friendlydateparser 0.2.4

A flexible date parsing library that can handle complex natural language expressions, making date manipulation easy and intuitive.

friendlydateparser

This Python module provides methods for parsing text into date, and datetime objects.

For instance, it can parse expressions like:

tuesday, october 15, 2024 14:45 Europe/Paris
2 days before the last day of next month
1h15m after next sunday at midnight CEST
the second monday of 2012

The aim is to be able to accept expressions witch, even if complex, express date references which are common in everyday life.

API

parse_date(text, now=None, month_first=True)

Parses date information from a given string and returns a date object. The function can handle different formats including relative date descriptions (e.g., "next month", "last Friday") or explicit dates (e.g., "10/3/2017", "15th of July").

• Parameters:

  • text (str): The text containing date information to be parsed.
  • now (datetime.date or datetime.datetime, optional): The reference date to use for relative expressions. Defaults to the current date if not specified.
  • month_first (bool, optional): Indicates whether the month appears first in numerical dates (e.g., 10/3 is treated as October 3rd if month_first=True). Defaults to True.

• Returns: A datetime.date object.

• Example:

  date_obj = parse_date("the first of next month")
  # Assuming today is 2023-10-10, returns: datetime.date(2023, 11, 1)
  

parse_datetime(text, now=None, month_first=True, default_tz=None)

Parses both date and time information from a given string and returns a datetime object. The function handles a wide range of date and time formats, including explicit and relative formats.

• Parameters:

  • text (str): The text containing datetime information to be parsed.
  • now (datetime.date or datetime.datetime, optional): The reference date to use for relative expressions. Defaults to the current date and time if not specified.
  • month_first (bool, optional): Indicates whether the month appears first in numerical dates (e.g., 10/3 is treated as October 3rd if month_first=True). Defaults to True.
  • default_tz: default timezone when the given string doesn't specify one. Defaults to None which causes the function to return a naive datetime object when a TZ does not appear explicitly in the given text.

• Returns: A datetime.datetime object.

• Example:

  datetime_obj = parse_datetime("january 1, 2017 at 14:30")
  # Returns: datetime.datetime(2017, 1, 1, 14, 30)
  

Supported Formats

The module can parse a wide variety of date and time formats, such as:

• Explicit Dates: Formats such as 3-october-2017, 10/3/2017, 2017/12/3, march 15, 2017, 15/july/2023 are accepted, with the ability to handle both mm/dd/yyyy and dd/mm/yyyy formats based on the month_first parameter.

• Incomplete dates: 1 october, 10/3, feb 2020, 2024, nov. The current year is used to fill in missing information on the right, while missing values on the left are filled with "ones" or "zeros" as appropriate. For instance, feb 2020 becomes the 1st of February of 2020 at 00:00:00, october becomes the 1st of October of the current year at 00:00:00.

  Note that very ambiguous expressions like 9 are just rejected.

• Relative Dates: the first of next month, next week, the day after tomorrow, october 1 last year.

• Relative Weekdays: Phrases like monday next week, last monday, second sunday of 2023, last friday of october.

• Year and month week numbers: wed week 20 2018, week 2 october 2017, last week october.

• Time and Timezones: today at 12pm, last monday 12h40m, 10/2/2022 40:30:12.1 CEST, tomorrow at midnight Europe/Paris.

• Deltas (before and after): 3 weeks before yesterday, 1d 4h after next monday at noon, 1 month ago. Time units can be abbreviated, for instance, both days, day, ds and d are accepted as days. On the other hand months can only be given as month or months as m and ms mean minutes.

  Negative numbers are also accepted: 3 weeks -1 day before today.

• Relative references (by): in some cases it is useful to reference some relative date from another date. For instance, a common case is when you want to express an idea like "the last Monday including today if today is Monday". We can express that with by as last monday by tomorrow. Other examples are this month by sunday, monday by the first of next month.

• ISO8601: 2024-12-31T13:01+02:00, 2024-W01-8T10:12Z, 2008-365.

And some extra considerations:

• The only inflexible rule when parsing is the order of the components, which must follow (delta, date, time, timezone).

• Besides that, the module tries to parse anything which makes sense without being ambiguous.

• Common particles such as the, of, commas, dashes, slashes, etc. can be included when they are common but also usually excluded. For instance, october the second of 2015, october the second, 2015, october 2nd 2015 are all parsed correctly.

• Space can be included freely everywhere except:

  • Between the digits of a number.
  • In ordinal numbers composed by several words (dashes must be used to join them as in twenty-third).
  • In timezone names (Europe/Paris).

  For instance, valid expressions are 1d-1h before october/23, 1 d - 1 h before october / 23.

  On the other hand, space is required between consecutive words. For instance, fridaynextweek is not a valid expression.

• Abbreviations for weekdays, month names and units are accepted as long as they are unambiguous.

• Ordinals can be written as numbers with the appropriate suffix (1st, 25th, etc.). Ordinals from 1st to 99th can also be written as words.

Note: If you think a new format should be supported, just ask for it!

Error Handling

• If the input text is incomplete or not recognizable as a valid date/time, a ValueError may be raised.

Example Usage

from friendlydateparser import parse_date, parse_datetime

# Parsing a date
date_obj = parse_date("the last day of next jan")
# Returns: datetime.date(2024, 1, 31)

# Parsing a datetime
datetime_obj = parse_datetime("march 15, 2017 11:59 PM")
# Returns: datetime.datetime(2017, 3, 15, 23, 59)

See Also

• datetime: For working with date and time objects in Python.

• time: Provides various time-related functions, such as working with timestamps or sleep functions. This can be useful in combination with date and time parsing.

• dateutil.relativedelta: For relative date calculations, like adding or subtracting months or years.

• calendar: For calendar-related functions, such as determining leap years or getting the number of days in a month.

• pytz: For handling time zones in Python. It can be useful when parsing dates and times that involve different time zones or when converting between local times and UTC.

• pendulum: A library that provides easy-to-use functions for parsing, formatting, and manipulating dates and times, with a focus on improved datetime management and natural language parsing.

• dateparser: A library similar to this module, which parses natural language dates, even supporting multiple languages.

• Arrow: A datetime alternative with a focus on usability. The dehumanize feature allows one to parse time deltas expressed in natural language.

License

Copyright (c) 2024 Salvador Fandiño García

This project is licensed under the MIT License. See the LICENSE file for more details.