Python Trainer

What Python is, where it belongs in an automation stack, and how it actually runs. Installation and version management, the interpreter and REPL, module search paths, and virtual environments — because the first real Python problem most teams hit is dependency isolation, not syntax.

Topics: What Python is and why it dominates infrastructure tooling · Choosing an interpreter version and managing several with pyenv · The interpreter, the REPL and script execution · Python syntax and semantics — indentation, truthiness, naming · Virtual environments and dependency isolation · sys.path, imports and how modules are found

• Assignments: (1) Install and pin a Python version, then create an isolated environment for the course project; (2) Diagnose three deliberately broken import failures
• Capstone: Produce a reproducible development environment definition your team could adopt unchanged

The core of the language, taught against operational examples rather than exercises. Variables and data types, operators and expressions, conditional statements and the loop constructs, plus the mutability rules that cause most surprising behaviour in shared automation code.

Topics: Variables and data types · Operators and expressions · Conditional statements · Loops — for, while, else, break and continue · Mutability, identity and copying · Truthiness and the comparison traps

• Assignments: (1) Rewrite a 60-line Bash script as readable Python; (2) Fix a script that mutates a shared default argument
• Capstone: Build a host-inventory report that reads raw data and prints a formatted summary

How Python code is organised once it stops being a single file. Function definition, arguments and defaults, scope and closures, then modules, packages, __init__ and the layout that keeps a growing internal tool navigable.

Topics: Functions, arguments, defaults and keyword-only parameters · Return values, unpacking and multiple returns · Scope, closures and the LEGB rule · Modules and packages · Entry points and __main__ · Type hints and what they buy in shared code

• Assignments: (1) Refactor a flat script into a package with a clear entry point; (2) Add type hints to an existing module and run a type checker over it
• Capstone: Convert a single-file automation script into an installable package

Classes as they are actually used in operational code — modelling a resource, a client, a configuration object — rather than as an academic exercise. Attributes, methods, inheritance, composition, dunder methods and dataclasses, with an honest look at when a function is the better answer.

Topics: Classes, instances and attributes · Methods, class methods and static methods · Inheritance versus composition · Dunder methods — repr, eq, context managers · dataclasses and named tuples · When not to use a class

• Assignments: (1) Model a cloud resource as a class with a working __repr__ and equality; (2) Write a context manager that guarantees cleanup on failure
• Capstone: Design a small client library wrapping an internal API, with tested error types

The structures every automation script leans on, and the numerical and tabular libraries used for capacity, cost and reliability reporting. Lists, tuples, dictionaries and sets with their performance characteristics, then NumPy arrays and Pandas DataFrames for cleaning, preparing and visualising operational data.

Topics: Lists and tuples · Dictionaries and dictionary methods · Sets and set algebra for diffing state · Comprehensions and generators · NumPy arrays · Pandas DataFrames · Data cleaning, preparation and visualisation

• Assignments: (1) Diff two inventories using set operations and report additions, removals and changes; (2) Load a month of billing data into a DataFrame and produce a per-service cost summary
• Capstone: Build a reporting script that turns raw exported data into a chart and a written summary

Everything an automation tool does to disk. Reading and writing files with correct encoding and buffering, walking and manipulating directories with pathlib, atomic writes and temporary files, and serialising state to JSON, CSV, YAML and pickle — including when each of those is the wrong choice.

Topics: Reading and writing files · Working with directories and pathlib · Encodings, newlines and binary mode · Serialization and deserialization — JSON, CSV, YAML, pickle · Atomic writes and temporary files · File permissions and ownership from Python

• Assignments: (1) Write a config loader that merges defaults, a file and environment variables; (2) Implement an atomic write that survives being killed mid-run
• Capstone: Build a state file mechanism an automation tool can safely resume from

Turning unstructured text into structured records — the single most common Python task in an operations team. Pattern syntax, metacharacters, groups, compilation and the difference between match, search and findall, applied to real log formats.

