pip2pi builds a PyPI-compatible package repository from pip requirements

PyPI can go down, package maintainers can remove old tarballs, and downloading tarballs can take a long time. pip2pi helps to alleviate these problems by making it blindingly simple to maintain a PyPI-compatible repository of packages your software depends on.

Status

These tools were developed to be used internally, and they appear to work for me. A quick glance at the code will make it obvious that they are far from robust (ex, they probably won't work on Windows and they make a few calls to shell commands that could be implemented in Python)... But they should work, and they shouldn't eat your data or steal private keys or anything.

Requirements

0. pip
1. A requirements.txt file for your project (optional, but useful)
2. An HTTP server (optional, but useful)

Setup

Install pip2pi:

$ pip install pip2pi

And create the directory which will contain the tarballs of required packages, preferably somewhere under your web server's document root:

$ mkdir /var/www/packages/

Mirroring Packages

To mirror a package and all of its requirements, use pip2tgz:

$ pip2tgz packages/ foo==1.2
...
$ ls packages/
foo-1.2.tar.gz
bar-0.8.tar.gz

Note that pip2tgz passes package arguments directly to pip, so packages can be specified in any format that pip recognizes:

$ cat requirements.txt
foo==1.2
http://example.com/baz-0.3.tar.gz
$ pip2tgz packages/ -r requirements.txt bam-2.3/
...
$ ls packages/
foo-1.2.tar.gz
bar-0.8.tar.gz
baz-0.3.tar.gz
bam-2.3.tar.gz

Building a Package Index

A directory full of .tar.gz files can be turned into PyPI-compatible "simple" package index using the dir2pi command:

$ ls packages/
bar-0.8.tar.gz
baz-0.3.tar.gz
foo-1.2.tar.gz
$ dir2pi packages/
$ find packages/
packages/
packages/bar-0.8.tar.gz
packages/baz-0.3.tar.gz
packages/foo-1.2.tar.gz
packages/simple
packages/simple/bar
packages/simple/bar/bar-0.8.tar.gz
packages/simple/baz
packages/simple/baz/baz-0.3.tar.gz
packages/simple/foo
packages/simple/foo/foo-1.2.tar.gz

But that's a lot of work...

If running two commands seems like too much work... Take heart! The pip2pi command will run both of them for you... And it will use rsync to copy the new packages and index to a remote host!

$ pip2pi example.com:/var/www/packages/ foo==1.2
...
$ curl -I http://example.com/packages/simple/foo/foo-1.2.tar.gz | head -n1
HTTP/1.1 200 OK

But that's still too much work...

Take heart! Your shell's alias command can help. Add an alias like this to your shell's runtime configuration file (hint: ~/.bashrc or similar):

alias pip2acmeco="pip2pi dev.acmeco.com:/var/www/packages/"

Now updating your package index will be as simple as:

$ pip2acmeco foo==1.2 -r bar/requirements.txt

Using Your New Package Index

To use the new package index, pass the --index-url= argument to pip:

$ pip install --index-url=http://example.com/packages/simple/ foo

Or, once it has been mirrored, prefix you requirements.txt with --index-url=...:

$ cat requirements.txt
--index-url=http://example.com/packages/simple/
foo==1.2

Without a web server

You can use your package index offline, too:

$ pip install --index-url=file:///var/www/packages/simple foo==1.2

Some Tips

When installing packages from source via python setup.py install or python setup.py install, you may need to create a setup.cfg, which points to your package index. Here are some examples for an offline package index in your Windows, Linux, or Mac file system:

[easy_install]
# Windows
# index_url = file:///C:/pip2pi/simple/

# Linux
# index_url = file:///home/myusername/.pip2pi/simple/

# Mac
index_url = file:///Users/myusername/.pip2pi/simple/

Note the triple ///` after ``file: -- two for the protocol, the third for the root of the local file system.

Keywords

• Mirror PyPI
• Offline PyPI
• Create offline PyPI mirror