **kwargs) → dict[source]¶

Run a WDL workflow on given inputs, and check that the output matches given expected values.

Parameters

wdl_path – The WDL script to execute.
inputs – Object that will be serialized to JSON and provided to Cromwell as the workflow inputs.
expected – Dict mapping output parameter names to expected values.
kwargs –
Additional keyword arguments, mostly for debugging: * workflow_name: The name of the workflow in the WDL script. If None,

the name of the WDL script is used (without the .wdl extension).
- inputs_file: Path to the Cromwell inputs file to use. Inputs are
  written to this file only if it doesn’t exist.
- imports_file: Path to the WDL imports file to use. Imports are
  written to this file only if it doesn’t exist.
- java_args: Additional arguments to pass to Java runtime.
- cromwell_args: Additional arguments to pass to cromwell run.

Returns

Dict of outputs.

Raises

ExecutionFailedError – if there was an error executing Cromwell
AssertionError – if the actual outputs don’t match the expected outputs

class pytest_wdl.executors.cromwell.Failures(num_failed: int, failed_task: str, failed_task_exit_status: Optional[str] = None, failed_task_stdout: Union[str, pathlib.Path, None] = None, failed_task_stderr: Union[str, pathlib.Path, None] = None)[source]¶

Bases: object

property failed_task_stderr¶

property failed_task_stdout¶

pytest_wdl.executors.miniwdl module¶

class pytest_wdl.executors.miniwdl.MiniwdlExecutor(import_dirs: Optional[Sequence[pathlib.Path]] = None)[source]¶

Bases: pytest_wdl.executors.Executor

Manages the running of WDL workflows using Cromwell.

static log_source(logger: logging.Logger, exn: Exception)[source]¶

static read_miniwdl_command_std(path: Optional[str] = None) → Optional[str][source]¶

run_workflow(wdl_path: pathlib.Path, inputs: Optional[dict] = None, expected: Optional[dict] = None, **kwargs) → dict[source]¶

Run a WDL workflow on given inputs, and check that the output matches given expected values.

Parameters

wdl_path – The WDL script to execute.
inputs – Object that will be serialized to JSON and provided to Cromwell as the workflow inputs.
expected – Dict mapping output parameter names to expected values.
kwargs –
Additional keyword arguments, mostly for debugging: * workflow_name: Name of the workflow to run. * task_name: Name of the task to run if a workflow isn’t defined. * inputs_file: Path to the Cromwell inputs file to use. Inputs are

written to this file only if it doesn’t exist.

Returns

Dict of outputs.

Raises

Exception – if there was an error executing Cromwell
AssertionError – if the actual outputs don’t match the expected outputs

Module contents¶

exception pytest_wdl.executors.ExecutionFailedError(executor: str, target: str, status: str, inputs: Optional[dict] = None, executor_stdout: Optional[str] = None, executor_stderr: Optional[str] = None, failed_task: Optional[str] = None, failed_task_exit_status: Optional[int] = None, failed_task_stdout: Optional[str] = None, failed_task_stderr: Optional[str] = None, msg: Optional[str] = None)[source]¶

Bases: pytest_wdl.executors.ExecutorError

property exit_status_str¶

class pytest_wdl.executors.Executor[source]¶

Bases: object

Base class for WDL workflow executors.

abstract run_workflow(wdl_path: pathlib.Path, inputs: Optional[dict] = None, expected: Optional[dict] = None, **kwargs) → dict[source]¶

Run a WDL workflow on given inputs, and check that the output matches given expected values.

Parameters

wdl_path – The WDL script to execute.
inputs – Object that will be serialized to JSON and provided to Cromwell as the workflow inputs.
expected – Dict mapping output parameter names to expected values.
kwargs – Additional executor-specific keyword arguments (mostly for debugging)

Returns

Dict of outputs.

Raises

ExecutionFailedError – if there was an error executing the workflow
AssertionError – if the actual outputs don’t match the expected outputs

exception pytest_wdl.executors.ExecutorError(executor: str, msg: Optional[str] = None)[source]¶: Bases: Exception

class pytest_wdl.executors.InputsFormatter[source]¶

Bases: object

format_inputs(inputs_dict: dict, namespace: Optional[str] = None) → dict[source]¶

format_value(value: Any) → Any[source]¶

Convert a primitive, DataFile, Sequence, or Dict to a JSON-serializable object. Currently, arbitrary objects can be serialized by implementing an as_dict() method, otherwise they are converted to strings.

