Welcome to crocodile
Fill your life with one-liners, take your code to artistic level of brevity and readability while simultaneously being more productive by typing less boilerplate lines of code that are needless to say.
This package extends many native Python classes to equip you with an uneasy-to-tame power. The major classes extended are:
-
listis extended toList- Forget that
forloops exist, because with this class,forloops are implicitly used to apply a function to all items. Inevitably while programming, one will encounter objects of the same type and you will be struggling to get a tough grab on them.Listis a powerful structure that put at your disposal a grip, so tough, that the objects you have at hand start behaving like one object. Behaviour is ala-JavaScript implementation offorEachmethod of Arrays.
- Forget that
-
dictis extended toStruct.- Combines the power of dot notation like classes and key access like dictionaries.
-
pathlib.Pathis extended toPPobjects are incredibly powerful for parsing paths, no more than one line of code is required to do any operation. Take a shufti at this:
path = tb.P("dataset/type1/meta/images/file3.ext") >> path[0] # allows indexing! P("dataset") >> path[-1] # nifty! P("file3.ext") >> path[2:-1] # even slicing! P("meta/images/file3.ext")This and much more, is only on top of the indespensible
pathlib.Pathfunctionalities. -
Additionally, the package provides many other new classes, e.g.
ReadandSave. Together withP, they provide comprehensive support for file management. Life cannot get easier with those. Every class inherits attributes that allow saving and loading in one line.
Furthermore, those classes are inextricably connected. For example, globbing a path P object returns a List object. You can move back and forth between List and Struct and DataFrame with one method, and so on.
- Deep Learning Modules.
- A paradigm that facilitates working with deep learning models that is based on a tri-partite scheme:
- HyperParameters: facilitated through
HParamsclass. - Data: facilitated though
DataReaderclass. BaseModelis a frontend for bothTensorFlow&Pytorchbackends. The wrapper worked in tandem.
- HyperParameters: facilitated through
- The aforementioned classes cooperate together to offer sealmess workflow during creation, training, and saving models.
- A paradigm that facilitates working with deep learning models that is based on a tri-partite scheme:
Install
In the commandline:
pip install crocodile.
Being a thin extension on top of almost pure Python, you need to worry not about your venv, the package is not aggressive in requirements, it installs itself peacefully, never interfere with your other packages. If you do not have numpy, matplotlib and pandas, it simply throws ImportError at runtime, that's it.
Getting Started
That's as easy as taking candy from a baby; whenever you start a Python file, preface it with following in order to unleash the library:
import crocodile.toolbox as tb
A Taste of Power
Suppose you want to know how many lines of code in your repository. The procedure is to glob all .py files recursively, read string code, split each one of them by lines, count the lines, add up everything from all strings of code.
To achieve this, all you need is an eminently readable one-liner.
tb.P.cwd().search("*.py", r=True).read_text().split('\n').apply(len).to_numpy().sum()
How does this make perfect sense?
searchreturnsListofPpath objectsread_textis aPmethod, but it is being run againstListobject. Behind the scenes, responsible black magic fails to find such a method inListand realizes it is a method of items inside the list, so it runs it against them and thus read all files and containerize them in anotherListobject and returns it.- A similar story applies to
splitwhich is a method of strings in Python. - Next,
applyis a method ofList. Sure enough, it lives up to its apt name and applies the passed functionlento all items in the list and returns anotherListobject that contains the results. .to_numpy()convertsListtonumpyarray, then.sumis a method ofnumpy, which gives the final result.
Methods naming convention like apply and to_numpy are inspired from the popular pandas library, resulting in almost non-existing learning curve.
Friendly interactive tutorial.
Please refer to Here on the main git repo.
Full docs:
Click Here
Author
Alex Al-Saffar. email
1 maintainer
Metadata
1 maintainer
Release files for crocodile 5.6
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | |
|---|---|---|---|
| crocodile-5.6.tar.gz | 82.5 kB | Details |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Reset |
|---|---|---|---|---|
| crocodile-5.6-py3-none-any.whl | Python 3 | none | any | Details |
Total release size: 166.9 kB
Release files / crocodile-5.6.tar.gz
| Download URL | crocodile-5.6.tar.gz |
|---|---|
| Size | 82.5 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
ad8b4970872a813b70dbc1db6c813acd8b9f36c316a97786e4148e7670a8efcd
|
|
BLAKE2b-256 checksum How to use checksums |
8f1d0a637867180f18f60b9acc16bf497aab5f9feeec5aa6591255aa6c8f3815
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/3.7.1 importlib_metadata/4.10.1 pkginfo/1.8.2 requests/2.27.1 requests-toolbelt/0.9.1 tqdm/4.62.3 CPython/3.9.10
|
Release files / crocodile-5.6-py3-none-any.whl
| Download URL | crocodile-5.6-py3-none-any.whl |
|---|---|
| Size | 84.4 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
c57c3508552a8f0b138170cd971af981b39dd5fcceaeb4c9c589f2f24a3ecb39
|
|
BLAKE2b-256 checksum How to use checksums |
1b73aa16ac430d2b9d6395a16443aa11b6c99365a5479519c688d7283ee44772
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/3.7.1 importlib_metadata/4.10.1 pkginfo/1.8.2 requests/2.27.1 requests-toolbelt/0.9.1 tqdm/4.62.3 CPython/3.9.10
|
