Cyril Matthey-Doret

Task Automation: Just Make it Easy

2025-03-14T00:00:00+00:00
When working in a git repository, we often use recipes to automate frequent tasks, such as building, linting, formatting, testing or benchmarking.
This provides a written trace of such "recipes" and a way to discover, remember and run the commands easily in the right order.
While this can be achieved with custom shell scripts, it would be error prone and require much boilerplate code.
Instead, there are "command runner" tools that are made for this purpose.
Make is a Build System</h2>
Most projects use `make</code></a> as a "command runner", because it is well known and available on most systems.`
`However, make</code> was intended as a build system, not a generic command runner: its purpose is to organize the compilation of source files, and this is reflected by its features:`
`recipes are designed around file dependencieiave file Recipes in a "Makefile" are expected to have file inputs and outputs.</li> running arbitrary commands require explicitely marking rules with .PHONY</code>.</li>`
`no built-in way to pass arguments to recipes.</li> </ul> Additionally, some default behaviours get in the way of using make</code> as a task runner, for example if a .PHONY</code> recipe has the same name as a file or directory (e.g. test/</code>), it will refuse to run it.`
`Just Use a Task Runner</h2>`just</code></a> is only one of many tools that were designed specifically as task runners. Other examples include task</a> or doit</a>. In brief, it has saner default than make</code> for use as a task runner and provides useful options which remove the need for glue code: print recipe comments as help message</li> group recipes into sub commands</li> automatically load .env files, or make it a requirement</li> pass variadic or named arguments to recipes</li> </ul> The main con of just</code>, is that it is likely not available on the system, however it is available on most package registries and can be used as a standalone binary. Example</h2> Here is a small example justfile</code> showcasing some useful just features: #!/usr/bin/env just set shell := ["bash", "-cue"] # <- enforce specific shell in recipes set positional-arguments # <- allow passing positional args set dotenv-required # <- source .env, exit if missing alias bench := benchmark # default behaviour: show list of recipes _default: # <- underscore prefix hides recipe from the list @just --list --no-aliases # <- @ hides command from output # Install dependencies install: uv sync # Run tests test args: install # <- install, then test uv run pytest {{args}} # <- all args passed to pytest # Measure performances benchmark model="large" dataset="iris": install # <- args with default values uv run ./benchmark.py --model {{model}} --dataset {{dataset}} </code></pre> If the above is saved in ./justfile</code>, we can use just</code> as follows: ➜ just Available recipes: benchmark model="large" dataset="iris" # Measure performances install # Install dependencies test args # Run tests ➜ just test --collect-only uv sync Resolved 6 packages in 1ms Audited 4 packages in 0.02ms uv run pytest --collect-only # <- all args passed to pytest ============================== test session starts =============================== platform linux -- Python 3.13.1, pytest-8.3.5, pluggy-1.5.0 rootdir: /tmp/foobar configfile: pyproject.toml collected 2 items <Dir foobar> <Dir tests> <Module test_foobar.py> <Function test_foo> <Function test_bar> =========================== 2 tests collected in 0.01s =========================== ➜ just bench small uv sync Resolved 6 packages in 1ms Audited 4 packages in 0.02ms uv run ./benchmark.py --model small --dataset iris benchmarking small model on iris dataset </code></pre> Conclusion</h2> Only use make</code> if you actually need a build system, otherwise use a proper task runner tool. I personally think the convenience is well worth having one extra dependency, and this is a non-issue if the project comes with its environment (e.g. ~conda~ micromamba</a>, docker</a>, nix</a>, pixi</a>). In any case, reimplementing it with shell script glue is risky and time consuming. Resources</h1> https://theorangeone.net/posts/just-stop-using-makefile/



Simplifying Research Code With Modern Python Features
2023-01-13T00:00:00+00:00
Most scientific software is developed by researchers whose main area of expertise is not software engineering, and who may not have much time to dedicate to development. This often results in a tendency to ignore more advanced language features, such as failing to create custom types and instead over-rely on the language's built-in types. </p>
While it is arguably better to avoid over-engineering to keep the code simple and accessible to a wider audience, using more advanced features can also improve readability and maintainability in larger codebases. In such situations, sticking to the built-ins can result in a code smell known as primitive obsession</em>.</p>
Fortunately, python provides several features to easily take advantage of custom types and interfaces with little overhead. In this post, I'll be showcasing a few modern(ish) python features that can help keep your code readable, and save you time in the long term.</p>

I thank Fabian Egli</a> for his feedback, which helped improve</a> this post.</p>
</blockquote>
🪆 The problem</h2>
A few issues may arise when over-relying on built-ins to represent complex domain-specific models:</p>

Lack of custom interfaces can force more complex logic or repetitions</li>
Code becomes harder to read due to the absence of explicitly named fields and methods</li>
Maintainability suffers: changing a field may require replacing it at multiple places, as opposed to a single definition.</li>
</ul>
Let's take a look at an example that uses only built-ins organized into nested structures:</p>
genes</span></span> =</span> [</span>
    (</span>"</span></span></span>geneA"</span></span></span>,</span> (</span>"</span></span></span>chr1"</span></span></span>,</span> 1002</span>,</span> 1200</span>,</span> "</span></span></span>+"</span></span></span>)</span></span>)</span></span>,</span>
    (</span>"</span></span></span>geneB"</span></span></span>,</span> (</span>"</span></span></span>chr2"</span></span></span>,</span> 39001</span>,</span> 39950</span>,</span> "</span></span></span>-"</span></span></span>)</span></span>)</span></span>,</span>
]</span></span>
genome</span></span> =</span> [</span>(</span>"</span></span></span>chr1"</span></span></span>,</span> 141123</span>,</span> [</span>genes</span></span></span>[</span></span>0</span></span>]</span></span>]</span></span>)</span></span>,</span> (</span>"</span></span></span>chr2"</span></span></span>,</span> 131312</span>,</span> [</span>genes</span></span></span>[</span></span>1</span></span>]</span></span>]</span></span>)</span></span>]</span></span>
read</span></span> =</span> (</span>"</span></span></span>chr2"</span></span></span>,</span> 39800</span>,</span> 40002</span>,</span> "</span></span></span>-"</span></span></span>)</span></span>