Parameters: value – The value to format.
Returns: The serializable value.

classmethod get_instance() → pytest_wdl.executors.InputsFormatter[source]¶

class pytest_wdl.executors.JavaExecutor(java_bin: Union[str, pathlib.Path, None] = None, java_args: Optional[str] = None)[source]¶

Bases: pytest_wdl.executors.Executor

Manages the running of WDL workflows using a Java-based executor.

Parameters

java_bin – Path to the java executable.
java_args – Default Java arguments to use; can be overidden by passing java_args=… to run_workflow.

static resolve_jar_file(file_name_pattern: str, jar_path: Optional[pathlib.Path] = None, env_var: Optional[str] = None)[source]¶

pytest_wdl.executors.get_target_name(wdl_path: Optional[pathlib.Path] = None, wdl_doc: Optional[WDL.Tree.Document] = None, task_name: Optional[str] = None, workflow_name: Optional[str] = None, **kwargs) → Tuple[str, bool][source]¶

Get the execution target. The order of priority is:

task_name
workflow_name
wdl_doc.workflow.name
wdl_doc.task[0].name
wdl_file.stem

Parameters

wdl_path – Path to a WDL file
wdl_doc – A miniwdl-parsed WDL document
task_name – The task name
workflow_name – The workflow name
**kwargs – Additional keyword arguments to pass to parse_wdl

Returns

A tuple (target, is_task), where is_task is a boolean indicating whether the target is a task (True) or a workflow (False).

Raises

ValueError if 1) neither task_name nor workflow_name is specified and the –
WDL document contains no workflow and multiple tasks; or 2) all of the –
parameters are None. –

pytest_wdl.executors.parse_wdl(wdl_path: pathlib.Path, import_dirs: Optional[Sequence[pathlib.Path]] = (), check_quant: bool = False, **_) → WDL.Tree.Document[source]¶

pytest_wdl.executors.read_write_inputs(inputs_file: Union[str, pathlib.Path, None] = None, inputs_dict: Optional[dict] = None, inputs_formatter: Optional[pytest_wdl.executors.InputsFormatter] = <pytest_wdl.executors.InputsFormatter object>, **kwargs) → Tuple[dict, Optional[pathlib.Path]][source]¶

If inputs_file is specified and it exists, read its contents. Otherwise, if inputs_dict is specified, format it using inputs_formatter (if specified) and write it to inputs_file or a temporary file.

Parameters

inputs_file –
inputs_dict –
inputs_formatter –
kwargs –

Returns

The (formatted) inputs dict and the resolved inputs file. If both inputs_dict and inputs_file are None, returns ({}, None).

pytest_wdl.providers package¶

Submodules¶

pytest_wdl.providers.dx module¶

class pytest_wdl.providers.dx.DxInputsFormatter(wdl_doc: WDL.Tree.Document, project_id: str = None, data_project_id: Optional[str] = None, folder: str = '/', data_folder: Optional[str] = None, **_)[source]¶

Bases: object

format_inputs(inputs_dict: dict, namespace: Optional[str] = None) → dict[source]¶

format_value(value: Any, path: Tuple[str, ...], nested: bool = False) → Tuple[Any, bool][source]¶

Convert a primitive, DataFile, Sequence, or Dict to a JSON-serializable object. Currently, arbitrary objects can be serialized by implementing an as_dict() method, otherwise they are converted to strings.

Parameters

value – The value to format.
path – The path to the current value.
nested – Whether the value is nested within a complex type

Returns

The tuple (val, is_complex), where val is serializable value and is_complex is True if the value is a complex type.

class pytest_wdl.providers.dx.DxResponse(file_id: str, project_id: Optional[str] = None)[source]¶

Bases: pytest_wdl.url_schemes.Response

download_file(destination: pathlib.Path, show_progress: bool = False, digests: Optional[dict] = None)[source]¶

Download a file to a specific destination.

