Functionality added to datetime:
- Initialize from date/time strings, intuiting format (no need to call strptime() for supported formats)
- Initialize current date/time from no arguments (no need to call .now())
- Easy way to jump forward and backward a day at a time
- Easy way to get date of next Monday, next Friday, etc.
- Iterate over a range of dates
- Easy conversion from timezone to UTC and back
Import the module:
>>> from dtm import dt
Get the current date/time:
>>> a = dt()
>>> a
dt(1571952371, tz='America/New_York')
>>> a()
'2019-10-24-17:26:11'
Get a target date/time based on a default format:
>>> b = dt("2004.0226 12:17:30")
>>> b()
'2004-02-26-12:17:30'
Jump forward or backward by one or more days:
>>> c = b.next_day(6)
>>> c()
'2004-03-03-12:17:30'
>>> d = c.previous_day(17)
>>> d()
'2004-02-15-12:17:30'
Get the date of the Monday after April 3, 2019, and the Friday before that:
>>> e = d.next_weekday('mon')
>>> e()
'2019-04-08-12:17:30'
>>> f = e.last_weekday('fri')
>>> f()
'2019-04-05-12:17:30'
Iterate over a range of dates:
>>> for day in b.dt_range(c):
>>> day()
'2004-02-26-12:17:30'
'2004-02-27-12:17:30'
'2004-02-28-12:17:30'
'2004-02-29-12:17:30'
'2004-03-01-12:17:30'
'2004-03-02-12:17:30'
'2004-03-03-12:17:30'
'2004-03-04-12:17:30'
Convert between timezones:
>>> q = dt(2019, 3, 10, 3, 0, 0)
>>> q("%F %T %A %Z")
'2019-03-10 03:00:00 Sunday EDT'
>>> q("%F %T %Z %A", tz='utc')
'2019-03-10 07:00:00 UTC Sunday'
>>> q("%F %T %Z %A", tz='pst8pdt')
'2019-03-09 23:00:00 PST Saturday'
>>> q("%F %T %Z %A", tz='cet')
'2019-03-10 08:00:00 CET Sunday'
>>> q("%F %T %Z %A", tz='Asia/Jakarta')
'2019-03-10 14:00:00 WIB Sunday'
dtspec
Any date/time specification. Such values may be in the form of an
epoch, a dt object, a datetime object, a 3 to 9 element tuple of ints,
or a string containing a date/time spec in a specific format.
default local timezone (DLTZ)
The default timezone used by time.localtime(). That is, the default
timezone of the local machine (LM).
epoch
"The Epoch" is the beginning of time from the perspective of the
underlying date/time software Python depends upon. For most Unix-style
operating systems, this is 1970-01-01T00:00:00 UTC.
epoch value
An epoch value is a number of seconds since "The Epoch" (see above).
Epoch values always represent UTC, not local times.
local machine (LM)
The machine on which the software is running
timeref
A reference to a moment in time. A timeref can be represented as a date
and time in UTC format, a date and time localized to a specific non-UTC
timezone, the number of seconds since the epoch (1970-01-01 00:00:00
UTC), etc. Note that timerefs represented by a number of seconds since
the epoch is always implicitly in the UTC frame of reference.
UTC
Universal Coordinated Time. The global reference time which is
generally the same as local time in Greenwich, England.
The main things the dtm module exports are the 'dt' class and the 'td' class. 'dt' objects represent moments in time. A 'dt' object's moment is stored internally as a UTC time reference. 'td' objects represent a span or period of time, stored internally as a number of seconds. Each object includes the methods needed to reformat the internal information for ease of use and understanding.
Objects of this class represent a moment in time, stored as an epoch time (which represents UTC by definition). They also contain a timezone so they can be localized to specific locations on the globe.
The recommended method for importing the dt class is:
>>> from dtm import dt
The dt object constructor will accept several argument schemes
>>> myobj = dt()
Create a dt object containing the current date and time. This is analogous to
>>> myobj = datetime().now()
>>> myobj = dt(2011, 10, 9)
At least year, month, and day are required. Hour, minute, and second are all optional, although somewhat interdependent. For example, because they are positional, you can't provide second without providing minute.
