Nimue RT

Nimue RT is a regression testing framework designed to simplify and streamline the process of tracking and comparing Python modules execution outcomes. It supports a wide range of file formats, execution environments, and custom plugins to provide a flexible and powerful testing environment.

Table of Contents

• Installation
• Quick Start
• Usage
• Configuration
• Plugins
• License

Installation

To install Nimue RT, you can use pip:

pip install nimue-rt

Quick Start

Here's a simple example to get you started with Nimue RT:

1. Create a Nimue configuration file (e.g., example.nimue):

version: "1.0"

options:
  trace:
    venv_path: "/path/to/venv"
    retry_on_failure: 3
    failure_exit_code: 96
    remove_transient: always
    console: always

trace:
  module:
    version:
      attributes_match:
        - '.*version.*'

2. Run a trace:

nimue trace --config example.nimue path/to/archive.nrt your_module_name

3. Run a regression test:

nimue test --config example.nimue path/to/archive.nrt

Usage

Tracing a Module Execution

Nimue RT allows you to trace the execution of a Python module and save the trace data for future regression testing.

nimue trace --config example.nimue path/to/archive.nrt your_module_name

Running a Regression Test

Once you have a trace archive, you can use Nimue RT to perform non-regression tests.

nimue test --config example.nimue path/to/archive.nrt

Command Line Interface

You can use the following commands with Nimue:

• nimue trace [options] archive module_name [module_args...]: Trace a module's execution and store it in an archive.
• nimue test [options] archive: Run a regression test using the trace archive.

Options

• -c, --config: Specify a custom Nimue configuration file.
• --version: Show the current version of Nimue RT.
• -v, --verbose: Set the log level to INFO.
• -vv, --very-verbose: Set the log level to DEBUG.

Configuration

Nimue RT uses a YAML configuration file to define various aspects of tracing and testing. The configuration can include:

• Global options for tracing and testing.
• Module-specific configurations for versioning and environment variables.
• File-specific ignore patterns.
• Plugin configurations for custom comparison logic.

The .nimue configuration file is the core of Nimue RT's functionality, enabling you to define the behavior for tracing and testing modules. This section provides a comprehensive overview of the options and structure available in a .nimue configuration file, using examples/parscival/parscival.nimue as a reference.

The .nimue configuration file typically contains the following main sections:

1. version: Defines the version of the configuration file format.
2. options: Specifies global options that apply to both tracing and testing.
3. trace: Defines settings and behaviors for tracing a module's execution.
4. test: Configures the behavior and settings for regression testing based on the traces.

Below is a detailed breakdown of each section and its configuration options.

version

The version key indicates the version of the configuration format being used.

version: "1.0"

This should be updated if there are breaking changes or new features added to the configuration format in future releases of Nimue RT.

options

The options section defines global settings that apply to both the trace and test phases.

trace

The trace sub-section contains options for tracing module executions.

options:
  trace:
    venv_path: "/tmp/parrun/.venv"
    retry_on_failure: 3
    failure_exit_code: 96
    remove_transient: always
    console: always

• venv_path: The path to the virtual environment that should be used during tracing.
• retry_on_failure: Number of retries allowed if the trace fails.
• failure_exit_code: The exit code that indicates a failure during tracing.
• remove_transient: Specifies when to remove transient data. Options are:
  • always: Always remove transient data after tracing.
  • on_success: Remove transient data only if tracing is successful.
  • on_error: Remove transient data only if tracing fails.
• console: Defines when to show the console output. Options are:
  • always: Always show console output.
  • on_error: Show console output only if an error occurs.
  • never: Never show console output.

test

The test sub-section contains options for running tests based on the traced execution.

  test:
    venv_path: "/tmp/partest/.venv"
    retry_on_failure: 1
    failure_exit_code: 96
    remove_transient: on_success
    console: on_error
    show_test_log: on_error
    max_report_lines: 50

• venv_path: The path to the virtual environment that should be used during testing.
• retry_on_failure: Number of retries allowed if the test fails.
• failure_exit_code: The exit code that indicates a failure during testing.
• remove_transient: Specifies when to remove transient data. Options are the same as in the trace section.
• console: Defines when to show the console output during testing.
• show_test_log: Specifies when to show the test log. Options are:
  • always: Always show the test log.
  • on_error: Show the test log only if an error occurs.
  • never: Never show the test log.
• max_report_lines: Limits the number of lines displayed in the test report.

trace

The trace section defines how the module tracing should be performed. This includes configurations for environment variables, file handling, and exit codes.

trace:
  module:
    version:
      attributes_match:
        - '.*version.*'

module

The module sub-section allows you to specify how the module's version information should be captured.