Parameters

destination – Destination path
show_progress – Whether to show a progress bar
digests – Optional dict mapping hash names to digests. These are used to validate the downloaded file.

Raises

DigestsNotEqualError –

class pytest_wdl.providers.dx.DxUrlHandler[source]¶

Bases: pytest_wdl.url_schemes.UrlHandler

property handles¶

property scheme¶

urlopen(request: urllib.request.Request) → pytest_wdl.url_schemes.Response[source]¶

class pytest_wdl.providers.dx.DxWdlExecutor(import_dirs: Optional[Sequence[pathlib.Path]] = None, java_bin: Union[str, pathlib.Path, None] = None, java_args: Optional[str] = None, dxwdl_jar_file: Union[str, pathlib.Path, None] = None, dxwdl_cache_dir: Union[str, pathlib.Path, None] = None)[source]¶

Bases: pytest_wdl.executors.JavaExecutor

run_workflow(wdl_path: pathlib.Path, inputs: Optional[dict] = None, expected: Optional[dict] = None, **kwargs) → dict[source]¶

Run a WDL workflow on given inputs, and check that the output matches given expected values.

Parameters

wdl_path – The WDL script to execute.
inputs – Object that will be serialized to JSON and provided to Cromwell as the workflow inputs.
expected – Dict mapping output parameter names to expected values.
kwargs – Additional executor-specific keyword arguments (mostly for debugging)

Returns

Dict of outputs.

Raises

ExecutionFailedError – if there was an error executing the workflow
AssertionError – if the actual outputs don’t match the expected outputs

pytest_wdl.providers.dx.login(logout: bool = False, interactive: bool = True)[source]¶

Checks that the user is logged into DNAnexus, otherwise log them in.

Parameters

logout – Whether to log out before exiting the context. Ignored if the user is already logged in.
interactive – Whether to allow interactive login.

Raises

PluginError – if interactive is False and the user cannot be logged in non-interactively

Module contents¶

Submodules¶

pytest_wdl.config module¶

class pytest_wdl.config.UserConfiguration(config_file: Optional[pathlib.Path] = None, cache_dir: Optional[pathlib.Path] = None, remove_cache_dir: Optional[bool] = None, execution_dir: Optional[pathlib.Path] = None, proxies: Optional[Dict[str, Union[str, Dict[str, str]]]] = None, http_headers: Optional[List[dict]] = None, show_progress: Optional[bool] = None, executors: Optional[str] = None, executor_defaults: Optional[Dict[str, dict]] = None, provider_defaults: Optional[Dict[str, dict]] = None)[source]¶

Bases: object

Stores pytest-wdl configuration. If configuration options are specified both in the config file and as arguments to the constructor, the latter take precedence.

Parameters

config_file – JSON file from which to load default values.
cache_dir – The directory in which to cache localized files; defaults to using a temporary directory that is specific to each module and deleted afterwards.
remove_cache_dir – Whether to remove the cache directory; if None, takes the value True if a temp directory is used for caching, and False, if a value for cache_dir is specified.
execution_dir – The directory in which to run workflows. Defaults to None, which signals that a different temporary directory should be used for each workflow run.
proxies – Mapping of proxy type (typically ‘http’ or ‘https’ to either an environment variable, or a dict with either/both keys ‘env’ and ‘value’, where the value is taken from the environment variable (‘env’) first, and from ‘value’ if the environment variable is not specified or is unset.
http_headers – A list of dicts, each of which defines a header. The allowed keys are ‘pattern’, ‘name’, ‘env’, and ‘value’, where pattern is a URL pattern to match, ‘name’ is the header name and ‘env’ and ‘value’ are interpreted the same as for proxies. If no pattern is provided, the header is used for all URLs.
show_progress – Whether to show progress bars when downloading remote test data files.
executors – Default set of executors to run.
executor_defaults – Mapping of executor name to dict of executor-specific configuration options.

cleanup() → None[source]¶: Preforms cleanup operations, such as deleting the cache directory if self.remove_cache_dir is True.

get_executor_defaults(executor_name: str) → dict[source]¶

Get default configuration values for the given executor.

Parameters: executor_name – The executor name
Returns: A dict with the executor configuration values, if any.

get_provider_defaults(provider_name: str) → dict[source]¶

Get default configuration values for the given provider.

