Skip to main content

python

Python coding standards for Scenescape — imports, indentation, patterns, and conventions.

Quellinformationen

Repository
open-edge-platform/scenescape
Letzte Quellaktivität
18. Juni 2026 um 10:00
Erkannte Sprache von SKILL.md
Englisch
Sterne
47
Forks
51

Installationsoptionen

Standardmäßig ist der Prompt ausgewählt, der zuerst die Quelle prüft. Sie können zu einem direkten Befehl wechseln oder eine lokale Kopie herunterladen.

Quelldateien prüfen

Lesen Sie SKILL.md und alle von SkillsMP angezeigten Begleitdateien, bevor Sie sich für eine Installation entscheiden.

SKILL.md wird angezeigt

SKILL.md
Quellanweisungen · Schreibgeschützte Vorschau
name
python
description
Python coding standards for Scenescape — imports, indentation, patterns, and conventions.
# Python Coding Standards for Scenescape ## Import Organization Organize imports in this order: 1. Standard library imports 2. Third-party imports (external packages) 3. Local application imports (scene_common, manager, controller, etc.) ```python # Standard library import json import os from pathlib import Path # Third-party import numpy as np from django.http import HttpResponse # Local from scene_common import log from scene_common.geometry import Point from manager.models import Scene ``` ## Naming Conventions - **Classes**: `PascalCase` (e.g., `SceneController`, `RESTClient`) - **Functions/Methods**: `snake_case` (e.g., `get_scene_data`, `validate_message`) - **Constants**: `UPPER_SNAKE_CASE` (e.g., `TOPIC_BASE`, `CHUNK_SIZE`) - **Private members**: Prefix with single underscore `_` (e.g., `_process_data`, `_internal_state`) ## Code Style ### Linting & Formatting - **Linters**: pylint and flake8 (both must pass) - **Formatter**: autopep8 - **Commands**: ```bash make lint-python # Run both pylint and flake8 make format-python # Auto-format with autopep8 ``` ### Indentation - Use **2 spaces** (never tabs) - Checked via `make indent-check` ### Line Length - Target: 100 characters (soft limit) - Hard limit: 120 characters ### Docstrings Use docstrings for classes and public methods: ```python def process_detection(self, detection_data): """Process detection data from sensor. Args: detection_data: Dictionary containing detection information Returns: Processed detection object or None on failure """ ``` ## Django Patterns ### Models - Import from `scene_common` for shared geometry/camera classes - Use `ListField` (from `manager.fields`) for list/array storage - provides database portability (PostgreSQL and non-PostgreSQL) and robust handling of edge cases - Implement `__str__` for admin interface readability ```python from django.db import models from scene_common.geometry import Region as ScenescapeRegion from manager.fields import ListField class Scene(models.Model): name = models.CharField(max_length=255) map = models.FileField(upload_to='maps/') coordinates = ListField(default=list) # Works with PostgreSQL and SQLite def __str__(self): return self.name ``` ### Views - Use class-based views for CRUD operations (`CreateView`, `UpdateView`, `DeleteView`) - Use `@login_required` decorator for protected views - Return `JsonResponse` for AJAX endpoints ### API Endpoints (Django REST Framework) - Use serializers from `manager.serializers` - Implement custom permissions (e.g., `IsAdminOrReadOnly`) - Return proper HTTP status codes ```python from rest_framework.views import APIView from rest_framework.response import Response from rest_framework import status class SceneAPI(APIView): def get(self, request, scene_id): # Implementation return Response(data, status=status.HTTP_200_OK) ``` ## Scene Common Library Patterns ### MQTT/PubSub ```python from scene_common.mqtt import PubSub pubsub = PubSub(mqtt_auth, client_cert, root_cert, mqtt_broker, keepalive=60) pubsub.onMessage = self.handle_message pubsub.connect() ``` ### Schema Validation ```python from scene_common.schema import SchemaValidation schema_val = SchemaValidation(schema_file) if not schema_val.validateMessage("detector", message_data): log.error("Validation failed") ``` ### REST Client ```python from scene_common.rest_client import RESTClient client = RESTClient(rest_url, rest_auth, root_cert) result = client.getScene(scene_id) ``` ### Logging - Use f-strings for interpolated log messages. - Do not use printf-style placeholders (for example `%s`, `%d`) in log statements. ```python from scene_common import log log.info("Processing started") log.error(f"Failed to process scene {scene_id}") log.debug(f"Debug information: {debug_payload}") ``` ## Type Hints Use type hints for function signatures: ```python from typing import Optional, List, Dict, Union from pathlib import Path def process_images( image_paths: List[Path], config: Optional[Dict] = None ) -> Union[np.ndarray, None]: pass ``` ## Error Handling Use specific exceptions and log appropriately: ```python try: result = dangerous_operation() except FileNotFoundError as e: log.error(f"File not found: {e}") return None except Exception as e: log.error(f"Unexpected error: {e}") raise ``` ## Performance Patterns ### Avoid Deep Copies When Possible ```python # Good - shallow copy data = original_data.copy() # Use only when needed data = copy.deepcopy(original_data) ``` ### Use List Comprehensions ```python # Preferred results = [process(item) for item in items if item.valid] # Over results = [] for item in items: if item.valid: results.append(process(item)) ``` ## Common Anti-Patterns to Avoid ❌ **Don't use mutable default arguments**: ```python # Bad def process(data, cache={}): pass # Good def process(data, cache=None): if cache is None: cache = {} ``` ❌ **Don't catch bare exceptions unless re-raising**: ```python # Bad try: risky_operation() except: pass # Good try: risky_operation() except SpecificException as e: log.error(f"Failed: {e}") ``` ❌ **Don't use `from module import *`**: ```python # Bad from scene_common.geometry import * # Good from scene_common.geometry import Point, Region ``` ## Virtual Environment Setup Development uses `.venv` in project root: ```bash python3 -m venv .venv source .venv/bin/activate pip install -r requirements-runtime.txt ``` VS Code configuration should point to `.venv/bin/python`. ## Dependencies - Add to `requirements-runtime.txt` for runtime dependencies - Add to `requirements-build.txt` for build-time dependencies - Rebuild Docker image after dependency changes
Auf GitHub ansehen