Skip to content

stephen-zhao/datetime_matcher

Repository files navigation

datetime-matcher 📆←💬

PyPI PyPI - Status PyPI - Python Version PyPI - License

datetime-matcher is python module that enables an extension of regex which allows matching, extracting, and reformatting stringified datetimes.

It does so by providing an interface eerily similar to python's native re module.

It's mighty useful for doing things like bulk-renaming files with datetimes in their filenames. But don't let us tell you what it's good for—give it a try yourself!

Quick Links

Getting Started

Install it from pypi by running

pip install datetime-matcher

Then, get it into your code by importing and instantiating

from datetime_matcher import DatetimeMatcher
dtm = DatetimeMatcher()

Finally, run your data through it to perform subsitutions (or any of our many other supported operations!)

oh_my_would_you_look_at_the_time = [
  'TheWallClock_1982-Feb-27.jpeg',
  'TheWristWatch_2003-Aug-11.jpg',
  'TheSmartWatch_2020-Mar-10.jpeg',
]

pattern = r'(\w+)_%Y-%b-%d\.jpe?g'
replace = r'%Y%m%d-\1.jpg'

its_all_clear_now = dtm.sub(pattern, replace, text) for text in oh_my_would_you_look_at_the_time

assert its_all_clear_now[0] == '19820227-TheWallClock.jpg'
assert its_all_clear_now[1] == '20030811-TheWristWatch.jpg'
assert its_all_clear_now[2] == '20200310-TheSmartWatch.jpg'

Example

Use Case in String Substitution with Datetime Reformatting

Let's say we have several filenames of the following format that we want to rename:

'MyLovelyPicture_2020-Mar-10.jpeg'

We want to change them to look like this string:

'20200310-MyLovelyPicture.jpg'

The Unclean Way to Do It, without datetime-matcher

Using the standard library re.sub, we run into an issue:

text = 'MyLovelyPicture_2020-Mar-10.jpeg'

pattern = r'(\w+)_([0-9]{4}-\w{3}-[0-9]{2})\.jpe?g' # ❌ messy
replace = r'(??????)-\1.jpg'                        # ❌ what do we put for ??????

result = re.sub(pattern, replace, text)             # ❌ This does't work

We have to manually run datetime.strptime with a custom parser string to extract the date, and then manually insert it back into the replacement string before running a non-generic search-and-replace using the customized replacement string.

Yuck.

The Clean Way to Do It, with datetime-matcher

We can do the following for a quick and easy substitution with reformatting.

from datetime_matcher import DatetimeMatcher
dtmatcher = DatetimeMatcher()

text = 'MyLovelyPicture_2020-Mar-10.jpeg'

pattern = r'(\w+)_%Y-%b-%d\.jpe?g'              # ✅ regex + strptime
replace = r'%Y%m%d-\1.jpg'                      # ✅ template + strftime

result = dtmatcher.sub(pattern, replace, text)  # ✅ magical substitution

assert result == '20200310-MyLovelyPicture.jpg' # ✅ This works like a charm

Dfregex Syntax Informal Spec

The syntax for dfregex is nearly identical to that of conventional python regex. There is only one addition and one alteration to support datetime format codes. This is far from a formal spec, but expect that currently supported syntaxes, within the current major semantic version, will NOT be removed unless provided reasonable notification and a generous deprecation period.

The Datetime Format Codes

The percentage character indicates the beginning of a datetime format code. These codes are the standard C-style ones used in the built-in datetime module for strftime.

For a list of standard codes, see the Python docs.

Minus the exceptions below, and barring platform-specific support, strftime.org is a good alternative list.

NOTE: The following codes are currently not supported: %Z, %c, %x, %X

The Percent Literal (%)

The percentage literal in conventional regex (%) must be escaped in dfregex (\%) because an unescaped one marks the beginning of a datetime format code and otherwise would be ambiguous.

Development

This project has an extensive Makefile for development automation. To get started quick: after cloning this project, run make all. This should create a virtual environment, install all the required dev-time packages, lint, build, and test the project.

Virtual Environment Setup

You only have to run these once. And make all covers these steps automatically, but this is here for your reference.

Use make reinit-venv to create a new virtual environment from scratch. This will live inside the project's root. All subsequent make commands will automatically invoke python from inside this virtual environment. You may rerun this to completely wipe the virtual environment and start from scratch.

Use make init-piptools to bootstrap the virtual environment.

Dependency Management

To add more runtime dependencies on pypi packages, add them to requirements.in. Then use make install-requirements to install them.

To add more devtime dependencies (these will not be deemed dependencies in the distributable/built version of the package), add them to any *.in file within the requirements-devtime.d/ directory. Then use make install-requirements to install them.

To upgrade the pinned dependencies, run make upgrade-requirements. This will re-resolve the requirements to the latest versions, pin the newly resolved versions, and install them.

Building

make all is the one-stop shop for building the project. It will fix auto-fixable linting errors, raise any additional linting errors, build, run unit tests, and produce a coverage report.

To only check lints or fix lints, use make lint or make fix-lint.

To only build, use make build.

To only test, use make test.

Publishing

Use make publish-test to publish to the test pypi repository.

Use make publish to publish to the prod repository.

Cleaning

Use make clean to remove build artifacts.

Use make reinit-venv if you need to completely reset your virtual environment.