def</span></span> read_in_gene</span></span></span>(</span></span>read</span>,</span> genome</span>)</span></span>:</span></span>
    for</span> chrom</span> in</span></span> genome</span></span></span>:</span></span>
        if</span> chrom</span></span></span>[</span></span>0</span></span>]</span></span> ==</span> read</span></span></span>[</span></span>0</span></span>]</span></span>:</span></span>
            for</span> gene</span> in</span></span> chrom</span></span></span>[</span></span>2</span></span>]</span></span></span>:</span></span>
                starts_in</span></span> =</span> gene</span></span></span>[</span></span>1</span></span>]</span></span>[</span></span>1</span></span>]</span></span> <=</span> read</span></span></span>[</span></span>1</span></span>]</span></span> <=</span> gene</span></span></span>[</span></span>1</span></span>]</span></span>[</span></span>2</span></span>]</span></span>
                ends_in</span></span> =</span> gene</span></span></span>[</span></span>1</span></span>]</span></span>[</span></span>1</span></span>]</span></span> <=</span> read</span></span></span>[</span></span>2</span></span>]</span></span> <=</span> gene</span></span></span>[</span></span>1</span></span>]</span></span>[</span></span>2</span></span>]</span></span>
                same_sign</span></span> =</span> gene</span></span></span>[</span></span>1</span></span>]</span></span>[</span></span>3</span></span>]</span></span> ==</span> read</span></span></span>[</span></span>3</span></span>]</span></span>
                if</span> starts_in</span></span> or</span> ends_in</span></span>:</span></span>
                    return</span> True</span>
    return</span> False</span>
</span></code></pre>

The code in this post was tested under python 3.10.4</p>
</blockquote>
We define a genome as a complex nested structure of built-in types (list, tuple, str, int). Then we write a function which will check whether an input read falls on a gene from a given genome. All intervals are 0-based and open.</p>
Even with intuitive variable names (gene</code> and not g</code>) it takes a lot of mental effort to understand the nesting, and translate the numeric indices (e.g. gene[1][1]</code>) into their meaning 🪆. This can make the code harder to understand and maintain.</p>
🥉 Type hints</h2>

TypeAlias was introduced by PEP-613</a> in python 3.10 (2020)</p>
</blockquote>
Despite being a dynamically typed language, python provides type hints to specify the types of variables. Type hints are not enforced at runtime, but they go a long way to help the user (or yourself) understand how to use your code and structures. They are also well integrated with common IDEs, which will display hints and warn you in cases of inconsistencies.</p>

Tip: If you use the Pylance extension in VScode, static analysis does not use type checking by default. To enable it, go to settings > extensions > Pylance</code> and set Analysis: Type checking mode</code> to "strict" or "basic".</p>
</blockquote>
Type hints can be provided to specify the function signature. In the example above, this would be very hard to read due to the nesting:</p>
from</span></span> typing</span></span> import</span></span></span></span></span> Literal</span></span>

def</span></span> read_in_gene</span></span></span>(</span></span>
    read</span></span>:</span> tuple</span></span></span>[</span></span>str</span></span>, int</span></span>, int</span></span>, str</span></span>, Literal</span></span></span>[</span></span>'</span></span></span>+'</span></span></span>, '</span></span></span>-'</span></span></span></span>]</span></span></span>]</span></span></span>,</span>
    genome</span></span>:</span> list</span></span></span>[</span></span>tuple</span></span></span>[</span></span>str</span></span>, int</span></span>, list</span></span></span>[</span></span>str</span></span>, tuple</span></span></span>[</span></span>str</span></span>, int</span></span>, int</span></span>, Literal</span></span></span>[</span></span>'</span></span></span>+'</span></span></span>, '</span></span></span>-'</span></span></span></span>]</span></span></span>]</span></span></span>]</span></span></span>]</span></span></span>]</span></span></span>,</span>
)</span></span> </span>-></span> bool</span></span></span>:</span></span>
    ...</span>
</span></code></pre>
Moreover, if we decided to make a change in one of those structures, we would need to edit the type hints in the signature of every function that uses them. This is where TypeAlias</code> and NewType</code> can be useful:</p>
from</span></span> typing</span></span> import</span></span></span></span></span> Literal</span>,</span> NewType</span>,</span> TypeAlias</span></span>

#</span> Use NewType to denote the meaning of a type
</span>Name</span></span> =</span> NewType</span></span></span></span>(</span>"</span></span></span>Name"</span></span></span>,</span> str</span></span>)</span></span>
Size</span></span> =</span> NewType</span></span></span></span>(</span>"</span></span></span>Size"</span></span></span>,</span> int</span></span>)</span></span>

#</span> Use TypeAlias to define equivalents to complex signatures
</span>Interval</span></span>:</span> TypeAlias</span></span> =</span> tuple</span></span></span>[</span></span>str</span></span>, int</span></span>, int</span></span>, Literal</span></span></span>[</span></span>'</span></span></span>+'</span></span></span>, '</span></span></span>-'</span></span></span></span>]</span></span></span>]</span></span>
Gene</span></span>:</span> TypeAlias</span></span> =</span> tuple</span></span></span>[</span></span>Name</span></span>, Interval</span></span></span>]</span></span>
Chrom</span></span>:</span> TypeAlias</span></span> =</span> tuple</span></span></span>[</span></span>Name</span></span>, Size</span></span>, list</span></span></span>[</span></span>Gene</span></span></span>]</span></span></span>]</span></span>


def</span></span> read_in_gene</span></span></span>(</span></span>read</span></span>:</span> Interval</span></span></span>,</span> genome</span></span>:</span> list</span></span></span>[</span></span>Chrom</span></span></span>]</span></span></span>)</span></span> </span>-></span> bool</span></span></span>:</span></span>
    ...</span>
</span></code></pre>
The function signature is much easier to read, and we know that the genome is represented as a list of chromosomes right away. In an IDE such as VScode or PyCharm, placing the mouse over Chrom</code> will show the alias definition:</p>
(type alias) Chrom: Type[tuple[Name, Size, list[tuple[Name, tuple[str, int, int, Literal['+', '-']]]]]]
</span></code></pre>
If we needed to change some data structures, we would only need to edit the hints in the alias definition.</p>
Type hints are a great way of documenting code, and they can even be combined with tools like mypy</a> or pyright</a> to check for errors! In this example, however, the hints themselves are also hard to read because of the heavily nested structures, we need a better solution.</p>
🥈 Named tuples</h2>

NamedTuple was introduced by PEP-526</a> in python 3.6 (2016)</p>
</blockquote>
Named tuples improve code readability by allowing name-based field access in your tuples. They also inherit regular tuple features:</p>
from</span></span> collections</span></span> import</span></span></span></span></span> namedtuple</span></span>

