Deply - keep your python architecture clean

---

Deply is a standalone Python tool for enforcing architectural patterns and dependencies in large Python projects. By analyzing code structure and dependencies, this tool ensures that architectural rules are followed, promoting cleaner, more maintainable, and modular codebases.

Features

• Layer-Based Analysis: Define project layers and restrict their dependencies to enforce modularity.
• Dynamic Layer Configuration: Easily configure collectors for each layer using file patterns, class inheritance, and logical conditions.
• Cross-Layer Dependency Rules: Specify rules to disallow certain layers from accessing others.
• Extensible and Configurable: Customize layers and rules for any Python project setup.

Installation

To install Deply, use pip:

pip install deply

Configuration

Before running the tool, create a configuration file (deply.yaml or similar) that specifies the rules and target files to enforce.

Example Configuration (deply.yaml)

deply:
  paths:
    - /path/to/your/project

  exclude_files:
    - ".*\\.venv/.*"

  layers:
    - name: models
      collectors:
        - type: bool
          any_of:
            - type: class_inherits
              base_class: "django.db.models.Model"
            - type: class_inherits
              base_class: "django.contrib.auth.models.AbstractUser"

    - name: views
      collectors:
        - type: file_regex
          regex: ".*/views_api.py"

    - name: app1
      collectors:
        - type: directory
          directories:
            - "app1"

    - name: services
      collectors:
        - type: bool
          must:
            - type: class_name_regex
              class_name_regex: ".*Service$"
          must_not:
            - type: file_regex
              regex: ".*/excluded_folder_name/.*"
            - type: decorator_usage
              decorator_name: "deprecated_service"

    - name: auth_protected
      collectors:
        - type: decorator_usage
          decorator_name: "login_required"

  ruleset:
    views:
      disallow_layer_dependencies:
        - models

    tasks:
      enforce_function_decorator_usage:
        - type: function_decorator_name_regex
          decorator_name_regex: "app.task"

    models:
      enforce_inheritance:
        - type: bool
          any_of:
            - type: class_inherits
              base_class: "django.db.models.Model"
            - type: class_inherits
              base_class: "django.contrib.auth.models.AbstractUser"

Commands and Usage

Deply uses a command-line interface to analyze your project.

Basic Usage

Run Deply by executing the following command in your terminal:

deply analyze

By default, Deply looks for a configuration file named deply.yaml in the current directory.

Command-Line Arguments

• deply analyze: Analyzes the project dependencies based on the configuration.
  • --config: Path to the configuration YAML file. Default is deply.yaml.
  • --report-format: Format of the output report. Choices are text, json, html. Default is text.
  • --output: Output file for the report. If not specified, the report is printed to the console.
• -h, --help: Displays help information about Deply and its commands.

Examples

• Analyze the project using the default configuration file:

  deply analyze

• Analyze the project with a specific configuration file:

  deply analyze --config=custom_config.yaml

• Display help information:

  deply --help

Default Behavior

• Configuration File: If no --config argument is provided, Deply looks for deply.yaml in the current directory.
• Paths: If the paths option is not specified in the configuration file, Deply uses the directory where the configuration file resides.

Sample Output

If violations are found, the tool will output a summary of architectural violations grouped by layer, along with details of each violation, such as the file, line number, and violation message.

/path/to/your_project/your_project/app1/views_api.py:74 - Layer 'views' is not allowed to depend on layer 'models'

Collectors

Collectors define how code elements are collected into layers. Deply supports several types of collectors:

BoolCollector

The BoolCollector allows combining other collectors with logical operations such as must, any_of, and must_not. This enables you to define complex conditions for collecting code elements without deep nesting.

• Configuration Options:
  • type: "bool"
  • must (optional): A list of collectors where all must match.
  • any_of (optional): A list of collectors where at least one must match.
  • must_not (optional): A list of collectors where none must match.

Example using must, any_of, and must_not:

- type: bool
  must:
    - type: class_name_regex
      class_name_regex: ".*Service$"
  any_of:
    - type: decorator_usage
      decorator_name: "transactional"
    - type: decorator_usage
      decorator_name: "cacheable"
  must_not:
    - type: file_name_regex
      regex: ".*test.*"

Explanation:

• Collects classes whose names end with Service.
• Additionally, the class must either use the @transactional decorator or the @cacheable decorator.
• Excludes any classes from files with names matching .*test.*.

ClassInheritsCollector

Collects classes that inherit from a specified base class.

• Configuration Options:
  • type: "class_inherits"
  • base_class: The fully qualified name of the base class to match.
  • exclude_files_regex (optional): Regular expression to exclude certain files.

Example:

- type: class_inherits
  base_class: "django.db.models.Model"

ClassNameRegexCollector

Collects classes whose names match a specified regular expression.

• Configuration Options:
  • type: "class_name_regex"
  • class_name_regex: Regular expression to match class names.
  • exclude_files_regex (optional): Regular expression to exclude certain files.