Topics: What regular expressions are and where they belong · Using regular expressions in Python — the re module · Patterns, metacharacters and character classes · Groups, named groups and backreferences · Compiling patterns and matching efficiently · Greedy versus lazy matching and catastrophic backtracking

• Assignments: (1) Parse a web server access log into structured records and aggregate by status code; (2) Rewrite a pattern that backtracks catastrophically on long input
• Capstone: Build a log analyser that extracts, aggregates and alerts on a real log format

Where Python meets the operating system. Running external commands safely, streaming and capturing output, handling exit codes and timeouts, responding to signals, and wrapping the result in a command-line interface with proper logging — plus a short look at desktop GUI options for the rare case a tool needs one.

Topics: subprocess — run, Popen, argument lists and why shell=True is dangerous · Capturing, streaming and interleaving stdout and stderr · Exit codes, timeouts and killing child processes · Signals, graceful shutdown and cleanup · Building a CLI with argparse — subcommands, defaults and help · Structured logging and verbosity levels · GUI programming in Python — Tkinter and PyQt, and when a desktop tool is justified

• Assignments: (1) Wrap a long-running shell command with a timeout, streamed output and a correct exit code; (2) Add subcommands, --dry-run and log levels to an existing script
• Capstone: Ship an operations CLI with subcommands, dry-run mode, logging and a non-zero exit path

Talking to the rest of the estate. HTTP requests with authentication, pagination and retry semantics, then the socket layer underneath for the cases where HTTP is not the protocol — health checkers, port scanners and custom agents.

Topics: What networking looks like from an automation script · Creating and sending HTTP requests in Python · Authentication, headers, sessions and connection reuse · Pagination, rate limiting, timeouts and retry with backoff · Creating and managing TCP and UDP sockets · Popular Python networking libraries — requests, httpx, urllib3, paramiko

• Assignments: (1) Write a paginated API client that retries on 429 and 5xx with exponential backoff; (2) Build a TCP health checker that reports latency across a host list
• Capstone: Build a resilient client for an internal API, with tests that simulate failures

When automation needs to remember something. The DB-API, connecting and interacting with relational databases, parameterised queries and transactions, connection pooling, and the SQL injection failure mode that appears the moment someone builds a query with string formatting.

Topics: What database programming is and when automation needs it · Connecting to and interacting with databases from Python · Popular Python database libraries — sqlite3, psycopg, PyMySQL, SQLAlchemy · SQL queries, parameterisation and injection · Transactions, commits and rollback · Connection pooling and long-running jobs

• Assignments: (1) Store and query run history in SQLite from an automation tool; (2) Fix a query built by string concatenation and prove the injection is closed
• Capstone: Add durable, queryable run history to the CLI you built earlier

Exposing automation to other teams. What web development means in a platform context — a small internal service, not a product — using Flask or FastAPI, then deployment behind a real server, and the application security issues that matter when the endpoint can change infrastructure.

Topics: What web development means for a platform team · Creating and deploying web applications with Flask and FastAPI · Routing, request validation and JSON responses · Popular Python web frameworks and how they differ · Serving with WSGI/ASGI behind a real server · Web application security — authentication, authorisation, input validation, secrets

• Assignments: (1) Expose an existing automation script as an authenticated HTTP endpoint; (2) Add request validation and rate limiting to that endpoint
• Capstone: Deliver a small internal self-service API that another team could call safely

The delivery end of the course. Driving cloud control planes with the vendor SDKs, deploying Python to serverless runtimes, using Python for CI/CD and infrastructure as code, testing with pytest, and packaging so the tool installs cleanly for the next engineer.

Topics: What cloud computing changes for automation code · Deploying Python applications to the cloud · Cloud SDKs — boto3, azure-sdk, google-cloud, paginators and credential chains · Serverless computing — Lambda and Cloud Functions, cold starts and packaging · Python for DevOps tasks and CI/CD pipeline steps · Infrastructure as code from Python — CDK and Pulumi · Testing with pytest — fixtures, parametrisation, mocking cloud calls · Packaging, dependency pinning and internal distribution · Where machine learning fits in operations tooling

• Assignments: (1) Write a boto3 script that tags and reports on untagged resources across an account; (2) Package a tool and install it from an internal index in a pipeline
• Capstone: Ship a tested, packaged Python tool that runs on a schedule in CI against a real cloud account