Gene</span></span> =</span> namedtuple</span></span></span></span>(</span>'</span></span></span>Gene'</span></span></span>,</span> [</span>'</span></span></span>name'</span></span></span>,</span> '</span></span></span>chrom'</span></span></span>,</span> '</span></span></span>start'</span></span></span>,</span> '</span></span></span>end'</span></span></span>,</span> '</span></span></span>sign'</span></span></span>]</span></span>)</span></span>
gene_1</span></span> =</span> Gene</span></span></span></span>(</span>'</span></span></span>Gene1'</span></span></span>,</span> '</span></span></span>chr1'</span></span></span>,</span> 312</span>,</span> 344</span>,</span> '</span></span></span>+'</span></span></span>)</span></span>

#</span> Allows standard tuple operations
</span>len</span></span></span></span>(</span>gene_1</span></span>)</span></span>
chrom</span></span>, start</span></span>, end</span></span> =</span> gene_1</span></span></span>[</span></span>1</span>:</span>4</span></span>]</span></span>

#</span> Adds name-based access semantics
</span>gene_1</span>.</span>end</span></span> -</span> gene_1</span>.</span>start</span></span>
</span></code></pre>
typing.NamedTuple</code> is a typed variant of namedtuple</code> which can be subclassed to take advantage of type annotations mentioned earlier. The definitions are more verbose, but much easier to understand:</p>
from</span></span> typing</span></span> import</span></span></span></span></span> Literal</span>,</span> NamedTuple</span>,</span> Optional</span></span>


class</span></span> Interval</span></span>(</span></span></span>NamedTuple</span>)</span></span>:</span></span>
    """</span>An optionally stranded, 0-based, open genomic interval."""</span></span>

    chrom</span></span>:</span> str</span></span>
    start</span></span>:</span> int</span></span>
    end</span></span>:</span> int</span></span>
    sign</span></span>:</span> Optional</span></span></span>[</span></span>Literal</span></span></span>[</span></span>'</span></span></span>+'</span></span></span>, '</span></span></span>-'</span></span></span></span>]</span></span></span>]</span></span> =</span> None</span>


class</span></span> Gene</span></span>(</span></span></span>NamedTuple</span>)</span></span>:</span></span>
    """</span>A gene, consisting of a named entity with genomic coordinates."""</span></span>

    name</span></span>:</span> str</span></span>
    coord</span></span>:</span> Interval</span></span>


class</span></span> Chrom</span></span>(</span></span></span>NamedTuple</span>)</span></span>:</span></span>
    """</span>A chromosome represented as a collection of genes"""</span></span>

    name</span></span>:</span> str</span></span>
    size</span></span>:</span> int</span></span>
    genes</span></span>:</span> list</span></span></span>[</span></span>Gene</span></span></span>]</span></span>

genes</span></span> =</span> [</span>
    Gene</span></span></span></span>(</span>"</span></span></span>geneA"</span></span></span>,</span> Interval</span></span></span></span>(</span>"</span></span></span>chr1"</span></span></span>,</span> 1002</span>,</span> 1200</span>,</span> "</span></span></span>+"</span></span></span>)</span></span>)</span></span>,</span>
    Gene</span></span></span></span>(</span>"</span></span></span>geneB"</span></span></span>,</span> Interval</span></span></span></span>(</span>"</span></span></span>chr1"</span></span></span>,</span> 39001</span>,</span> 40050</span>,</span> "</span></span></span>-"</span></span></span>)</span></span>)</span></span>,</span>
]</span></span>
chroms</span></span> =</span> [</span>Chrom</span></span></span></span>(</span>"</span></span></span>chr1"</span></span></span>,</span> 141123</span>,</span> [</span>genes</span></span></span>[</span></span>0</span></span>]</span></span>]</span></span>)</span></span>,</span> Chrom</span></span></span></span>(</span>"</span></span></span>chr2"</span></span></span>,</span> 131312</span>,</span> [</span>genes</span></span></span>[</span></span>1</span></span>]</span></span>]</span></span>)</span></span>]</span></span>
read</span></span> =</span> Interval</span></span></span></span>(</span>chrom</span>=</span>"</span></span></span>chr2"</span></span></span>,</span> start</span>=</span>39800</span>,</span> end</span>=</span>40002</span>,</span> sign</span>=</span>"</span></span></span>-"</span></span></span>)</span></span>
</span></code></pre>
In this case, named tuple semantics also make it easier to figure out what's going on in the logic, but it still feels verbose and repetitive: </p>
def</span></span> read_in_gene</span></span></span>(</span></span>read</span></span>:</span> Interval</span></span></span>,</span> genome</span></span>:</span> list</span></span></span>[</span></span>Chrom</span></span></span>]</span></span></span>)</span></span> </span>-></span> bool</span></span></span>:</span></span>
    for</span> chrom</span> in</span></span> genome</span></span></span>:</span></span>
        if</span> chrom</span>.</span>name</span></span> ==</span> read</span>.</span>chrom</span></span>:</span></span>
            for</span> gene</span> in</span></span> chrom</span>.</span>genes</span></span></span>:</span></span>
                starts_in</span></span> =</span> gene</span>.</span>coord</span>.</span>start</span></span> <=</span> read</span>.</span>start</span></span> <=</span> gene</span>.</span>coord</span>.</span>end</span></span>
                ends_in</span></span> =</span> gene</span>.</span>coord</span>.</span>start</span></span> <=</span> read</span>.</span>end</span></span> <=</span> gene</span>.</span>coord</span>.</span>end</span></span>
                same_sign</span></span> =</span> gene</span>.</span>coord</span>.</span>sign</span></span> ==</span> read</span>.</span>sign</span></span>
                if</span> (</span>starts_in</span></span> or</span> ends_in</span></span>)</span></span> and</span> same_sign</span></span>:</span></span>
                    return</span> True</span>
    return</span> False</span>
</span></code></pre>
🥇 Data classes</h2>

dataclass was introduced by PEP-557</a> in python 3.7 (2017)</p>
</blockquote>
Dataclasses offer the flexibility of python classes with minimal administrative overhead. 
First, we can define an interval dataclass. The simplest definition we can write looks almost identical to the previous NamedTuple</code> subclass:</p>
from</span></span> dataclasses</span></span> import</span></span></span></span></span> dataclass</span>,</span> field</span></span>
from</span></span> typing</span></span> import</span></span></span></span></span> Literal</span>,</span> Optional</span></span>