Example:

- type: class_name_regex
  class_name_regex: ".*Service$"
  exclude_files_regex: ".*excluded_folder_name.*"

FunctionNameRegexCollector

Collects functions whose names match a specified regular expression.

Configuration Options:

• type: "function_name_regex"
• function_name_regex: Regular expression to match function names.
• exclude_files_regex (optional): Regular expression to exclude certain files.

Example:

- type: function_name_regex
  function_name_regex: "^helper_.*"

DecoratorUsageCollector

Collects functions or classes that use a specific decorator.

• Configuration Options:
  • type: "decorator_usage"
  • decorator_name (optional): The name of the decorator to match exactly.
  • decorator_regex (optional): Regular expression to match decorator names.
  • exclude_files_regex (optional): Regular expression to exclude certain files.

Example using decorator_name:

- type: decorator_usage
  decorator_name: "login_required"

Example using decorator_regex:

- type: decorator_usage
  decorator_regex: "^auth_.*"

DirectoryCollector

Collects code elements from specified directories.

• Configuration Options:
  • type: "directory"
  • directories: List of directories (relative to the project root) to include.
  • recursive (optional): Whether to search directories recursively (true by default).
  • element_type (optional): Type of code elements to collect ("class", "function", "variable").
  • exclude_files_regex (optional): Regular expression to exclude certain files.

Example:

- type: directory
  directories:
    - "app1"
    - "utils"
  recursive: true
  element_type: "function"

FileRegexCollector

Collects code elements from files matching a specified regular expression.

• Configuration Options:
  • type: "file_regex"
  • regex: Regular expression to match file paths.
  • element_type (optional): Type of code elements to collect ("class", "function", "variable").
  • exclude_files_regex (optional): Regular expression to exclude certain files.

Example:

- type: file_regex
  regex: ".*/views_api.py$"
  element_type: "class"

Rules

Rules define the architectural constraints and standards that your layers must adhere to. By default, Deply supports a set of rules that you can apply to layers to ensure clean, modular, and maintainable architectures.

Rules are defined under the ruleset section in deply.yaml, mapping layer names to rules that apply to those layers. Each rule can restrict dependencies, enforce naming conventions, decorator usage, inheritance patterns, and more.

Disallow Layer Dependencies

Prevent certain layers from depending on others:

views:
  disallow_layer_dependencies:
    - models

This ensures the views layer cannot depend on the models layer. If views references models directly, a violation will be reported.

Class Naming Rules

Use rules to enforce class naming patterns:

models:
  enforce_class_naming:
    - type: class_name_regex
      class_name_regex: ".*View$"

This ensures that classes in models must end with View. Violations occur if a class name doesn't match.

Function Naming Rules

Enforce function naming conventions:

views:
  enforce_function_naming:
    - type: function_name_regex
      function_name_regex: ".*view$"

All functions in views must end with view. Non-matching functions trigger violations.

Class and Function Decorator Usage Rules

Ensure that classes or functions have specific decorators:

auth_protected:
  enforce_class_decorator_usage:
    - type: decorator_name_regex
      decorator_name_regex: "@dataclass"

This ensures that classes in auth_protected have a decorator matching @dataclass.

For functions, a similar rule can be applied:

tasks:
  enforce_function_decorator_usage:
    - type: function_decorator_name_regex
      decorator_name_regex: "app.task"

Functions in tasks must have at least one decorator whose name matches "app.task".

Inheritance Rules

Require classes to inherit from certain base classes:

models:
  enforce_inheritance:
    - type: class_inherits
      base_class: "django.db.models.Model"

All classes in models must inherit from django.db.models.Model. If a class does not, a violation is reported.

Boolean Logic (BoolRule)

Use a bool rule type to combine multiple conditions with logical operations must, any_of, and must_not:

models:
  enforce_inheritance:
    - type: bool
      any_of:
        - type: class_inherits
          base_class: "django.db.models.Model"
        - type: class_inherits
          base_class: "django.contrib.auth.models.AbstractUser"

In this example, classes in models must inherit from either django.db.models.Model or django.contrib.auth.models.AbstractUser. If none of these conditions are met, a violation occurs.

• must: All specified sub-rules must pass.
• any_of: At least one sub-rule must pass.
• must_not: No sub-rule should pass (they should fail to avoid violation).

This logical flexibility allows complex rule scenarios without deep nesting or duplication.

By mixing and matching these rules, you can enforce your architectural guidelines effectively, ensuring that every layer and component adheres to the desired standards. As your project evolves, you can refine or add rules to maintain architectural integrity and prevent unwanted dependencies and naming patterns.

Running Tests

To test the tool, use unittest:

python -m unittest discover tests

Feel free to contribute to this roadmap or suggest features by opening an issue or submitting a pull request! Together, we can make Deply a powerful tool for the Python community. 😊

License

See the LICENSE file for details.