Parameters: provider_name – The provider name
Returns: A dict with the provider configuration values, if any.

pytest_wdl.config.cleanup()[source]¶

pytest_wdl.config.default_user_config_file() → pathlib.Path[source]¶

pytest_wdl.config.get_instance() → pytest_wdl.config.UserConfiguration[source]¶

pytest_wdl.config.set_instance(config: Optional[pytest_wdl.config.UserConfiguration] = None, path: Optional[pathlib.Path] = None)[source]¶

pytest_wdl.core module¶

pytest_wdl.core.DATA_TYPES = {'bam': <pytest_wdl.plugins.PluginFactory object>, 'json': <pytest_wdl.plugins.PluginFactory object>, 'vcf': <pytest_wdl.plugins.PluginFactory object>}¶: Data type plugin modules from the discovered entry points.

class pytest_wdl.core.DataDirs(basedir: pathlib.Path, module, function: Callable, cls: Optional[Type] = None)[source]¶

Bases: object

Provides data files from test data directory structure as defined by the datadir and datadir-ng plugins. Paths are resolved lazily upon first request.

property paths¶

class pytest_wdl.core.DataManager(data_resolver: pytest_wdl.core.DataResolver, datadirs: pytest_wdl.core.DataDirs)[source]¶

Bases: object

Manages test data, which is defined in a test_data.json file.

Parameters

data_resolver – Module-level config.
datadirs – Data directories to search for the data file.

get_dict(*names: str, **params) → dict[source]¶

Creates a dict with one or more entries from this DataManager.

Parameters

*names – Names of test data entries to add to the dict.
**params – Mapping of workflow parameter names to test data entry names.

Returns

Dict mapping parameter names to test data entries for all specified names.

get_list(*names: str) → list[source]¶

class pytest_wdl.core.DataResolver(data_descriptors: dict, user_config: pytest_wdl.config.UserConfiguration)[source]¶

Bases: object

Resolves data files that may need to be localized.

resolve(name: str, datadirs: Optional[pytest_wdl.core.DataDirs] = None)[source]¶

pytest_wdl.core.EXECUTORS = {'cromwell': <pytest_wdl.plugins.PluginFactory object>, 'dxwdl': <pytest_wdl.plugins.PluginFactory object>, 'miniwdl': <pytest_wdl.plugins.PluginFactory object>}¶: Executor plugin modules from the discovered entry points.

pytest_wdl.core.create_data_file(user_config: pytest_wdl.config.UserConfiguration, type: Union[str, dict, None] = 'default', name: Optional[str] = None, path: Union[str, pathlib.Path, None] = None, url: Optional[str] = None, contents: Union[str, dict, None] = None, env: Optional[str] = None, http_headers: Optional[dict] = None, digests: Optional[dict] = None, datadirs: Optional[pytest_wdl.core.DataDirs] = None, **kwargs) → pytest_wdl.data_types.DataFile[source]¶

pytest_wdl.core.create_executor(executor_name: str, import_dirs: Sequence[pathlib.Path], user_config: pytest_wdl.config.UserConfiguration)[source]¶

pytest_wdl.fixtures module¶

Fixtures for writing tests that execute WDL workflows using Cromwell.

Note: This library is being transitioned to python3 only, and to use `pathlib.Path`s instead of string paths. For backward compatibility fixtures that produce a path may still return string paths, but this support will be dropped in a future version.

pytest_wdl.fixtures.default_executors(user_config: pytest_wdl.config.UserConfiguration) → Sequence[str][source]¶

pytest_wdl.fixtures.import_dirs(request: _pytest.fixtures.FixtureRequest, project_root: Union[str, pathlib.Path], import_paths: Union[str, pathlib.Path, None]) → List[Union[str, pathlib.Path]][source]¶

Fixture that provides a list of directories containing WDL scripts to make avaialble as imports. Uses the file provided by import_paths fixture if it is not None, otherwise returns a list containing the parent directory of the test module.