@</span>dataclass</span></span></span></span>
class</span></span> Interval</span></span>:</span></span>
    """</span>An optionally stranded, 0-based, open genomic interval."""</span></span>

    chrom</span></span>:</span> str</span></span>
    start</span></span>:</span> int</span></span>
    end</span></span>:</span> int</span></span>
    sign</span></span>:</span> Optional</span></span></span>[</span></span>Literal</span></span></span>[</span></span>'</span></span></span>+'</span></span></span>, '</span></span></span>-'</span></span></span></span>]</span></span></span>]</span></span> =</span> None</span>

</span></code></pre>
Again, placing the mouse over any occurrence of Interval</code> should show the type hints, along with the docstring annotation:</p>
`(class) Interval(chrom: str, start: int, end: int, sign: Literal['+', '-'] | None = None)`
`An optionally stranded genomic interval.`
</span></code></pre>
Now, let's pretend that we want the ability to sort intervals (i.e. define interval1 > interval2</code>). With a traditional class, this would require writing dunder methods __lt__</code> and/or __gt__</code> to define the behaviours of <</code> and ></code> operators.</p>
With dataclasses, most dunder methods (__init__</code>, __repr__</code>, __eq__</code>, ...) are automatically generated. Their behavior can be customized with arguments to the @dataclass</code> decorator and the field()</code> helper function.</p>
For example, below, we use order=True</code> to say that instances of Interval</code> can be ordered. This will automatically define dunder methods __lt__</code> and __gt__</code>. We then use field(compare=True)</code> to specify which fields should be used in the comparison.</p>
@</span></span></span>dataclass</span></span></span></span>(</span></span>order</span>=</span>True</span></span>)</span></span>
class</span></span> Interval</span></span>:</span></span>
    """</span>An optionally stranded, 0-based, open genomic interval."""</span></span>

    chrom</span></span>:</span> str</span></span> =</span> field</span></span></span></span>(</span>compare</span>=</span>True</span>)</span></span>
    start</span></span>:</span> int</span></span> =</span> field</span></span></span></span>(</span>compare</span>=</span>True</span>)</span></span>
    end</span></span>:</span> int</span></span> =</span> field</span></span></span></span>(</span>compare</span>=</span>True</span>)</span></span>
    sign</span></span>:</span> Optional</span></span></span>[</span></span>Literal</span></span></span>[</span></span>'</span></span></span>+'</span></span></span>, '</span></span></span>-'</span></span></span></span>]</span></span></span>]</span></span> =</span> field</span></span></span></span>(</span>default</span>=</span>None</span>,</span> compare</span>=</span>False</span>)</span></span>
</span></code></pre>
Dunder methods can also be manually implemented for specific operators. For the sake of this example, we are going to define the in</code> operator by defining __contains__</code>. This will be used to define whether another interval (at least partially) falls in the current interval.</p>

For the sake of showing dataclasses, I overrid the in</code> operator, but in practice, this can be confusing. A method like Interval.overlaps(self, other)</code> would be a better choice.</p>
</blockquote>
It makes sense to move this logic into the Interval</code> definition, because we will likely need to reuse it in many different functions.</p>
@</span></span></span>dataclass</span></span></span></span>(</span></span>order</span>=</span>True</span></span>)</span></span>
class</span></span> Interval</span></span>:</span></span>
    """</span>An optionally stranded, 0-based, open genomic interval."""</span></span>

    chrom</span></span>:</span> str</span></span> =</span> field</span></span></span></span>(</span>compare</span>=</span>True</span>)</span></span>
    start</span></span>:</span> int</span></span> =</span> field</span></span></span></span>(</span>compare</span>=</span>True</span>)</span></span>
    end</span></span>:</span> int</span></span> =</span> field</span></span></span></span>(</span>compare</span>=</span>True</span>)</span></span>
    sign</span></span>:</span> Optional</span></span></span>[</span></span>Literal</span></span></span>[</span></span>'</span></span></span>+'</span></span></span>, '</span></span></span>-'</span></span></span></span>]</span></span></span>]</span></span> =</span> field</span></span></span></span>(</span>default</span>=</span>None</span>,</span> compare</span>=</span>False</span>)</span></span>

    def</span></span> __contains__</span></span></span>(</span></span>self</span>,</span> other</span></span>:</span> Interval</span></span></span>)</span></span> </span>-></span> bool</span></span></span>:</span></span>
        """</span>Checks if another interval overlaps with self."""</span></span>
        starts_in</span></span> =</span> self</span>.</span>start</span></span> <=</span> other</span>.</span>start</span></span> <=</span> self</span>.</span>end</span></span>
        ends_in</span></span> =</span> self</span>.</span>start</span></span> <=</span> other</span>.</span>end</span></span> <=</span> self</span>.</span>end</span></span>
        same_sign</span></span> =</span> self</span>.</span>sign</span></span> ==</span> other</span>.</span>sign</span></span>
        return</span> (</span>starts_in</span></span> or</span> ends_in</span></span>)</span></span> and</span> same_sign</span></span>
</span></code></pre>
Now we can sort intervals and check for overlap!</p>
></span>></span>></span> read_1</span></span> =</span> Interval</span></span></span></span>(</span>'</span></span></span>c1'</span></span></span>,</span> 10</span>,</span> 15</span>)</span></span>
></span>></span>></span> read_2</span></span> =</span> Interval</span></span></span></span>(</span>'</span></span></span>c1'</span></span></span>,</span> 12</span>,</span> 17</span>)</span></span>
></span>></span>></span> read_1</span></span> in</span> read_2</span></span>
True</span>
></span>></span>></span> sorted</span></span></span></span>(</span>[</span>read_2</span></span>,</span> read_1</span></span>]</span></span>)</span></span>
[</span>Interval</span></span></span></span>(</span>chrom</span>=</span>'</span></span></span>c1'</span></span></span>,</span> start</span>=</span>10</span>,</span> end</span>=</span>15</span>,</span> sign</span>=</span>None</span>)</span></span>,</span>
 Interval</span></span></span></span>(</span>chrom</span>=</span>'</span></span></span>c1'</span></span></span>,</span> start</span>=</span>12</span>,</span> end</span>=</span>17</span>,</span> sign</span>=</span>None</span>)</span></span>]</span></span>
</span></code></pre>
We can reuse this Interval</code> to define Gene</code> and Chrom</code>. Since a gene is just a named interval here, we can subclass</em> Interval</code> to inherit all its attributes and methods, and simply add an optional name</code> attribute (or any gene-specific behavior we may want):</p>
Since a chromosome contains a set of genes (which are a subclass of interval), we can easily define a helper method to check if a given interval overlaps any gene in a chromosome. This is shown below in Chromosome.gene_overlaps()</code>.</p>
@</span></span></span>dataclass</span></span></span></span>(</span></span>order</span>=</span>True</span></span>)</span></span>
class</span></span> Gene</span></span>(</span></span></span>Interval</span>)</span></span>:</span></span>
    """</span>A named genomic interval."""</span></span>

    name</span></span>:</span> Optional</span></span></span>[</span></span>str</span></span></span>]</span></span> =</span> field</span></span></span></span>(</span>default</span>=</span>None</span>,</span> compare</span>=</span>False</span>)</span></span>