• version.attributes_match: A list of regular expressions that match the attributes related to versioning within the module. These are used to capture the module's version information.

environment

The environment sub-section defines how environment variables should be handled during tracing.

  environment:
    capture:
      - FOO
      - BAR
    module_prefixes: true

• capture: A list of environment variable names or prefixes that should be captured during tracing.
• module_prefixes: A boolean value indicating whether to capture environment variables that start with the module name or its variations (e.g., lowercased, capitalized).

exit_codes

The exit_codes sub-section configures how different exit codes should be handled during tracing.

  exit_codes:
    store_skip:
    store_keep:
      - 0

• store_skip: A list of exit codes for which the trace data should not be stored.
• store_keep: A list of exit codes for which the trace data should always be stored.

files

The files sub-section defines patterns for files to ignore during tracing.

  files:
    read:
      ignore:
        - "/bin/"
        - "/lib/"
    write:
      ignore:
        - "/tmp/"

• read.ignore: A list of patterns for files that should be ignored when reading during tracing.
• write.ignore: A list of patterns for files that should be ignored when writing during tracing.

test

The test section configures how regression testing should be performed using the traces.

test:
  module:
    version:
      attributes_match:
        - '.*version.*'

module

The module sub-section allows you to specify how the module's version information should be compared during testing.

• version.attributes_match: A list of regular expressions that match the attributes related to versioning within the module.

comparing

The comparing sub-section defines how different aspects of the module and its execution should be compared between the trace and the test run.

  comparing:
    module:
      version:
        - plugin: 'comparing.metadata.semver'
          policy: warn
          enabled: true
          params:
            version: '${exec.module.version.__version__}'
            expression: '>=${nrt.module.version.__version__}'

• plugin: The name of the plugin used to compare the module version.
• policy: The policy to apply if the comparison fails. Options are reject or warn.
• enabled: A boolean indicating whether this comparison should be performed.
• params: Parameters passed to the plugin, allowing dynamic evaluation using variables such as ${exec.module.version.__version__}.

before_run and after_run

These sub-sections allow you to define commands or scripts that should be run before or after the test.

  before_run:
  after_run:

These sections are placeholders for commands that might be run to prepare the environment or clean up after the test.

exit_codes

The exit_codes sub-section configures how different exit codes should be handled during testing.

  exit_codes:
    compare_skip:
    compare_test:

• compare_skip: A list of exit codes for which the comparison should be skipped.
• compare_test: A list of exit codes for which the comparison should be performed.

comparing.trace

The trace sub-section within comparing defines how the trace data should be compared during testing.

      console:
        stdout:
          - plugin: 'comparing.console.content'
            policy: warn
            enabled: true
            params:
              min_similarity: 0.9
              clean_patterns:
                - '\d{4}-\d{2}-\d{2} \d{2}:\d{2}:\d{2}'

• console.stdout: Defines how to compare the standard output of the traced execution.
  • plugin: The name of the plugin used to compare the console output.
  • policy: The policy to apply if the comparison fails.
  • enabled: A boolean indicating whether this comparison should be performed.
  • params: Parameters passed to the plugin, such as min_similarity to define the acceptable similarity ratio and clean_patterns to define regex patterns to clean the output before comparison.

files

The files sub-section within comparing.trace defines how files read or written during execution should be compared.

      files:
        read:
          hdf5:
            - plugin: 'comparing.files.hdf5'
              policy: reject
              enabled: true
              params:

• read and written: Define the file types to be compared, such as hdf5, csv, json, yaml, and default.
• plugin: The name of the plugin used for file comparison.
• policy: The policy to apply if the comparison fails.
• enabled: A boolean indicating whether this comparison should be performed.
• params: Additional parameters passed to the plugin.

Plugins

Nimue RT supports custom plugins for comparing files, console outputs, exit codes, and metadata. Plugins can be used to extend the default behavior and integrate specific comparison logic.

Available Plugins

• File Comparators:

  • comparing.files.default: Default file comparator.
  • comparing.files.yaml: YAML file comparator.
  • comparing.files.json: JSON file comparator.
  • comparing.files.hdf5: HDF5 file comparator.
  • comparing.files.csv: CSV file comparator.

• Console Comparators:

  • comparing.console.content: Compares console output content.

• Exit Code Comparators:

  • comparing.exitcode.value: Compares exit codes using expressions.

• Metadata Comparators:

  • comparing.metadata.semver: Compares semantic versions using semver.

Custom Plugins

You can create custom plugins by defining them in Python and registering them with Nimue RT's plugin system.

License

Nimue RT is licensed under the MIT License. See the LICENSE file for more information.