Skip to main content

Workflow Custom Code Writing Standards

Background: For security reasons, [python custom skill code] && [advanced orchestration - code node] require standardized code writing instructions.

Supported Python Packages​

1. Built-in Functions and Types​

  • dict
  • sorted
  • filter
  • map
  • enumerate

2. Standard Library Modules​

Data Processing​

  • json
  • decimal
  • uuid
  • base64
  • hashlib

String Processing​

  • re (regular expressions)
  • string
  • textwrap
  • difflib (difference comparison)

Data Structures and Algorithms​

  • copy
  • bisect (binary search)
  • heapq (heap queue)
  • statistics (statistical calculations)

Mathematical Computation​

  • math
  • operator (operator functions)

Date and Time​

  • datetime
  • time
  • calendar

Other Utilities​

  • random (random number generation)
  • requests (HTTP requests)

3. Type Annotations (typing module)​

  • typing (the module itself)
  • Any
  • Union
  • Optional
  • Literal
  • Final
  • ClassVar
  • TypeVar
  • Generic
  • Protocol
  • runtime_checkable
  • overload
  • cast
  • TYPE_CHECKING
  • NoReturn
  • List
  • Dict
  • Set
  • Tuple
  • FrozenSet
  • Deque
  • Counter
  • ChainMap
  • OrderedDict
  • DefaultDict
  • MutableMapping
  • MutableSequence
  • MutableSet
  • Mapping
  • Sequence
  • AbstractSet
  • Collection
  • Container
  • Iterable
  • Iterator
  • Reversible
  • Sized
  • Hashable
  • Callable
  • Awaitable
  • Coroutine
  • AsyncIterable
  • AsyncIterator
  • AsyncGenerator
  • Generator
  • ContextManager
  • AsyncContextManager

Syntax Standards​

Due to the use of secure environment isolation and other restrictions, writers need to follow relevant standards when writing code. Premise: Only built-in syntax and built-in libraries, as well as a limited set of standard libraries, are supported.

  1. Using the math module
import math

def calculate(a, b):
    result = math.fsum([a, b])
    return result

Incorrect usage:

from math import fsum

def calculate(a, b):
    result = fsum([a, b])
    return result

Explanation: Due to current import restrictions in the code block, only parent package imports are supported when using import. The syntax from a import b is not supported.

  1. Using datetime

Correct usage:

import typing
from datetime import datetime


def get_weather(location: str, date: typing.Optional[str] = None) -> dict:
    # Handle default date
    if date is None:
        formatted_date = datetime.datetime.now().strftime("%Y-%m-%d")
    else:
        try:
            # Try to parse common date formats
            if len(date) == 8 and date.isdigit():
                # Handle YYYYMMDD format
                date_obj = datetime.datetime.strptime(date, "%Y%m%d")
            else:
                # Try standard format YYYY-MM-DD
                date_obj = datetime.datetime.strptime(date, "%Y-%m-%d")
            formatted_date = date_obj.strftime("%Y-%m-%d")
        except ValueError:
            # If the format does not match, use the original input (may cause API call failure)
            formatted_date = date

    # This is mock data; in actual use, you may call a weather API
    weather_data = {
        "location": location,
        "date": formatted_date,
        "forecast": "sunny",
        "temperature": "25°C",
        "wind": "light breeze"
    }
    return weather_data

Incorrect usage:

from typing import Optional, List, Dict, Any, Union, Callable
from datetime import datetime
...省略...