Coding Style Guide

Like most Python projects, we try to adhere to PEP 8 (Style Guide for Python Code) and PEP 257 (Docstring Conventions) with the modifications documented here. Be sure to read all documents if you intend to contribute code to ObsPy.

We rely on flake8 for code style checks, it can be installed using conda install or pip install. Thanks to .flake8 configuration file included in the repository, check of your contributions with flake8 can be done with a single command flake8.

Pre-Commit Hooks

In order to further ease the development process you can use pre-commit hooks. Pre-commit hooks can be configured in your repository with use of pre-commit framework. In order to do that, execute two commands:

pip install pre-commit
pre-commit install

That will install in your git repository all pre-commit hooks configured in .pre-commit-hooks.yaml file. That way, flake8 will immediately complain about problems with the coding style and the changes staged for committing can be adapted accordingly (even with git commit hooks installed, they can be ignored on a per-commit basis using git commit -n).

Import Conventions

Like the Python projects NumPy, SciPy and matplotlib, we try to improve readability of the code by importing the following modules in an unified manner:

>>> import numpy as np
>>> import matplotlib.pylab as plt

Import statements in source code are grouped by standard library imports, followed by third party packages and finally obspy imports. Inside blocks from ... imports come after import ... statements, and both should be sorted alphabetically:

import inspect
import math
import warnings
from copy import copy, deepcopy

import numpy as np
from decorator import decorator

from obspy import read, Stream
from obspy.core import compatibility
from obspy.core.utcdatetime import UTCDateTime
from obspy.core.util import AttribDict, create_empty_data_chunk, NUMPY_VERSION
from obspy.core.util.base import _get_function_from_entry_point
from obspy.core.util.decorator import raise_if_masked, skip_if_no_data

Naming

Names to Avoid

single character names except for counters or iterators
dashes (-) in any package/module name
__double_leading_and_trailing_underscore__ names (reserved by Python)

Naming Convention

Use meaningful variable/function/method names; these will help other people a lot when reading your code.
Prepending a single underscore (_) means an object is “internal” / “private”, which means that it is not supposed to be used by end-users and the API might change internally without notice to users (in contrast to API changes in public objects which get handled with deprecation warnings for one release cycle).
Prepending a double underscore (__) to an instance variable or method effectively serves to make the variable or method private to its class (using name mangling).
Place related classes and top-level functions together in a module. Unlike Java, there is no need to limit yourself to one class per module.
Use CamelCase for class names, but snake_case for module names, variables and functions/methods.

Type	Public	Internal / Private
Packages	`snake_case`
Modules	`snake_case.py`	`_snake_case`
Classes / Exceptions	`CamelCase`	`_CamelCase`
Functions / Methods	`snake_case()`	`_snake_case()`
Variables / Attributes	`snake_case`	`_snake_case`
Constants	`CAPS_WITH_UNDER`	`_CAPS_WITH_UNDER`

Doc Strings / Comments

One-liner Doc Strings: both """ are in new lines

def some_method():
    """
    This is a one line doc string.
    """
    print("test")

Multiple line Doc Strings: both """ are in new lines - also you should try provide a meaningful one-liner description at the top, followed by two linebreaks with further text.

def some_method():
    """
    This is just the short story.

    The long story is, this docstring would not have been able to fit in
    one line. Therefore we have to break lines.
    """
    print("test")

Comments at the end of code lines should come after (at least) two spaces:
```
x = x + 1  # Compensate for border
```

Comments start with a single # followed by a single space. The same goes for multi-line block comments:

# Compensate for border
x = x + 1
# The next line needs some more longish explanation which does not fit
# on a single line.
foobar = (foo + bar) ** 3 - 1

Function/Method Definitions

In docstrings which annotate functions and methods, the following reStructuredText fields are recognized and formatted nicely:

param: Description of a parameter.
type: Type of a parameter.
raises, raise: That (and when) a specific exception is raised.
var: Description of a variable.
returns, return: Description of the return value.
rtype: Return type.

The field names must consist of one of these keywords and an argument (except for returns and rtype, which do not need an argument). This is best explained by an example:

def format_exception(etype, value, tb, limit=None):
    """
    Format the exception with a traceback.

    :param etype: exception type
    :param value: exception value
    :param tb: traceback object
    :param limit: maximum number of stack frames to show
    :type limit: integer or None
    :rtype: list[str]
    :return: Traceback messages.
    """

which renders like this:

format_exception(etype, value, tb, limit=None)

Format the exception with a traceback.

Parameters:

etype – exception type
value – exception value
tb – traceback object
limit (int or None) – maximum number of stack frames to show

Return type:

list[str]

Returns:

Traceback messages.

Reference Conventions

As with numpy.ndarrays or Python lists, we try to reduce the memory consumption by using references where ever possible. In the following example a is appended to b as reference, that is the reason why b get changed when we change a:

>>> a = [1, 2, 3, 4]
>>> b = [5, 6]
>>> b.append(a)
>>> a[0] = -99
>>> print(b)
[5, 6, [-99, 2, 3, 4]]

Tests

test methods names must start with test_ followed by a mixedCase part

Tests which are expected to fail, because there is a known/unfixed bug should be commented with an XXX: followed by an valid ticket number, e.g.

def test_do_something():
    """
    XXX: This test does something.

    But fails badly. See ticket #number.
    """
    print("test")
    ...
    # XXX: here it fails
    ...

Citations

References to publications (journal articles, books, etc.) should be properly reproducible. A bibtex entry in obspy/misc/docs/source/bibliography should be made for each single publication (ideally with an URL or DOI), using first author and year as article identifier:

@article{Beyreuther2010,
  author = {Beyreuther, Moritz and Barsch, Robert and Krischer,
            Lion and Megies, Tobias and Behr, Yannik and
            Wassermann, Joachim},
  title = {ObsPy: A Python Toolbox for Seismology},
  volume = {81},
  number = {3},
  pages = {530-533},
  year = {May/June 2010},
  doi = {10.1785/gssrl.81.3.530},
  URL = {http://www.seismosoc.org/publications/SRL/SRL_81/srl_81-3_es/},
  eprint = {http://srl.geoscienceworld.org/content/81/3/530.full.pdf+html},
  journal = {Seismological Research Letters}
}

This entry can then be referenced (using the bibtex article identifier) in docstrings in the source code with the following Sphinx syntax to be converted to a link to the bibliography section:

def some_function():
    """
    Function to do something.

    See [Beyreuther2010]_ for details.
    """
    return None

Miscellaneous

Lines shouldn’t exceed a length of 79 characters. No, it’s not because we’re mainly using VT100 terminals while developing, rather because the diffs look nicer on short lines, especially in side-by-side mode.
never use multiple statements on the same line, e.g. if check: a = 0.
Prefer list comprehension to the built-in functions filter() and map() when appropriate.