Replacing Unstable Shell RPM Checks with Python’s rpm‑python Library

1. The Problem with Shell Scripts

A colleague’s Shell script that compared local RPM package versions with a central repository proved unstable and hard to maintain because RPM version strings have many formats, making simple string parsing impossible.

To solve this, the author searched for a Python interface that could directly access RPM metadata on Linux machines. yum install -y rpm-python After installing the rpm-python package (located in /usr/lib64/python2.7/site-packages/), the library can be imported in Python scripts.

>>> import rpm
>>> rpm.__path__
['/usr/lib64/python2.7/site-packages/rpm']

2. Discovering a Helpful Resource

Searching online revealed a website that aggregates real‑world code examples for Python libraries. By entering a library name, the site returns usage snippets extracted from open‑source projects, along with the number of projects that use the library.

The site (https://www.programcreek.com/python/) lists roughly 2,000 libraries and provides concrete examples that are more practical than the official Python documentation.

3. Using the rpm Library

Getting Information About an Uninstalled RPM Package

>> import rpm
>>> ts = rpm.TransactionSet()
>>> rpmhdr = ts.hdrFromFdno("/root/librbd1-devel-10.2.10-0.el7.centos.x86_64.rpm")
>>> rpmhdr["NAME"]
'librbd1-devel'
>>> rpmhdr["VERSION"]
'10.2.10'
>>> rpmhdr["RELEASE"]
'0.el7.centos'
>>> rpmhdr["ARCH"]
'x86_64'

Getting Information About an Installed RPM Package

>> import rpm
>>> ts = rpm.TransactionSet()
>>> query = ts.dbMatch("name", "librbd1")
>>> query.count()
1
>>> pkg_info = next(query)
>>> pkg_info["NAME"]
'librbd1'
>>> pkg_info["VERSION"]
'12.2.9.1.1'
>>> pkg_info["RELEASE"]
'0.el7.centos'
>>> pkg_info["ARCH"]
'x86_64'

These examples were extracted from the aforementioned website, which curates code snippets from real projects rather than the standard library documentation.

4. Key Takeaways

Python’s rpm-python library provides a reliable way to query both installed and offline RPM packages, avoiding the brittleness of Shell parsing.

The programcreek.com site is a valuable reference for discovering practical usage patterns of many Python libraries.