Miniwdl

Latest version: v1.12.0

Safety actively analyzes 629564 Python packages for vulnerabilities to keep your Python projects secure.

Scan your dependencies

Page 5 of 12

1.2.2

[![PyPI version](https://img.shields.io/pypi/v/miniwdl.svg)](https://pypi.org/project/miniwdl/) [![Anaconda-Server Badge](https://anaconda.org/conda-forge/miniwdl/badges/version.svg)](https://anaconda.org/conda-forge/miniwdl) **[Getting Started](https://miniwdl.readthedocs.io/en/latest/getting_started.html)**

Adds `miniwdl run --env X` to pass through environment variable `X` to the environment of all tasks, or `--env X=foo` to set `X` outright. (Can be supplied multiple times for different variables)

This is a non-standard side channel likely to cause portability problems for WDL tasks relying on it. It should be used only in certain circumstances where portability isn't a priority, for example passing through auth credentials for platform-specific APIs. Otherwise, it's usually better to pass explicit WDL inputs (and if necessary, set environment variables in the command script).

1.2.1

Maintenance release:

* Fix downloads failing on first attempt with `--no-cache` (put = true, get = false)
* Fix the status bar's "ready" counter when the call cache is used
* Set `--no-cache` for self-test
* Failing self-tests generate a log file that can be attached to a support request
* Suppress "stdout unused" warnings for downloader tasks

1.2.0

**Relax type-checking of inputs with defaults**

WDL inputs declared with default initializers, `input { T x = :default: }`, can now be called with optional values *whether or not* their declared type carries the `?` quantifier.

<details>
<summary>Implications & example</summary>

When called with `None`, the default applies if the declaration lacks the `?` quantifier. This change simplifies passing optional overrides through from workflow to task, while the task internally defines the appropriate default:<br/>
wdl
workflow w {
input {
String? s_override
}
call t { input: s = s_override }
}

task t {
input {
String s = "some default"
}
...
}

However, if caller expressly supplies `None` for an input that *is* declared optional, `String? s = "some default"`, the WDL specification does not yet explicate whether `s` should take `None` or the default in this case. For now our implementation has it take `None`, but `miniwdl check` also flags the ambiguity with `UnnecessaryQuantifier`. See ongoing discussion: https://github.com/openwdl/wdl/pull/464

The newly relaxed rule only applies for task/workflow call inputs; elsewhere, it remains necessary to use `select_first()` or `select_all()` to coerce an optional value to a non-optional type.
</details>

**Fix bug causing runs with `--no-cache` to still use stale cached download, and other rare race conditions**

**Usability improvements for logs and error messages**

1.1.5

* support task runtime.returnCodes per WDL 1.1 spec
* insource docker image used for AWS CLI

1.1.4

`miniwdl run`

* Add `-o stdout.json` option to robustly redirect the outputs or error JSON (normally printed to standard output) to a file
* unaffected if some plug-in or other code prints something else on standard output
* Add `--no-outside-imports` to block WDL import statements from directories outside of that of the top-level WDL file (or a `--path` directory)
* Improve certain error messages based on user feedback
* PyYAML dependency replaces ruamel.yaml

1.1.3

* Reverts v1.1.2 change to paths in run outputs JSON, due to unintended interference with caching of successive task calls in workflows
* Fix a rare error arising from concurrent tasks writing into the call cache at the same moment

Page 5 of 12

© 2024 Safety CLI Cybersecurity Inc. All Rights Reserved.