>>> myobj = dt("2011-10-09 20:07:06")
The code tries to intuit the format of the provided date/time string. Most popular formats should work. If no timezone is provided, the input date/time is considered to be in the local timezone. This value will be converted to UTC and stored as an epoch value.
>>> myobj = dt(epoch=1426905900)
>>> dt("%Y.%m%d %H:%M:%S")
'2015.0320 22:45:00'
The value provided must be numeric. It can be an integer or a float, so a value returned by time.time() or time.mktime() can be used. The value provided is stored without any timezone adjustment.
If you have a datetime object and want a dt object, the dt can be initialized directly from the datetime.
>>> myobj = dt(datetime(2011, 10, 9))
The microsecond member will be zeroed out.
Similarly, if you have a dt and want another, just pass in the one you've got:
>>> newobj = dt(myobj)
The tz argument is independent of any others. Any tz argument passed to the constructor specifies the locality of the input date/time specification.
Internally, the date/time value is converted to UTC and stored as a UTC epoch value.
All of dt's output methods accept a tz argument and will convert the internal UTC value to the specified output timezone. If no tz argument is specified on output, the internal UTC value is converted to the default local timezone (DLTZ) for the local machine (LM).
The standard comparison operators (==, !=, <, >, <=, >=) are supported for dt objects, which can be compared to other dt objects and datetime objects. A dt object, A, is less than another dt object, B, if A's UTC timeref falls earlier in time than B's UTC timeref.
Note that if a comparison puts a datetime or other value on the left side of the operation, the "reflection" variant of the operator method will be called. For __gt__(), the reflection is __lt__() and vice versa since x < y has the same truth table as y > x. Similarly, __ge__() reflects to __le__() and vice versa. The equality operator, __eq__(), is its own reflection since it is a commutative operation.
The dt object supports arithmetic with other dt objects, datetime objects, td and timedelta objects, and simple numbers (ints or floats). The following list summarizes what is produced by each interaction. For operations that are not supported, a TypeError exception is raised.
-
[dt] + [dt] -> TypeError
-
[dt] - [dt] -> [td]
-
[dt] + [td] -> [dt]
-
[dt] - [td] -> [dt]
-
[td] + [dt] -> [dt]
-
[td] - [dt] -> TypeError
-
[dt] + [datetime] -> TypeError
-
[dt] - [datetime] -> [td]
-
[datetime] + [dt] -> TypeError
-
[datetime] - [dt] -> [td]
-
[dt] + [timedelta] -> [dt]
-
[dt] - [timedelta] -> [dt]
-
[timedelta] + [dt] -> [dt]
-
[timedelta] - [dt] -> TypeError
-
[dt] + [int,float] -> [dt]
-
[dt] - [int,float] -> [dt]
-
[int,float] + [dt] -> [dt]
-
[int,float] - [dt] -> TypeError
The methods __radd__() and __rsub__() are called when the left hand operand is one of int, float, timedelta, or datetime. These "reflected" variants of __add__() and __sub__() do what is necessary to carry out the operation requested, so that even though int, float, datetime, and timedelta don't know anything about the dt class, they can participate in these operations on the left side of the operator, as one would want a commutative operation to behave.
The __call__() method adjusts the object's internal time to the indicated output timezone and formats it for output according to the provided format (if any).
>>> for day in dt(2011, 10, 1).dt_range(dt(2011, 10, 31)):
>>> # do whatever
Unlike other Python range functions, the dt_range() function is inclusive. That is, the above loop will process 2011-10-31 as well as the rest of the month. Most Python range functions terminate before processing the end value.
Return a dt object containing a timeref 24 hours later than the value in the object on which the method is called.
Return a dt object containing a timeref 24 hours earlier than the value in the object on which the method is called.
The argument, trgs, is one or more weekday names, either a string or a list of strings. The method searches forward in time to find the next occurrence of one of the entries in the list. If the target matches the weekday of the stored time ref, the result is a week later.