@</span>dataclass</span></span></span></span>
class</span></span> Chrom</span></span>:</span></span>
    """</span>A chromosome, represented as a collection of named genomic intervals."""</span></span>

    name</span></span>:</span> str</span></span>
    size</span></span>:</span> int</span></span>
    genes</span></span>:</span> list</span></span></span>[</span></span>Gene</span></span></span>]</span></span>
    
    def</span></span> gene_overlaps</span></span></span>(</span></span>self</span>,</span> interval</span></span>:</span> Interval</span></span></span>)</span></span> </span>-></span> bool</span></span></span>:</span></span>
        for</span> gene</span> in</span></span> self</span>.</span>genes</span></span></span>:</span></span>
            if</span> interval</span></span> in</span> gene</span></span>:</span></span>
                return</span> True</span>
        return</span> False</span>
</span></code></pre>
With the main overlap logic taken out, this function is much easier to read than the original version. </p>
gene_1</span></span> =</span> Gene</span></span></span></span>(</span>name</span>=</span>"</span></span></span>Gene1"</span></span></span>,</span> chrom</span>=</span>"</span></span></span>chr1"</span></span></span>,</span> start</span>=</span>1000</span>,</span> end</span>=</span>3000</span>,</span> sign</span>=</span>"</span></span></span>-"</span></span></span>)</span></span>
gene_2</span></span> =</span> Gene</span></span></span></span>(</span>name</span>=</span>"</span></span></span>Gene2"</span></span></span>,</span> chrom</span>=</span>"</span></span></span>chr1"</span></span></span>,</span> start</span>=</span>3500</span>,</span> end</span>=</span>4100</span>,</span> sign</span>=</span>"</span></span></span>-"</span></span></span>)</span></span>
chrom</span></span> =</span> Chrom</span></span></span></span>(</span>"</span></span></span>chr1"</span></span></span>,</span> 11000</span>,</span> genes</span>=</span>[</span>gene_1</span></span>,</span> gene_2</span></span>]</span></span>)</span></span>
read</span></span> =</span> Interval</span></span></span></span>(</span>"</span></span></span>chr1"</span></span></span>,</span> 2500</span>,</span> 3200</span>,</span> "</span></span></span>-"</span></span></span>)</span></span>


def</span></span> read_in_gene</span></span></span>(</span></span>read</span></span>:</span> Interval</span></span></span>,</span> genome</span></span>:</span> list</span></span></span>[</span></span>Chrom</span></span></span>]</span></span></span>)</span></span> </span>-></span> bool</span></span></span>:</span></span>
    for</span> chrom</span> in</span></span> genome</span></span></span>:</span></span>
        if</span> chrom</span></span>.</span></span>gene_overlaps</span></span></span></span>(</span>read</span></span>)</span></span>:</span></span>
            return</span> True</span>
    return</span> False</span>


read_in_gene</span></span></span></span>(</span>read</span></span>,</span> [</span>chrom</span></span>]</span></span>)</span></span>
</span></code></pre>
dataclasses are more flexible than named tuples. However, they are not as fast and don't inherit the default tuple behaviors (hashable, iterable, unpackable). For a more detailed comparison, see this stackoverflow answer by Oleksandr Yarushevski's</a>.</p>
Here is the original code for comparison:</p>

original:</summary>
genes</span></span> =</span> [</span>
    (</span>"</span></span></span>geneA"</span></span></span>,</span> (</span>"</span></span></span>chr1"</span></span></span>,</span> 1002</span>,</span> 1200</span>,</span> "</span></span></span>+"</span></span></span>)</span></span>)</span></span>,</span>
    (</span>"</span></span></span>geneB"</span></span></span>,</span> (</span>"</span></span></span>chr2"</span></span></span>,</span> 39001</span>,</span> 39950</span>,</span> "</span></span></span>-"</span></span></span>)</span></span>)</span></span>,</span>
]</span></span>
genome</span></span> =</span> [</span>(</span>"</span></span></span>chr1"</span></span></span>,</span> 141123</span>,</span> [</span>genes</span></span></span>[</span></span>0</span></span>]</span></span>]</span></span>)</span></span>,</span> (</span>"</span></span></span>chr2"</span></span></span>,</span> 131312</span>,</span> [</span>genes</span></span></span>[</span></span>1</span></span>]</span></span>]</span></span>)</span></span>]</span></span>
read</span></span> =</span> (</span>"</span></span></span>chr2"</span></span></span>,</span> 39800</span>,</span> 40002</span>,</span> "</span></span></span>-"</span></span></span>)</span></span>