Parameters

request – A FixtureRequest object
project_root – Project root directory
import_paths – File listing paths to imports, one per line

pytest_wdl.fixtures.import_paths(request: _pytest.fixtures.FixtureRequest) → Union[str, pathlib.Path, None][source]¶: Fixture that provides the path to a file that lists directories containing WDL scripts to make available as imports. This looks for the file at “tests/import_paths.txt” by default, and returns None if that file doesn’t exist.

pytest_wdl.fixtures.project_root(request: _pytest.fixtures.FixtureRequest, project_root_files: List[str]) → Union[str, pathlib.Path][source]¶: Fixture that provides the root directory of the project. By default, this assumes that the project has one subdirectory per task, and that this framework is being run from the test subdirectory of a task diretory, and therefore looks for the project root two directories up.

pytest_wdl.fixtures.project_root_files() → List[str][source]¶: Fixture that provides a list of filenames that are found in the project root directory. Used by the project_root fixture to locate the project root directory.

pytest_wdl.fixtures.user_config(user_config_file: Optional[pathlib.Path]) → pytest_wdl.config.UserConfiguration[source]¶

pytest_wdl.fixtures.user_config_file() → Optional[pathlib.Path][source]¶

Fixture that provides the value of ‘user_config’ environment variable. If not specified, looks in the default location ($HOME/pytest_user_config.json).

Returns: Path to the confif file, or None if not specified.

pytest_wdl.fixtures.workflow_data(request: _pytest.fixtures.FixtureRequest, workflow_data_resolver: pytest_wdl.core.DataResolver) → pytest_wdl.core.DataManager[source]¶

Provides an accessor for test data files, which may be local or in a remote repository.

Parameters

request – FixtureRequest object
workflow_data_resolver – Module-level test data configuration

Examples

def workflow_data_descriptor_file():: return “tests/test_data.json”
def test_workflow(workflow_data):: print(workflow_data[“myfile”])

pytest_wdl.fixtures.workflow_data_descriptor_file(request: _pytest.fixtures.FixtureRequest) → Union[str, pathlib.Path][source]¶

Fixture that provides the path to the JSON file that describes test data files.

Parameters: request – A FixtureRequest object

pytest_wdl.fixtures.workflow_data_descriptors(request: _pytest.fixtures.FixtureRequest, project_root: Union[str, pathlib.Path], workflow_data_descriptor_file: Union[str, pathlib.Path]) → dict[source]¶

Fixture that provides a mapping of test data names to values. If workflow_data_descriptor_file is relative, it is searched first relative to the current test context directory and then relative to the project root.

Parameters: workflow_data_descriptor_file – Path to the data descriptor JSON file.
Returns: A dict with keys as test data names and each value either a primitive, a map describing a data file, or a DataFile object.

pytest_wdl.fixtures.workflow_data_resolver(workflow_data_descriptors: dict, user_config: pytest_wdl.config.UserConfiguration) → pytest_wdl.core.DataResolver[source]¶

Provides access to test data files for tests in a module.

Parameters

workflow_data_descriptors – workflow_data_descriptors fixture.
user_config –

pytest_wdl.fixtures.workflow_runner(request: _pytest.fixtures.FixtureRequest, project_root: Union[str, pathlib.Path], import_dirs: List[Union[str, pathlib.Path]], user_config: pytest_wdl.config.UserConfiguration, default_executors: Sequence[str], subtests: pytest_subtests.SubTests)[source]¶

Provides a callable that runs a workflow. The callable has the same signature as Executor.run_workflow, but takes an additional keyword argument executors, a sequence of strings, which allows overriding the names of the executors to use.

If multiple executors are specified, the tests are run using the subtests fixture of the pytest-subtests plugin.

Parameters

request – A FixtureRequest object.
project_root – Project root directory.
import_dirs – Directories from which to import WDL scripts.
user_config – A UserConfiguration object.
default_executors – Names of executors to use when executor name isn’t passed to the workflow_runner callable.
subtests – A SubTests object.

pytest_wdl.localizers module¶

class pytest_wdl.localizers.JsonLocalizer(contents: dict)[source]¶

Bases: pytest_wdl.localizers.Localizer

localize(destination: pathlib.Path)[source]¶

Localize a resource to destination.

