如何在Python中将自然语言日期转换为标准时间段？求可用库或API

Absolutely, there are reliable Python libraries that can handle converting natural language date references like your examples into the standardized date range format you need. Let’s break down the best options with practical code examples:

1. dateparser – Intuitive Date Parsing #

dateparser is a great choice for its flexibility with different date formats and natural language inputs. It can easily parse month-year strings and handle range descriptions with a bit of string processing.

First, install it:

pip install dateparser

Here’s a function that handles both your example cases:

import dateparser
from datetime import timedelta

def nl_to_date_range(nl_input):
    # Handle single month-year (e.g., "June 2016" or "06/2016")
    if not any(phrase in nl_input.lower() for phrase in ["to", "between", "and"]):
        # Parse the date, forcing the first day of the month
        parsed_date = dateparser.parse(nl_input, settings={"PREFER_DAY_OF_MONTH": "first"})
        # Calculate the last day of the same month
        next_month = parsed_date.replace(day=28) + timedelta(days=4)
        last_day = next_month - timedelta(days=next_month.day)
        # Format as required
        return f"{parsed_date.strftime('%m/%d/%Y')}-{last_day.strftime('%m/%d/%Y')}"
    
    # Handle date ranges (e.g., "June 2016 to Dec 2016" or "between June 2016 and Dec 2016")
    else:
        # Clean up the input to extract start and end dates
        cleaned_input = nl_input.lower().replace("between ", "").replace(" and ", " to ")
        start_str, end_str = [part.strip() for part in cleaned_input.split("to")]
        
        # Parse start date (first day of month) and end date (we'll get last day)
        start_date = dateparser.parse(start_str, settings={"PREFER_DAY_OF_MONTH": "first"})
        end_date = dateparser.parse(end_str, settings={"PREFER_DAY_OF_MONTH": "first"})
        
        # Calculate last day of the end month
        next_month = end_date.replace(day=28) + timedelta(days=4)
        last_day_end = next_month - timedelta(days=next_month.day)
        
        return f"{start_date.strftime('%m/%d/%Y')}-{last_day_end.strftime('%m/%d/%Y')}"

# Test your examples
print(nl_to_date_range("June 2016"))  # Output: 06/01/2016-06/30/2016
print(nl_to_date_range("06/2016"))  # Output: 06/01/2016-06/30/2016
print(nl_to_date_range("June 2016 to Dec 2016"))  # Output: 06/01/2016-12/31/2016
print(nl_to_date_range("between June 2016 and Dec 2016"))  # Output: 06/01/2016-12/31/2016

2. parsedatetime – Robust Natural Language Support #

If you need to handle more complex natural language phrases beyond your examples, parsedatetime is a powerful option that understands context better.

Install it first:

pip install parsedatetime

Here’s an adapted function for your use case:

import parsedatetime as pdt
from datetime import datetime, timedelta

cal = pdt.Calendar()

def parse_date_range(nl_input):
    # Single month-year case
    if not any(phrase in nl_input.lower() for phrase in ["to", "between", "and"]):
        parsed_result, status = cal.parse(nl_input)
        if status == 1:  # Successful parse
            parsed_date = datetime(*parsed_result[:6])
            first_day = parsed_date.replace(day=1)
            # Get last day of the month
            next_month = first_day.replace(day=28) + timedelta(days=4)
            last_day = next_month - timedelta(days=next_month.day)
            return f"{first_day.strftime('%m/%d/%Y')}-{last_day.strftime('%m/%d/%Y')}"
        return "Failed to parse date"
    
    # Range case
    else:
        # Normalize input to use "to" as the separator
        cleaned_input = nl_input.lower().replace("between ", "").replace(" and ", " to ")
        start_str, end_str = [part.strip() for part in cleaned_input.split("to")]
        
        start_result, start_status = cal.parse(start_str)
        end_result, end_status = cal.parse(end_str)
        
        if start_status == 1 and end_status == 1:
            start_date = datetime(*start_result[:6]).replace(day=1)
            end_date = datetime(*end_result[:6])
            # Calculate last day of end month
            next_month = end_date.replace(day=28) + timedelta(days=4)
            last_day_end = next_month - timedelta(days=next_month.day)
            return f"{start_date.strftime('%m/%d/%Y')}-{last_day_end.strftime('%m/%d/%Y')}"
        return "Failed to parse date range"

# Test
print(parse_date_range("Dec 2016"))  # Output: 12/01/2016-12/31/2016
print(parse_date_range("06/2016 to 12/2016"))  # Output: 06/01/2016-12/31/2016

Both libraries work well for your specific cases. dateparser is often more straightforward for simple inputs, while parsedatetime shines with more varied natural language.

内容的提问来源于stack exchange，提问作者Zhenshan Jin