| Source Code | Release Notes

GlaceJS Tests Generator

Annotation

Tests Generator creates manual or automated test scenarios from steps description.

Features

• Tests generating via graph traversal and rules management.
• Machine learning in tests building via Markov chains based approach.

Installation

npm i -g glace-testgen

Utility test-gen will be available in console then.

Quick Start

1. Create file steps.yml and fill with content:

   -
     name: open app
     income: null
     outcome:
       app: true
   -
     name: authenticate
     income:
       app: true
     outcome:
       app:
         auth: true
   -
     name: close app
     income: 
       app: true
     outcome: null
   

2. Launch console and execute command:

   test-gen /path/to/steps.yml
   

3. It will generate next scenarios:

   Test case 1:
     - open app
     - authenticate
     - close app
   
   Test case 2:
     - open app
     - close app
   

4. Example of glace-js steps description for autotests generating.

5. Example of glace-proxy steps description for autotests generating.

6. More details about steps description you may find here.

Console Options

Utility usage:

test-gen path/to/steps/file [options]

Steps file may be .json or .yaml (.yml).

Arguments:

• --config [path], -c - Path to JSON file with CLI arguments. Default is cwd/config.json (if it exists).

Log:

• --stdout-log - Print log messages to stdout.
• --log [path] - Path to log file. Default is cwd/glace.log.
• --log-level [level] - Log level. Default is debug.

Test-gen:

• --gen-output-file <path> - Path to output file (yaml format).
• --gen-steps-filter <chunk> - Chunk of step name to filter tests.
• --gen-steps-uniq [number] - Number of steps in unique sequence to filter tests. Default is unlimited.
• --gen-steps-limit [number] - Maximum amount of steps per test. Default is unlimited.
• --gen-steps-usage <number> - Number of steps usage in test case.
• --gen-steps-files <sequence> - Space-separated sequence of paths to steps file (yaml or json format). As alternate to specify path to steps file in plugin mode.
• --gen-tests-limit [number] - Maximum amount of generated tests per iteration. Default is 1000000.
• --gen-tests-max <number> - Maximum amount of final tests.
• --gen-tests-reverse - Reverse order of final tests.
• --gen-tests-files <sequence> - Space-separated sequence of paths to files with pregenerated tests (yaml or json format).
• --gen-tests-only - Flag to exclude other found tests and launch only generated tests in plugin mode.
• --gen-tests-shuffle - Shuffle tests during generating. Provides more steps sequence randomization, but tests will be different in generating runs.
• --gen-train <path> - Load tests from path for training, launch generator in train mode and save result to file.
• --gen-train-result [path] - Path to file with training result. Default is cwd/train-result.json.
• --gen-load-train <path> - Path to file with pretrained model, which will be loaded before generating.
• --gen-train-before <path> - Path to file with tests for training before generating.
• --gen-names-only - Flag to print only step names.

Common:

• --version - Show version number.
• -h, --help - Show help.

How To

How to generate tests quickly

If you have several dozens of steps, tests building with unlimited parameters may require a long time. Use options --gen-steps-uniq, --gen-tests-limit, --gen-steps-limit for efficient optimisation of consumed time.

1. Download example with steps description.

2. Launch command

test-gen e2e.steps.yaml --gen-steps-uniq 2 --gen-tests-limit 150 --gen-output-file result

1. Wait for finish

Generating tests from steps...
50 tests are generated during 419ms

1. Open file result.yml to see generated tests.

How to train generator model

For training you need a file with test samples in yaml format.

1. Download example with steps description.

2. Generate 10 tests

test-gen e2e.steps.yaml --gen-steps-uniq 2 --gen-tests-limit 150 --gen-tests-max 10 --gen-output-file for_train

1. Launch training

test-gen --gen-train for_train.yml

1. Wait for finish

Training model...
Model is trained during 18ms

1. Training result is saved to train-result.json.

How to use trained model for generating

1. Train model from example above.

2. Launch command

test-gen e2e.steps.yaml --gen-steps-uniq 2 --gen-tests-limit 150 --gen-load-train train-result.json --gen-output-file result

1. Generated tests will be saved to file result.yml.

How to train model on fly in generating

1. Prepare test samples for training from example above.

2. Launch command

test-gen e2e.steps.yaml --gen-steps-uniq 2 --gen-tests-limit 150 --gen-train-before for_train.yml --gen-output-file result

1. Generated tests will be saved to file result.yml.

Tests and quality

• Project tests report is here
• Code coverage report is here