Parameters: destination – Path to file where the non-local resource is to be localized.

class pytest_wdl.localizers.LinkLocalizer(source: pathlib.Path)[source]¶

Bases: pytest_wdl.localizers.Localizer

Localizes a file to another destination using a symlink.

localize(destination: pathlib.Path)[source]¶

Localize a resource to destination.

Parameters: destination – Path to file where the non-local resource is to be localized.

class pytest_wdl.localizers.Localizer[source]¶

Bases: object

Abstract base of classes that implement file localization.

abstract localize(destination: pathlib.Path) → None[source]¶

Localize a resource to destination.

Parameters: destination – Path to file where the non-local resource is to be localized.

verify(path: pathlib.Path) → bool[source]¶

Verify that path exists and is valid.

Parameters: path – Path to verify.
Returns: True if the path is verified, else False

class pytest_wdl.localizers.StringLocalizer(contents: str)[source]¶

Bases: pytest_wdl.localizers.Localizer

Localizes a string by writing it to a file.

localize(destination: pathlib.Path)[source]¶

Localize a resource to destination.

Parameters: destination – Path to file where the non-local resource is to be localized.

class pytest_wdl.localizers.UrlLocalizer(url: str, user_config: pytest_wdl.config.UserConfiguration, http_headers: Optional[dict] = None, digests: Optional[dict] = None)[source]¶

Bases: pytest_wdl.localizers.Localizer

Localizes a file specified by a URL.

property http_headers¶

localize(destination: pathlib.Path)[source]¶

Localize a resource to destination.

Parameters: destination – Path to file where the non-local resource is to be localized.

property proxies¶

verify(path: pathlib.Path) → bool[source]¶

Verify that path exists and is valid.

Parameters: path – Path to verify.
Returns: True if the path is verified, else False

pytest_wdl.localizers.download_file(url: str, destination: pathlib.Path, http_headers: Optional[dict] = None, proxies: Optional[dict] = None, show_progress: bool = True, digests: Optional[dict] = None)[source]¶

pytest_wdl.plugins module¶

exception pytest_wdl.plugins.PluginError[source]¶: Bases: Exception

class pytest_wdl.plugins.PluginFactory(entry_point: pkg_resources.EntryPoint, return_type: Type[T])[source]¶

Bases: typing.Generic

Lazily loads a plugin class associated with a data type.

pytest_wdl.plugins.plugin_factory_map(return_type: Type[T], group: Optional[str] = None, entry_points: Optional[Iterable[pkg_resources.EntryPoint]] = None) → Dict[str, pytest_wdl.plugins.PluginFactory[~T][T]][source]¶

Creates a mapping of entry point name to PluginFactory for all discovered entry points in the specified group.

Parameters

group – Entry point group name
return_type – Expected return type
entry_points –

Returns

Dict mapping entry point name to PluginFactory instances

pytest_wdl.url_schemes module¶

class pytest_wdl.url_schemes.BaseResponse[source]¶

Bases: pytest_wdl.url_schemes.Response

download_file(destination: pathlib.Path, show_progress: bool = False, digests: Optional[dict] = None)[source]¶

Download a file to a specific destination.

Parameters

destination – Destination path
show_progress – Whether to show a progress bar
digests – Optional dict mapping hash names to digests. These are used to validate the downloaded file.

Raises

DigestsNotEqualError –

abstract get_content_length() → Optional[int][source]¶

abstract read(block_size: int)[source]¶

class pytest_wdl.url_schemes.Method(src_attr, dest_pattern)[source]¶

Bases: enum.Enum

An enumeration.

OPEN = ('urlopen', '{}_open')¶

REQUEST = ('request', '{}_request')¶

RESPONSE = ('response', '{}_response')¶

class pytest_wdl.url_schemes.Response[source]¶

Bases: object

abstract download_file(destination: pathlib.Path, show_progress: bool = False, digests: Optional[dict] = None)[source]¶

Download a file to a specific destination.

Parameters

destination – Destination path
show_progress – Whether to show a progress bar
digests – Optional dict mapping hash names to digests. These are used to validate the downloaded file.