def</span></span> read_in_gene</span></span></span>(</span></span>read</span>,</span> genome</span>)</span></span>:</span></span>
    for</span> chrom</span> in</span></span> genome</span></span></span>:</span></span>
        if</span> chrom</span></span></span>[</span></span>0</span></span>]</span></span> ==</span> read</span></span></span>[</span></span>0</span></span>]</span></span>:</span></span>
            for</span> gene</span> in</span></span> chrom</span></span></span>[</span></span>2</span></span>]</span></span></span>:</span></span>
                starts_in</span></span> =</span> gene</span></span></span>[</span></span>1</span></span>]</span></span>[</span></span>1</span></span>]</span></span> <=</span> read</span></span></span>[</span></span>1</span></span>]</span></span> <=</span> gene</span></span></span>[</span></span>1</span></span>]</span></span>[</span></span>2</span></span>]</span></span>
                ends_in</span></span> =</span> gene</span></span></span>[</span></span>1</span></span>]</span></span>[</span></span>1</span></span>]</span></span> <=</span> read</span></span></span>[</span></span>2</span></span>]</span></span> <=</span> gene</span></span></span>[</span></span>1</span></span>]</span></span>[</span></span>2</span></span>]</span></span>
                same_sign</span></span> =</span> gene</span></span></span>[</span></span>1</span></span>]</span></span>[</span></span>3</span></span>]</span></span> ==</span> read</span></span></span>[</span></span>3</span></span>]</span></span>
                if</span> starts_in</span></span> or</span> ends_in</span></span>:</span></span>
                    return</span> True</span>
    return</span> False</span>
</span></code></pre>
</details>
In summary, the complicated logic has been moved into the structure definitions, and for the better: each function is more modular (easier to read), and we can reuse them anytime we need new functions, instead of rewriting similar logic.</p>
Using these features introduces the overhead of defining data structures. In this example, named tuples would probably be a good middle ground. For larger projects and algorithms, these definitions and the associated documentation (docstrings and type hints) represent a small cost compared to the long term gain in efficiency and clarity.</p>
🍒 pydantic</h2>

pydantic</a> (1.10.4) compatible with python 3.7-3.11</p>
</blockquote>
This package is not part of python's standard library, but I really like it, and it is a very useful package if you need to serialize your classes.</p>
It provides a BaseModel</code> class with serialization methods such as .dict()</code>, .json()</code>, as well as additional constrained types to validate instances.</p>
In the example below, we define Interval</code> as a subclass of BaseModel</code> to access its neat features:</p>

Constrained types: note start</code> and end</code> are of type pos</code>, which we define as a constrained non-negative integer type with conint</code>.</li>
Instance validation: Until now, type hints were only informative. pydantic actually enforces type hints at instantiation time. If we attempt to create a gene with a negative coordinates, or add genes in a chromosome with a different chrom</code> attribute than the chromosome name</code>, it will raise an exception with a clear message.</li>
</ul>
Here is the full example rewritten with pydantic:</p>
from</span></span> typing</span></span> import</span></span></span></span></span> Any</span>,</span> Literal</span>,</span> Optional</span></span>
from</span></span> pydantic</span></span> import</span></span></span></span></span> BaseModel</span>,</span> conint</span>,</span> root_validator</span>,</span> NonNegativeInt</span></span>

#</span> Genomic positions can not be negative
</span>pos</span></span> =</span> conint</span></span></span></span>(</span>ge</span>=</span>0</span>)</span></span>


class</span></span> Interval</span></span>(</span></span></span>BaseModel</span>)</span></span>:</span></span>
    """</span>An optionally stranded, 0-based, open genomic interval."""</span></span>

    chrom</span></span>:</span> str</span></span>
    start</span></span>:</span> pos</span></span>
    end</span></span>:</span> pos</span></span>
    sign</span></span>:</span> Optional</span></span></span>[</span></span>Literal</span></span></span>[</span></span>"</span></span></span>-"</span></span></span>, "</span></span></span>+"</span></span></span></span>]</span></span></span>]</span></span>

    def</span></span> __contains__</span></span></span>(</span></span>self</span>,</span> other</span></span>:</span> Interval</span></span></span>)</span></span> </span>-></span> bool</span></span></span>:</span></span>
        """</span>Checks if another interval overlaps with self."""</span></span>
        starts_in</span></span> =</span> self</span>.</span>start</span></span> <=</span> other</span>.</span>start</span></span> <=</span> self</span>.</span>end</span></span>
        ends_in</span></span> =</span> self</span>.</span>start</span></span> <=</span> other</span>.</span>end</span></span> <=</span> self</span>.</span>end</span></span>
        same_sign</span></span> =</span> self</span>.</span>sign</span></span> ==</span> other</span>.</span>sign</span></span>
        return</span> starts_in</span></span> or</span> ends_in</span></span> and</span> same_sign</span></span>


class</span></span> Gene</span></span>(</span></span></span>Interval</span>)</span></span>:</span></span>
    """</span>A named genomic interval."""</span></span>

    name</span></span>:</span> Optional</span></span></span>[</span></span>str</span></span></span>]</span></span>


class</span></span> Chrom</span></span>(</span></span></span>BaseModel</span>)</span></span>:</span></span>
    """</span>A chromosome, represented as a collection of named genomic intervals."""</span></span>

    name</span></span>:</span> str</span></span>
    size</span></span>:</span> NonNegativeInt</span></span>
    genes</span></span>:</span> list</span></span></span>[</span></span>Gene</span></span></span>]</span></span> =</span> [</span>]</span></span>

