I've created an introductory post where I looked at the packages meta data in a very superficial way.
In the second post I analyzed the relationships between packages by looking at a dependency graph.
Now, in the beginning of 2020, I've updated the code and ran my analysis again. Have fun 🙂
Out of 208 492 packages, only 176 107 packages are used.
There are 90,682 authors of Python packages. The ten most active authors are:
| Author | Created packages |
|---|---|
| 8948 | |
| UNKNOWN | 1399 |
| The Guardians | 1140 |
| Tecnativa, Odoo Community Association (OCA) | 554 |
| OpenStack | 466 |
| hfpython | 460 |
| Paul Sokolovsky | 450 |
| Microsoft Corporation | 394 |
| Eficent, Odoo Community Association (OCA) | 329 |
| Akretion,Odoo Community Association (OCA) | 290 |
Maximum length
| Column | NULL values | Maximum Length | Entry |
|---|---|---|---|
| name | 0 | 80 | Aaaaaaaaaaaaaaaaaaa-aaaaaaaaa-aaaaaaasa-aaaaaaasa-aaaaasaa-aaaaaaasa-bbbbbbbbbbb |
| version | 560 | 70 | 1.0dev-BZR-r115-panta-elasticworld.org-20100520155735-sf3yrsr0pvyvlm8m |
| stable_version | 560 | 7 | UNKNOWN |
| release_url | 560 | 128 | https://pypi.org/project/softwarefabrica.django.appserver/1.0dev-BZR-r10-panta-elasticworld.org-20091023132843-vitk6k7e5qlvhej5/ |
| package_url | 560 | 106 | https://pypi.org/project/Aaaaaaaaaaaaaaaaaaa-aaaaaaaaa-aaaaaaasa-aaaaaaasa-aaaaasaa-aaaaaaasa-bbbbbbbbbbb/ |
| bugtrack_url | 560 | 4 | None |
| summary | 7717 | 210 | 采集百度手机卫士 360手机卫士 搜狗号码通三个平台对手机号码的标注信息. 1.1.0: 加入搜狗API采集方式，但与从网页采集结果是一致的。 1.2.0：采用多线程，将不同源分开采集，避免互相影响。 1.3.1：加入命令行工具，加入对IP的处理工具。 1.3.3：可以使用--file xxx.csv 命令指定文件。 1.3.4：改用百度和360手机版本，减少下载量；尝试遇到反爬退出。 1.3.5：完善使用代理下载。 |
| home_page | 17927 | 150 | http://127.0.0.1:8888 ... |
| author | 8948 | 255 | ['Diemer, Jens', 'Hall, Dave', 'van der Boor, Diederik', 'Chainz, Adam', 'Cornehl, Denis', 'Fahed, Sami', 'Hoppe, Johannes', 'Iakovlev, Maksim', 'Lagovas', 'Larin, Nikita', 'Lavin, Mark', 'Lisák, Peter', 'Marques de Araújo, Rodrigo Pinheiro', 'Mihelac, Bo |
| author_email | 16619 | 255 | [email protected], [email protected], [email protected], [email protected], [email protected], [email protected], [email protected], [email protected], [email protected], [email protected], [email protected] |
| license | 44374 | 255 | Copyright © 2012 Kenan Yildirim.[...] |
| keywords | 92634 | 4033 | algorithmic,algotrading,altcoin,altcoins,api,arbitrage,backtest,backtesting,bitcoin,bot,... |
| requires_python | 112090 | 94 | >=2.7, >=2.7.1, !=3.0, !=3.0.*, !=3.1, !=3.1.*, !=3.2, !=3.2.*, !=3.3, !=3.3.*, !=3.4, !=3.4.* |
| maintainer | 140232 | 145 | Brian Drawert, Kevin Sanft, Sean Matthew, George Hall, Dalton Nickerson, Samuel Hodges, Emma Weisgerber, Eliot Dixon, Ghilman Brock, W.R. Jackson |
| maintainer_email | 140745 | 137 | [email protected], [email protected], [email protected], [email protected] |
| platform | 126048 | 255 | [spam] |
| download_url | 129479 | 183 | softwarefabrica.django.appserver |
Platforms
| Platform | Count |
|---|---|
| 126048 | |
| UNKNOWN | 61530 |
| ANY | 12163 |
| LINUX | 1644 |
| ALL | 1146 |
| OS Independent | 1018 |
| None | 494 |
| Windows | 493 |
| POSIX | 458 |
| Posix; MacOS X; Windows | 340 |
| UNIX | 238 |
| Aix | 152 |
| Platform Independent | 130 |
| MacOS X | 102 |
| POSIX,Windows | 71 |
| WIN32 | 61 |
| Cross Platform | 53 |
| GNU/Linux | 51 |
| Independent | 49 |
| Posix; MacOS X | 46 |
Licenses
| License | Count |
|---|---|
| MIT | 53876 |
| 44374 | |
| UNKNOWN | 17531 |
| BSD | 14056 |
| MIT LICENSE | 8221 |
| GPL | 5720 |
| AGPL-3 | 5680 |
| GPLV3 | 3902 |
| APACHE 2.0 | 3340 |
| APACHE LICENSE 2.0 | 2587 |
Size
What is currently the biggest Python package?
| Package | Release Number | Size |
|---|---|---|
| mxnet-cu90 | 1.6.0b20200104 | 600MB |
| mxnet-cu92 | 1.6.0b20200104 | 598MB |
| de422 | 2009.1 | 545MB |
| mxnet-cu101mkl | 1.6.0b20191125 | 531MB |
| mxnet-cu101 | 1.6.0b20191125 | 496MB |
| cntk-gpu | 2.7 | 493MB |
| mxnet-cu100mkl | 1.6.0b20191125 | 490MB |
| mxnet-cu100 | 1.6.0b20191125 | 454MB |
| mxnet-cu92mkl | 1.6.0b20191118 | 452MB |
| mxnet-cu90mkl | 1.6.0b20191115 | 427MB |
| mxnet-cu91mkl | 1.3.1b20180926 | 389MB |
| h2o4gpu | 0.3.2 | 366MB |
| tensorflow-gpu | 2.1.0rc2 | 357MB |
| tensorflow | 2.1.0rc2 | 356MB |
| mxnet-cu91 | 1.3.1b20180926 | 346MB |
It is strange, that the deep learning frameworks are way bigger than the rest. I've tried to find a Stackoverflow Question.
Package type
| License | Count |
|---|---|
| sdist | 180307 |
| bdist_wheel | 106515 |
| bdist_egg | 8150 |
| bdist_wininst | 1201 |
| bdist_dumb | 804 |
| bdist_rpm | 44 |
| bdist_msi | 33 |
| bdist_dmg | 4 |
Dependencies
