birkin

Introducing uv

Python is a friendly language, used by a wide-variety of folk in the Library -- from back-end developers to folk exploring digital-humanities text-mining and web-scraping to others exploring AI to others doing data-cleanup.

But too often there's friction in the process of getting code running because of the need to install:

• various versions of python
• a virtual-environment (venv)
• various packages.

This workshop -- no programming experience required -- will introduce the python package-manager "uv", which has become incredibly popular because it removes lots of that friction, allowing people to focus more on their work/exploration/research, rather than "setup". It also helps collaboration, by making it easier for others using uv to use your scripts.

birkin

# /// script
# requires-python = "==3.12.*"
# dependencies = [
#   "polars",
#   "httpx"
# ]
# ///

"""
Example of accessing a public google-sheet ("Anyone with the link can view")

birkin

# /// script
# requires-python = "==3.12.*"
# ///

"""
Generates a secure unique ID

Usage:
    uv run ./unique_id.py  # default length is 10
    uv run ./unique_id.py --length 20

birkin

# /// script
# requires-python = "==3.8.*"
# dependencies = ["lxml==4.9.1"]
# ///

"""
Script to test lxml installation on redhat-7 server.

Uses python PEP-723 inline-script-metadata to allow uv to run without venv installs.

birkin

exempt-paths refactor suggestion

idea

Based on thinking and research, below is a suggestion for handling exempt-paths in turnstile_middleware.py.

Some nice things about this...

• the regex-compilation happens just once, when the webapp first loads, improving speed.
• it allows us to change a .env setting easily, without deploying new code.
• it allows the pattern-match code in turnstile_middleware.py to be simple, becase regex is handling the "starts-with" or "anywhere-within" check.

birkin

# /// script
# requires-python = "==3.12.*"
# dependencies = ["pydantic==2.10.*"]
# ///

"""
Pydantic example
- Pydantic is a data validation package for Python. Using rust, it's fast.
- <https://docs.pydantic.dev/latest/>

birkin

#!/bin/bash

## Purpose:
## This script hacks Django's sqlite package to enable it to work with an old version
##   of sqlite3 on an old version of RedHat.
##
## Flow:
## It checks if django and pysqlite3 and pysqlite3-binary are installed in the virtual environment.
## If the required packages are not found, the script alerts the user and exits.
## If the packages are found, it proceeds to update Django's `sqlite3/base.py` file.

birkin

"""
Generates XML output using Django's template system.
Accepts command-line arguments to specify a person's name and role.

Usage:
    $ uv run "./render_xml_via_template.py" --person "Birkin" --role "cheerleader"

Output:
    <?xml version="1.0" encoding="UTF-8"?>
    <root>

birkin

My October-2024 question

(continuation from irrelevant code-thread)...

If I had a PDF that conformed to this standard, and wanted to indicate that in metadata, what would be the a common digital-repository metadata xml spec where I'd list this conformance? And please provide an example or two of what the entry might look like.

---

ChatGPT response...

birkin

import pikepdf

def edit_decodeparms(pdf_path, output_path):
    with pikepdf.open(pdf_path) as pdf:
        for page_num, page in enumerate(pdf.pages, start=1):
            resources = page.get('/Resources', {})
            xobjects = resources.get('/XObject', {})
            for xobj_name, xobj_ref in xobjects.items():
                xobj = xobj_ref  # Use the object directly
                filters = xobj.get('/Filter', [])