    @</span></span></span>root_validator</span></span></span></span>(</span></span>allow_reuse</span>=</span>True</span></span>)</span></span>
    def</span></span> genes_on_chrom</span></span></span>(</span></span>cls</span>,</span> values</span></span>:</span> dict</span></span></span>[</span></span>str</span></span>, Any</span></span></span>]</span></span></span>)</span></span> </span>-></span> dict</span></span></span>[</span></span>str</span></span>, Any</span></span></span>]</span></span></span>:</span></span>
        """</span>Checks that input genes are on the right chromosome."""</span></span>
        genes</span></span> =</span> values</span></span>.</span></span>get</span></span></span></span>(</span>"</span></span></span>genes"</span></span></span>)</span></span>
        curr_chrom</span></span> =</span> values</span></span>.</span></span>get</span></span></span></span>(</span>"</span></span></span>name"</span></span></span>)</span></span>
        if</span> genes</span></span> is</span> None</span>:</span></span>
            return</span> values</span></span>
        for</span> gene</span> in</span></span> genes</span></span></span>:</span></span>
            if</span> gene</span>.</span>chrom</span></span> !=</span> curr_chrom</span></span>:</span></span>
                raise</span> ValueError</span></span></span></span>(</span>
                    f</span>"</span></span></span>{</span>gene</span>.</span>name</span></span></span></span>}</span></span> is on </span>{</span>gene</span>.</span>chrom</span></span></span></span>}</span></span> but should be on </span>{</span>curr_chrom</span></span></span></span>}</span></span>"</span></span></span>
                )</span></span></span>
        return</span> values</span></span>
</span></code></pre>
If we try to enter a gene with negative coordinates, we get a clear error:</p>
></span>></span>></span> gene_1</span></span> =</span> Gene</span></span></span></span>(</span>name</span>=</span>"</span></span></span>Gene1"</span></span></span>,</span> chrom</span>=</span>"</span></span></span>chr1"</span></span></span>,</span> start</span>=</span>-</span>1</span>,</span> end</span>=</span>3000</span>,</span> sign</span>=</span>"</span></span></span>-"</span></span></span>)</span></span>
[</span>...</span>]</span></span>
pydantic</span>.</span>error_wrappers</span>.</span>ValidationError</span></span>:</span> 1</span> validation</span></span> error</span></span> for</span> Gene</span></span>
start</span></span>
  ensure</span></span> this</span></span> value</span></span> is</span> greater</span></span> than</span></span> or</span> equal</span></span> to</span></span> 0</span> (</span>type</span></span>=value_error</span>.</span>number</span>.</span>not_ge</span></span>; limit_value</span></span>=0</span>)</span></span>
</span></code></pre>
Similarly, we cannot enter a gene from the chr1 into chr2:</p>
></span>></span>></span> gene_1</span></span> =</span> Gene</span></span></span></span>(</span>name</span>=</span>"</span></span></span>Gene1"</span></span></span>,</span> chrom</span>=</span>"</span></span></span>chr1"</span></span></span>,</span> start</span>=</span>-</span>1</span>,</span> end</span>=</span>3000</span>,</span> sign</span>=</span>"</span></span></span>-"</span></span></span>)</span></span>
></span>></span>></span> chrom_2</span></span> =</span> Chrom</span></span></span></span>(</span>name</span>=</span>"</span></span></span>chr2"</span></span></span>,</span> size</span>=</span>11000</span>,</span> genes</span>=</span>[</span>gene_1</span></span>]</span></span>)</span></span>
pydantic</span>.</span>error_wrappers</span>.</span>ValidationError</span></span>:</span> 1</span> validation</span></span> error</span></span> for</span> Chrom</span></span>
__root__</span></span>
  Gene2</span></span> is</span> on</span></span> chr2</span></span> but</span></span> should</span></span> be</span></span> on</span></span> chr1</span></span></span> </span>(</span>type</span>=</span>value_error</span></span>)</span></span>[</span></span>...</span></span>]</span></span>
</span></code></pre>
Finally, pydantic allows serializing the instance to json. This is great for interoperability, as many tools are compatible with this format, and it is also simple to convert common genomic formats to or from json:</p>
></span>></span>></span> gene_1</span></span> =</span> Gene</span></span></span></span>(</span>name</span>=</span>"</span></span></span>Gene1"</span></span></span>,</span> chrom</span>=</span>"</span></span></span>chr1"</span></span></span>,</span> start</span>=</span>1000</span>,</span> end</span>=</span>3000</span>,</span> sign</span>=</span>"</span></span></span>-"</span></span></span>)</span></span>
></span>></span>></span> gene_2</span></span> =</span> Gene</span></span></span></span>(</span>name</span>=</span>"</span></span></span>Gene2"</span></span></span>,</span> chrom</span>=</span>"</span></span></span>chr1"</span></span></span>,</span> start</span>=</span>3500</span>,</span> end</span>=</span>4100</span>,</span> sign</span>=</span>"</span></span></span>-"</span></span></span>)</span></span>
></span>></span>></span> chrom</span></span> =</span> Chrom</span></span></span></span>(</span>name</span>=</span>"</span></span></span>chr1"</span></span></span>,</span> size</span>=</span>11000</span>,</span> genes</span>=</span>[</span>gene_1</span></span>,</span> gene_2</span></span>]</span></span>)</span></span>
></span>></span>></span> print</span></span></span></span>(</span>chrom</span></span>.</span></span>json</span></span></span></span>(</span>indent</span>=</span>2</span>)</span></span>)</span></span>
{</span>
</span>  </span>"</span></span></span>name"</span></span></span></span>:</span></span> </span>"</span></span></span>chr1"</span></span></span></span>,</span>
  </span>"</span></span></span>size"</span></span></span></span>:</span></span> </span>11000</span></span>,</span>
  </span>"</span></span></span>genes"</span></span></span></span>:</span></span> </span>[</span>
    {</span>
</span>      </span>"</span></span></span>chrom"</span></span></span></span>:</span></span> </span>"</span></span></span>chr1"</span></span></span></span>,</span>
      </span>"</span></span></span>start"</span></span></span></span>:</span></span> </span>1000</span></span>,</span>
      </span>"</span></span></span>end"</span></span></span></span>:</span></span> </span>3000</span></span>,</span>
      </span>"</span></span></span>sign"</span></span></span></span>:</span></span> </span>"</span></span></span>-"</span></span></span></span>,</span>
      </span>"</span></span></span>name"</span></span></span></span>:</span></span> </span>"</span></span></span>Gene1"</span></span></span>
</span>    }</span></span>,</span>
    {</span>
</span>      </span>"</span></span></span>chrom"</span></span></span></span>:</span></span> </span>"</span></span></span>chr1"</span></span></span></span>,</span>
      </span>"</span></span></span>start"</span></span></span></span>:</span></span> </span>3500</span></span>,</span>
      </span>"</span></span></span>end"</span></span></span></span>:</span></span> </span>4100</span></span>,</span>
      </span>"</span></span></span>sign"</span></span></span></span>:</span></span> </span>"</span></span></span>-"</span></span></span></span>,</span>
      </span>"</span></span></span>name"</span></span></span></span>:</span></span> </span>"</span></span></span>Gene2"</span></span></span>
</span>    }</span></span>
  ]</span></span>
</span>}</span></span>
</span></code></pre>
Similarly, the parse_file method can instantiate the class directly from a json file!</p>
></span>></span>></span> Chrom</span></span>.</span></span>parse_file</span></span></span></span>(</span>'</span></span></span>test.json'</span></span></span>)</span></span>
Chrom</span></span></span></span>(</span>name</span>=</span>'</span></span></span>chr1'</span></span></span>,</span> size</span>=</span>11000</span>,</span> genes</span>=</span>[</span>Gene</span></span></span></span>(</span>chrom</span>=</span>'</span></span></span>chr1'</span></span></span>,</span> start</span>=</span>1000</span>,</span> end</span>=</span>3000</span>,</span> sign</span>=</span>'</span></span></span>-'</span></span></span>,</span> name</span>=</span>'</span></span></span>Gene1'</span></span></span>)</span></span>,</span> Gene</span></span></span></span>(</span>chrom</span>=</span>'</span></span></span>chr1'</span></span></span>,</span> start</span>=</span>3500</span>,</span> end</span>=</span>4100</span>,</span> sign</span>=</span>'</span></span></span>-'</span></span></span>,</span> name</span>=</span>'</span></span></span>Gene2'</span></span></span>)</span></span>]</span></span>)</span></span>
</span></code></pre>
The object can also be converted into a json schema using Chrom.schema_json(). This can be very useful for validating data in other tools.</p>