The time reference in the object is adjusted based on tz (if specified) and formatted according to fmt. If tz is not provided (or is None), the timezone value in the object is used to determine the timezone adjustment.
The string spec is parsed according to format fmt and interpreted in terms of the timezone indicated by the tz argument. If tz is not specified,
Return the version of the dtm package.
Return the weekday name of the stored time ref.
Search backward in time from the stored time ref to the most recent day matching a day in the weekday list, wkdays (may be a single string or a list of strings). If an entry in the list matches the weekday of the stored time ref, return self.
Return a list of abbreviated, lowercase weekday names.
Return the stored time adjusted to the stored timezone (or tz if provided) in ISO format (YYYY-mm-dd-HH:MM:SS).
Return the stored time adjusted to the stored timezone (or tz if provided) in YYYY.mmdd format.
Return the stored time adjust to the stored timezone (or tz if provided) in YYYY.mmdd.www format.
The dtm module also provides the 'td' class. Objects of this class represent a period of time with a specific length. td objects can be compared to other td objects, timedelta objects, and int or float values. td objects can be added to or subtracted from dt objects to arrive at other dt objects. They can be added to or subtracted from other time intervals (td or timedelta objects).
The recommended import statement for obtaining the td class is:
>>> from dtm import td
td objects can be initialized from the following:
- an int or float value,
- another td object, or
- a timedelta object
>>> myobj = td()
This initializes a time period of 0 length.
Int arguments are interpreted as days, hours, minutes, and seconds, in that order. To specify minutes in this format, a seconds value must also be provided. To specify hours, both minutes and seconds must be specified.
>>> period = td([[[days,] hours,] minutes,] seconds)
So, for example, the following object represents 5 seconds:
>>> short = td(5)
The following object represents 2 minutes, 5 seconds:
>>> not_long = td(2, 5)
The following object represents 5 hours, 10 minutes, 20 seconds:
>>> a_while = td(5, 10, 20)
Finally, the following object represents 2 days, 3 hours, 15 minutes, and 32 seconds:
>>> longer = td(2, 3, 15, 32)
The following named arguments are supported:
- 's', 'secs', 'seconds',
- 'm', 'mins', 'minutes',
- 'h', 'hrs', 'hours',
- 'd', 'days'
Examples:
>>> ten_sec = td(seconds=10)
>>> two_min = td(mins=2, secs=2)
>>> six_hour = td(h=6, m=17, s=3)
>>> week = td(d=7)
The standard comparison operators (==, !=, <, >, <=, >=) are supported for td objects. td objects can be compared to other td objects and timedelta objects. Reflected operations also work because of how Python reflects __ge__() to __le__() (and vice versa) and __gt__() to __lt__() (and verse visa).
The td object supports addition and subtraction with td objects, dt and datetime objects, timedelta objects, and ints or floats. Here's a summary of what each interaction produces. For unsupported operations, a TypeError exception is raised.
-
[td] + [td] -> [td]
-
[td] - [td] -> [td]
-
[td] + [datetime] -> [dt]
-
[td] - [datetime] -> TypeError
-
[datetime] + [td] -> [dt]
-
[datetime] - [td] -> [dt]
-
[td] + [timedelta] -> [td]
-
[td] - [timedelta] -> [td]
-
[timedelta] + [td] -> [td]
-
[timedelta] - [td] -> [td]
-
[td] + [int,float] -> [td]
-
[td] - [int,float] -> [td]
-
[int,float] + [td] -> [td]
-
[int,float] - [td] -> [td]
Note that the reflected variant methods, __radd__() and __rsub__() are defined to support the commutativtity of these operations.
The td object can be multiplied or divided by ints or floats. Any multiply or divide operation involving two td objects or a td with a dt, datetime, or timedelta is unsupported and raises a TypeError exception.
- [td] * [int,float] -> [td]
- [int,float] * [td] -> [td]
Note that __rmul__() and the analogous reflection methods for the other operations are implemented to handle reflected operations correctly (mostly by reporting that they are not supported).
