codespellit

#!/bin/bash
#emacs: -*- mode: shell-script; c-basic-offset: 4; tab-width: 4; indent-tabs-mode: nil -*-
#ex: set sts=4 ts=4 sw=4 et:

set -eu
source "$(readlink -f "$0" | xargs dirname)/common.sh"

codespell_version_full=$(codespell --version)
codespell_version=${codespell_version_full//.dev$/}

if [[ "$codespell_version_full" =~ \.dev ]]; then
	# for the action we
	codespell_dir=$(python -c 'import codespell_lib as l; print(l.__path__[0])')
	codespell_version=$(
		git -C "$codespell_dir" describe --tags --match 'v[0-9]*' | sed -e 's,-.*,,g' -e 's,^v,,g'
	)
	echo "WARNING! Running development version $codespell_version_full."
	echo "         For the action etc we will mention tagged version $codespell_version".
fi

function deduce-config() {
    # deduce where we have that config
    if [ -e ".codespellrc" ]; then
        echo .codespellrc
    elif grep -q '^\[tool.codespell\]' pyproject.toml 2>|/dev/null; then
        echo pyproject.toml
    elif grep -q '^\[codespell\]' setup.cfg 2>|/dev/null; then
        echo setup.cfg
    else
        echo "ERROR: Cannot figure out where codespell config is" >&2
    fi
}

function add-precommit() {
    if [ "$#" -ge 1 ]; then
        config="$1"
    else
        config=$(deduce-config)
    fi
    if [ -e .pre-commit-config.yaml ]; then
        # add pre-commit configuration there
        cat >> .pre-commit-config.yaml <<EOF

- repo: https://github.com/codespell-project/codespell
  # Configuration for codespell is in $config
  rev: v${codespell_version}
  hooks:
  - id: codespell
EOF

        if [ "$config" = "pyproject.toml" ]; then
            cat >> .pre-commit-config.yaml <<EOF
    additional_dependencies:
    - tomli
EOF
        fi
        git add .pre-commit-config.yaml
        git commit -m 'Add pre-commit definition for codespell'

        if ! python -c 'import yaml; yaml.safe_load(open(".pre-commit-config.yaml"))'; then
            echo "ERROR: added pre-commit definition ruined the yaml format, fix and rerun '!0 list'"
            exit 1
        fi
    else
        echo "No .pre-commit-config.yaml was found, nothing changed"
    fi
}

function list() {
    codespell "$@" || {
        echo
        echo "Problematic?:"
        echo
        codespell "$@" | grep ','
    }
}

function list-hits-sorted() {
    codespell "$@" | grep -e '==>' | sed -e 's,.*: *,,g' | sort | uniq -c | sort -n
}

# given filtered output from list-hits-sorted, apply those which were left
function fix-sorted() {  # shellcheck disable=SC2120
    if ! git diff-index --quiet HEAD --; then
        echo "Repository is dirty. Please commit your changes first."
        exit 1
    fi

    nfixes=$(cat "$@" | grep -e '==>' -c )
    cat "$@" | grep -e '==>' | while read -r _ fr _ to; do
        git grep -l "\<$fr\>" | tr '\n' '\000' | xargs -0 sed -i -e "s,\<$fr\>,$to,g"
    done

    git commit -m "Fix $nfixes typos found by codespell

Here are typos with their counts and introduced fix:

$(grep -e '==>' "$@")
" -a
}

# poor man shortcuts for just doing it using the regex etc
cmd="${1:-}"
case "$cmd" in
    list|list-hits-sorted|fix-sorted|deduce-config|add-precommit) shift; "$cmd" "$@"; exit $?;;
esac

git checkout -b enh-codespell

skips=".git*"
for f in '*.pdf' '*.svg' go.sum venv .venv venvs locale .tox versioneer.py package-lock.json vendor i18n '*-lock.yaml' '*.lock' '*.css' '*.min.*' '*-min.*' '*.pack.js' '*.niml' '*.gii'; do
    if find . -iname "$f" -print -quit | grep -q .; then
        skips+=",$f"
    fi
done

ignore_regex=''
if find . -iname '*.ipynb' -print -quit | grep -q .; then
  ignore_regex='^\s*"image/\S+": ".*'
fi

ignore_regex_prefix=''
if [ -z "$ignore_regex" ]; then
    ignore_regex_prefix="# "
fi

if [ -e pyproject.toml ]; then
	config="pyproject.toml"
	cat >> pyproject.toml <<EOF

[tool.codespell]
# Ref: https://github.com/codespell-project/codespell#using-a-config-file
skip = '$skips'
check-hidden = true
${ignore_regex_prefix}ignore-regex = '$ignore_regex'
# ignore-words-list = ''
EOF
elif [ -e setup.cfg ]; then
    config="setup.cfg"
    # Then config file as the "last" commit we can amend if needed
    cat >> setup.cfg <<EOF

[codespell]
# Ref: https://github.com/codespell-project/codespell#using-a-config-file
skip = $skips
check-hidden = true
${ignore_regex_prefix}ignore-regex = $ignore_regex
# ignore-words-list =
EOF
else
    config=".codespellrc"
    skips+=",$config"  # for check-hidden
    # Then config file as the "last" commit we can amend if needed
    cat > .codespellrc <<EOF
[codespell]
# Ref: https://github.com/codespell-project/codespell#using-a-config-file
skip = $skips
check-hidden = true
${ignore_regex_prefix}ignore-regex = $ignore_regex
# ignore-words-list =
EOF
fi

if [ -e .github ]; then
    mkdir -p .github/workflows
    cat > .github/workflows/codespell.yml << EOF
# Codespell configuration is within $config
---
name: Codespell

on:
  push:
    branches: [$branch]
  pull_request:
    branches: [$branch]

permissions:
  contents: read

jobs:
  codespell:
    name: Check for spelling errors
    runs-on: ubuntu-latest

    steps:
      - name: Checkout
        uses: actions/checkout@v4
      - name: Annotate locations with typos
        uses: codespell-project/codespell-problem-matcher@v1
      - name: Codespell
        uses: codespell-project/actions-codespell@v2
EOF
    git add .github/workflows/codespell.yml
    git commit -m "Add github action to codespell $branch on push and PRs"
else
    echo "No .github directory, no workflow provided"
fi


echo "Config in $config"
git add "$config"
# --no-verify since we might have trailing spaces which could trip some linters
git commit -m 'Add rudimentary codespell config' --no-verify "$config"

add-precommit "$config"
list "$@"