json schema</summary>
{</span>
  </span>"</span>title"</span></span></span>:</span> </span>"</span>Chrom"</span></span>,</span>
  </span>"</span>description"</span></span></span>:</span> </span>"</span>A chromosome, represented as a collection of named genomic intervals."</span></span>,</span>
  </span>"</span>type"</span></span></span>:</span> </span>"</span>object"</span></span>,</span>
  </span>"</span>properties"</span></span></span>:</span> </span>{</span>
    </span>"</span>name"</span></span></span>:</span> </span>{</span>
      </span>"</span>title"</span></span></span>:</span> </span>"</span>Name"</span></span>,</span>
      </span>"</span>type"</span></span></span>:</span> </span>"</span>string"</span></span>
    }</span></span>,</span>
    </span>"</span>size"</span></span></span>:</span> </span>{</span>
      </span>"</span>title"</span></span></span>:</span> </span>"</span>Size"</span></span>,</span>
      </span>"</span>minimum"</span></span></span>:</span> </span>0</span>,</span>
      </span>"</span>type"</span></span></span>:</span> </span>"</span>integer"</span></span>
    }</span></span>,</span>
    </span>"</span>genes"</span></span></span>:</span> </span>{</span>
      </span>"</span>title"</span></span></span>:</span> </span>"</span>Genes"</span></span>,</span>
      </span>"</span>default"</span></span></span>:</span> </span>[</span>]</span></span>,</span>
      </span>"</span>type"</span></span></span>:</span> </span>"</span>array"</span></span>,</span>
      </span>"</span>items"</span></span></span>:</span> </span>{</span>
        </span>"</span>$ref"</span></span></span>:</span> </span>"</span>#/definitions/Gene"</span></span>
      }</span></span>
    }</span></span>
  }</span></span>,</span>
  </span>"</span>required"</span></span></span>:</span> </span>[</span>
    "</span>name"</span></span>,</span>
    "</span>size"</span></span>
  ]</span></span>,</span>
  </span>"</span>definitions"</span></span></span>:</span> </span>{</span>
    </span>"</span>Gene"</span></span></span>:</span> </span>{</span>
      </span>"</span>title"</span></span></span>:</span> </span>"</span>Gene"</span></span>,</span>
      </span>"</span>description"</span></span></span>:</span> </span>"</span>A named genomic interval."</span></span>,</span>
      </span>"</span>type"</span></span></span>:</span> </span>"</span>object"</span></span>,</span>
      </span>"</span>properties"</span></span></span>:</span> </span>{</span>
        </span>"</span>chrom"</span></span></span>:</span> </span>{</span>
          </span>"</span>title"</span></span></span>:</span> </span>"</span>Chrom"</span></span>,</span>
          </span>"</span>type"</span></span></span>:</span> </span>"</span>string"</span></span>
        }</span></span>,</span>
        </span>"</span>start"</span></span></span>:</span> </span>{</span>
          </span>"</span>title"</span></span></span>:</span> </span>"</span>Start"</span></span>,</span>
          </span>"</span>minimum"</span></span></span>:</span> </span>0</span>,</span>
          </span>"</span>type"</span></span></span>:</span> </span>"</span>integer"</span></span>
        }</span></span>,</span>
        </span>"</span>end"</span></span></span>:</span> </span>{</span>
          </span>"</span>title"</span></span></span>:</span> </span>"</span>End"</span></span>,</span>
          </span>"</span>minimum"</span></span></span>:</span> </span>0</span>,</span>
          </span>"</span>type"</span></span></span>:</span> </span>"</span>integer"</span></span>
        }</span></span>,</span>
        </span>"</span>sign"</span></span></span>:</span> </span>{</span>
          </span>"</span>title"</span></span></span>:</span> </span>"</span>Sign"</span></span>,</span>
          </span>"</span>enum"</span></span></span>:</span> </span>[</span>
            "</span>-"</span></span>,</span>
            "</span>+"</span></span>
          ]</span></span>,</span>
          </span>"</span>type"</span></span></span>:</span> </span>"</span>string"</span></span>
        }</span></span>,</span>
        </span>"</span>name"</span></span></span>:</span> </span>{</span>
          </span>"</span>title"</span></span></span>:</span> </span>"</span>Name"</span></span>,</span>
          </span>"</span>type"</span></span></span>:</span> </span>"</span>string"</span></span>
        }</span></span>
      }</span></span>,</span>
      </span>"</span>required"</span></span></span>:</span> </span>[</span>
        "</span>chrom"</span></span>,</span>
        "</span>start"</span></span>,</span>
        "</span>end"</span></span>
      ]</span></span>
    }</span></span>
  }</span></span>
}</span></span>
</span></code></pre>
</details>


Welcome to my website
2022-11-02T00:00:00+00:00
Hi, I'm Cyril and this is my website! I may write about open data, programming, bioinformatics or various tech-related things.</p>
I made this website to share thoughts, programming tips and projects I'm working on, and things I learn along the way.</p>
Technical notes about this website</h3>
This site is free and open source software licensed under GPL-3.0</a>, the source code is available from GitHub</a>. It is built with the Zola</a> static site generator with the Abridge</a> theme.</p>
Cyril Matthey-Doret

Task Automation: Just Make it Easy

Resources</h1> https://theorangeone.net/posts/just-stop-using-makefile/</p>

Simplifying Research Code With Modern Python Features

🪆 The problem</h2> A few issues may arise when over-relying on built-ins to represent complex domain-specific models:</p> Lack of custom interfaces can force more complex logic or repetitions</li> Code becomes harder to read due to the absence of explicitly named fields and methods</li>

Welcome to my website

Technical notes about this website</h3> This site is free and open source software licensed under GPL-3.0</a>, the source code is available from GitHub</a>. It is built with the Zola</a> static site generator with the Abridge</a> theme.</p>

`Resources</h1> https://theorangeone.net/posts/just-stop-using-makefile/</p>`

Technical notes about this website</h3>
This site is free and open source software licensed under GPL-3.0</a>, the source code is available from GitHub</a>. It is built with the Zola</a> static site generator with the Abridge</a> theme.</p>