Returns the length of the period in fractional days
>>> a = td(h=6)
>>> a.days()
0.25
Returns the length of the period in fractional hours
>>> b = td(m=105)
>>> b.hours()
1.75
Returns the length of the period in fractional minutes
>>> c = td(s=75)
>>> c.minutes()
1.25
Returns the length of the period in int seconds
>>> d = td(m=3)
>>> d.seconds()
180
Formats the time period in the td object into a string showing days, hours, minutes, and seconds: NdHH:MM:SS.
>>> e = td(238796)
>>> e.dhhmmss()
'2d18:19:56'
Returns days, hours, minutes, and seconds as a tuple of ints.
>>> e.dhms()
(2, 18, 19, 56)
dtm # project directory
|
+- .coveragerc # parameters for test coverage
+- .gitignore # tell git which files to ignore
+- LICENSE # license for this software
+- Makefile # document command lines for tests and such
+- README.md # this file
+- CHANGELOG.md # change history
+- pytest.ini # parameters for pytest
+- requirements.txt # items this project depends on
+- setup.py # configuration management
|
+- dtm # source directory
| |
| +- __init__.py # the payload source code
|
+- tests # tests directory
| |
| +- conftest.py # test configurations
| +- test_0_quality.py # code quality tests
| +- test_1_dt.py # functionality tests
|
+- pytest.ini* # pytest configuration
+- .env* # set env upon cd into this directory (autoenv)
+- venv* # virtual environment
Files marked with * are not in git.
Note: I decided to use environment variables DTM_FORMATS and DTM_STR rather than reading a file every time the dt constructor is called because this class should not add much weight to packages that use it. Environment variables seem like the right balance between configurability and light-weight-ness.
The value of environment variable DTM_FORMATS will be prepended to the list of default parseable formats. For example, the following
$ export DTM_FORMATS="%d/%m/%Y; %d/%m/%y; %d/%m/%y %H:%M:%S"
would add European date formatting (day, month, year) in front of the default American (month, day, year) formats. This would result in 12/11/2019 being interprested as November 12, 2019 rather than December 11, 2019.
The value will be split on ';' and leading and trailing whitespace will be removed from each segment.
The value of environment variable DTM_STR, if set, will be used to format dt.str() output. For example,
$ export DTM_STR="%H:%M:%S %B %d, %Y"
would result in the following:
>>> a = dt()
>>> a()
'09:09:29 October 26, 2019'
Provide a dt method that returns a datetime object, foo, containing the dt object's time reference and timezone such that foo.strftime("%F %T %Z") produces the same output as mydt.strftime("%F %T %Z").
I would like to add timezone support as follows.
At construction, an optional timezone argument can be provided to the constructor method to indicate which timezone the new object should represent. The zone argument can be 'UTC', 'local', or any valid timezone name. If no zone is specified, the default will be 'local'.
If an input date/time string is provided, it will be interpreted as being in the specified timezone. For example,
>>> q = dt('2015.0320 14:45:00', z='CST6CDT')
Will compute the time as 2:45 pm on March 20, 2015, in the CDT zone. This will actually store the UTC time 2015.0320 20:45:00 (the local time plus six hours).
When this object is used to display date/time values, they will be converted from UTC to CDT (i.e., the object will remember which zone it was constructed in and display outputs in terms of that zone).
>>> q.strftime('%Y.%m%d %H:%M:%S')
'2015.0320 14:45:00'
This makes zone to zone conversions as easy as easy:
>>> q = dt('2004.1007 18:19:17', z='Paris')
>>> q.strftime('%Y.%m%d %H:%M:%S', z='EST')
'2004.1007 10:19:17'
Most output methods will also accept an optional zone argument, allowing for on-the-fly conversions. For example,
>>> q = dt('2015.0428 00:14:00') # local time is EDT
>>> q.strftime('%Y.%m%d %H:%M:%S', z='MST')
2015.0427 21:14:00
>>> q.strftime('%Y.%m%d %H:%M:%S', z='Europe/Paris')
2015.0428 06:14:00