Raises

DigestsNotEqualError –

class pytest_wdl.url_schemes.ResponseWrapper(rsp)[source]¶

Bases: pytest_wdl.url_schemes.BaseResponse

get_content_length() → Optional[int][source]¶

read(block_size: int) → bytes[source]¶

class pytest_wdl.url_schemes.UrlHandler[source]¶

Bases: urllib.request.BaseHandler

alias()[source]¶: Add aliases that are required by urllib for handled methods.

property handles¶

request(request: urllib.request.Request) → urllib.request.Request[source]¶

response(request: urllib.request.Request, response: pytest_wdl.url_schemes.Response) → pytest_wdl.url_schemes.Response[source]¶

abstract property scheme¶

urlopen(request: urllib.request.Request) → pytest_wdl.url_schemes.Response[source]¶

pytest_wdl.url_schemes.install_schemes()[source]¶

pytest_wdl.utils module¶

exception pytest_wdl.utils.DigestsNotEqualError[source]¶: Bases: AssertionError

pytest_wdl.utils.chdir(todir: pathlib.Path)[source]¶

Context manager that temporarily changes directories.

Parameters: todir – The directory to change to.

pytest_wdl.utils.compare_files_with_hash(file1: pathlib.Path, file2: pathlib.Path, hash_name: str = 'md5')[source]¶

pytest_wdl.utils.context_dir(path: Optional[pathlib.Path] = None, change_dir: bool = False, cleanup: Optional[bool] = None) → pathlib.Path[source]¶

Context manager that looks for a specific environment variable to specify a directory. If the environment variable is not set, a temporary directory is created and cleaned up upon return from the yield.

Parameters

path – The environment variable to look for.
change_dir – Whether to change to the directory.
cleanup – Whether to delete the directory when exiting the context. If None, the directory is only deleted if a temporary directory is created.

Yields

A directory path.

pytest_wdl.utils.ensure_path(path: Union[str, py._path.local.LocalPath, pathlib.Path], search_paths: Optional[Sequence[pathlib.Path]] = None, canonicalize: bool = True, exists: Optional[bool] = None, is_file: Optional[bool] = None, executable: Optional[bool] = None, create: bool = False) → pathlib.Path[source]¶

Converts a string path or py.path.local.LocalPath to a pathlib.Path.

Parameters

path – The path to convert.
search_paths – Directories to search for path if it is not already absolute. If exists is True, looks for the first search path that contains the file, otherwise just uses the first search path.
canonicalize – Whether to return the canonicalized version of the path - expand home directory shortcut (~), make absolute, and resolve symlinks.
exists – If True, raise an exception if the path does not exist; if False, raise an exception if the path does exist.
is_file – If True, raise an exception if the path is not a file; if False, raise an exception if the path is not a directory.
executable – If True and is_file is True and the file exists, raise an exception if it is not executable.
create – Create the directory (or parent, if is_file = True) if it does not exist. Ignored if exists is True.

Returns

A pathlib.Path object.

pytest_wdl.utils.env_map(d: dict) → dict[source]¶: Given a mapping of keys to value descriptors, creates a mapping of the keys to the described values.

pytest_wdl.utils.find_executable_path(executable: str, search_path: Optional[Sequence[pathlib.Path]] = None) → Optional[pathlib.Path][source]¶

Finds ‘executable’ in search_path.

Parameters

executable – The name of the executable to find.
search_path – The list of directories to search. If None, the system search path (defined by the $PATH environment variable) is used.

Returns

Absolute path of the executable, or None if no matching executable was found.

pytest_wdl.utils.find_in_classpath(glob: str) → Optional[pathlib.Path][source]¶

Attempts to find a .jar file matching the specified glob pattern in the Java classpath.

Parameters: glob – JAR filename pattern
Returns: Path to the JAR file, or None if a matching file is not found.

pytest_wdl.utils.find_project_path(*filenames: Union[str, pathlib.Path], start: Optional[pathlib.Path] = None, return_parent: bool = False, assert_exists: bool = False) → Optional[pathlib.Path][source]¶

Starting from path folder and moving upwards, search for any of filenames and return the first path containing any one of them.

Parameters

*filenames – Filenames to search. Either a string filename, or a sequence of string path elements.
start – Starting folder
return_parent – Whether to return the containing folder or the discovered file.
assert_exists – Whether to raise an exception if a file cannot be found.

