modify-csv

Apply a user-provided Python function to every cell in one CSV data row, every cell in one column, or one cell. The CSV's first record is treated as its header and is never modified. Row and column indexes are zero-based; columns can also be addressed by their exact header text.

Install

Install a published release from PyPI:

python -m pip install modify-csv

Or install a tagged version directly from the private GitHub repository using SSH authentication:

python -m pip install "git+ssh://git@github.com/martinjose/modify-csv.git@v0.1.0"

Git and access to the private repository are required for the second form. Each GitHub Release also provides a universal wheel and source archive.

For local development, install an editable copy:

python -m pip install -e .

This installs the modify-csv command. You can also run the package without an installation using PYTHONPATH=src python -m modify_csv.

Script contract

The supplied script must define exactly this callable shape:

from typing import Any


def eval(content: Any, row: int, col: int) -> Any:
    return content

Any is provided by Python's standard-library typing module. The function name, parameter names, and three-argument shape are enforced.

content is the current CSV field (initially a string), while row and col are zero-based numeric indexes into the data records and columns. The returned value is passed to Python's CSV writer. The script is ordinary trusted Python and executes with the same permissions as the command.

Examples

Apply a script file to a column selected by header:

modify-csv apply input.csv output.csv --script-file normalize.py --column email

Apply inline code to data row 2:

modify-csv apply input.csv output.csv \
  --script 'def eval(content, row, col): return content.strip()' \
  --row 2

Apply inline code to the cell at data row 0 and column 1:

modify-csv apply input.csv output.csv \
  --script 'def eval(content, row, col): return int(content) * 2' \
  --cell 0 1

Exactly one of --script or --script-file, and exactly one of --row, --column, or --cell, is required. Existing output files are protected; pass --force to replace one.

Generate a ready-to-edit transformation script:

modify-csv template transform.py

Use --force to replace an existing template. Run modify-csv --help or the help for either subcommand to see the required function signature. The original form without the apply subcommand remains supported for compatibility.

Releasing

Versions follow YYYY.M.RELEASE, where RELEASE starts at 1 each month. For example, the first release in October 2026 is 2026.10.1.

Preview the next version without changing the repository:

python scripts/release.py next

To update pyproject.toml, create the release commit, and create an annotated tag locally for inspection:

python scripts/release.py prepare

The command prints the exact push command to use after reviewing the commit and tag. To prepare and push both atomically in one step instead, run:

python scripts/release.py release

Release preparation requires a clean main branch. The package version in pyproject.toml remains the release source of truth. Pushing the generated tag starts the GitHub Actions release process:

git push --atomic origin main v2026.10.1

The tag-triggered release workflow tests Python 3.10 and 3.13 on Linux, macOS, and Windows. It then builds and validates a platform-independent wheel and source archive, publishes them to TestPyPI through trusted publishing, verifies that the package installs and runs from TestPyPI, and stages the exact artifacts in a draft GitHub Release.

After reviewing the successful run, open Actions, select Release, choose Run workflow, and enter the tag. This explicit manual action verifies and publishes the staged artifacts to PyPI, then publishes the draft GitHub Release. A mismatched version and tag, missing draft release, or changed artifact causes the production run to fail.

Trusted publishers must be configured independently on TestPyPI and PyPI. Both use owner martinjose, repository modify-csv, and workflow release.yml. The TestPyPI publisher uses environment testpypi; the production publisher uses environment pypi. The manually dispatched production run serves as the approval gate for private repositories without environment reviewer support.