Returns

A Path, or None if no folder is found that contains any of filenames. If return_parent is False and more than one of the files is found one of the files is randomly selected for return.

Raises

FileNotFoundError if the file cannot be found and assert_exists is True. –

pytest_wdl.utils.hash_file(path: pathlib.Path, hash_name: str = 'md5') → str[source]¶

pytest_wdl.utils.is_executable(path: pathlib.Path) → bool[source]¶

Checks if a path is executable.

Parameters: path – The path to check
Returns: True if path exists and is executable by the user, otherwise False.

pytest_wdl.utils.resolve_file(filename: Union[str, pathlib.Path], project_root: pathlib.Path, assert_exists: bool = True) → Optional[pathlib.Path][source]¶

Finds filename under project_root or in the project path.

Parameters

filename – The filename, relative path, or absolute path to resolve.
project_root – The project root dir.
assert_exists – Whether to raise an error if the file cannot be found.

Returns

A pathlib.Path object, or None if the file cannot be found and assert_exists is False.

Raises

FileNotFoundError if the file cannot be found and assert_exists is True. –

pytest_wdl.utils.resolve_value_descriptor(value_descriptor: Union[str, dict]) → Optional[source]¶

Resolves the value of a value descriptor, which may be an environment variable name, or a map with keys env (the environment variable name) and value (the value to use if env is not specified or if the environment variable is unset.

Parameters: value_descriptor –

Returns:

pytest_wdl.utils.safe_string(s: str, replacement: str = '_') → str[source]¶

Makes a string safe by replacing non-word characters.

Parameters

s – The string to make safe
replacement – The replacement stringj

Returns

The safe string

pytest_wdl.utils.tempdir(change_dir: bool = False, tmproot: Optional[pathlib.Path] = None, cleanup: Optional[bool] = True) → pathlib.Path[source]¶

Context manager that creates a temporary directory, yields it, and then deletes it after return from the yield.

Parameters

change_dir – Whether to temporarily change to the temp dir.
tmproot – Root directory in which to create temporary directories.
cleanup – Whether to delete the temporary directory before exiting the context.

pytest_wdl.utils.verify_digests(path: pathlib.Path, digests: dict)[source]¶

Module contents¶

Fixtures for writing tests that execute WDL workflows using Cromwell. For testability purposes, the implementaiton of these fixtures is done in the pytest_wdl.fixtures module.

Welcome to pytest-wdl’s documentation!¶

User manual¶

Dependencies¶

Installation¶

Install from PyPI¶

Install from source¶

Install optional dependencies¶

Quick Start¶

Project setup¶

Fixtures¶

Test data¶

Files¶

URL Schemes¶

Data Types¶

default¶

vcf¶

bam*:¶

Executors¶

Executor-specific workflow_runner arguments¶

Cromwell¶

Miniwdl¶

dxWDL¶

Known limitations¶

Cromwell¶

miniwdl¶

dxWDL¶

Configuration¶

Project-specific configuration¶

Environment-specific configuration¶

Configuration file¶

Proxies¶

HTTP(S) Headers¶

Executor-specific configuration¶

Provider-specific configuration¶

Fixtures¶

Plugins¶

Creating new data types¶

Creating new executors¶

Supporting alternative URL schemes¶

pytest_wdl¶

pytest_wdl package¶

Subpackages¶

pytest_wdl.data_types package¶

Submodules¶

pytest_wdl.data_types.bam module¶

pytest_wdl.data_types.json module¶

pytest_wdl.data_types.vcf module¶

Module contents¶

pytest_wdl.executors package¶

Submodules¶

pytest_wdl.executors.cromwell module¶

pytest_wdl.executors.miniwdl module¶

Module contents¶

pytest_wdl.providers package¶

Submodules¶

pytest_wdl.providers.dx module¶

Module contents¶

Submodules¶

pytest_wdl.config module¶

pytest_wdl.core module¶

pytest_wdl.fixtures module¶

pytest_wdl.localizers module¶

pytest_wdl.plugins module¶

pytest_wdl.url_schemes module¶

pytest_wdl.utils module¶

Module contents¶

Indices and tables¶

Executor-specific `workflow_runner` arguments¶