check-dates

#!/usr/bin/env bash

# Check that dates are in chronological order.
#
# Input:
# - a Markdown (*.md) file, that contains a chronology in list format
# - a line matches if it has these characters, in order:
#   - a hyphen (-)
#   - a space ( )
#   - some digits, possibly with a decimal point, e.g.: 13.8, 1984
#   - a suffix, e.g.: bya, mya, kya, BCE, CE
#   - a colon (:)
#
# Copyright 2020 Benjamin Thornton.  Released as open source software under the
# ISC License: https://github.com/smenjas/history-notes/blob/master/LICENSE

# Which file do we want to check?
target="README.md"

# Where shall we store our temporary files?
tmp_dir="tmp"
if [ ! -d "$tmp_dir" ]; then
    mkdir "$tmp_dir"
fi

# Extract a list of dates to analyze.
dates_path="$tmp_dir/dates.txt"
#if [ ! -e "$dates_path" ]; then # Uncomment to cache output.
    echo "Writing all dates to: $dates_path"
    # Search for the pattern described above, cut the date and suffix,
    # remove the trailing colon, and write the output to a file.
    egrep -o '^- [0-9\.,]+ [A-Za-z]+:' "$target" \
        | cut -d ' ' -f 2,3 \
        | sed 's/:$//' \
        > "$dates_path"
#fi

# Which suffixes are these dates using?
suffixes=$(cat "$dates_path" | cut -d ' ' -f 2 | uniq)
echo "Found these suffixes:"
echo "$suffixes"
echo

for suffix in $suffixes; do
    # Split up the list according to suffix.
    suffix_path="${tmp_dir}/${suffix}.txt"
    #if [ ! -e "$suffix_path" ]; then # Uncomment to cache output.
        echo "Copying dates ending in $suffix to: $suffix_path"
        grep -w "$suffix" "$dates_path" > "$suffix_path"
    #fi

    # Display the first and last date for this suffix, for context.
    first=$(head -n 1 "$suffix_path")
    last=$(tail -n 1 "$suffix_path")
    num_lines=$(cat "$suffix_path" | wc -l | xargs)
    echo "$suffix_path 1: $first"
    echo "$suffix_path $num_lines: $last"

    # Sort the list, and save both for comparison.
    sorted_path="${tmp_dir}/${suffix}-sorted.txt"
    #if [ ! -e "$sorted_path" ]; then # Uncomment to cache output.
        echo "Sorting dates ending in $suffix to: $sorted_path"
        first_num=$(echo "$first" | cut -d ' ' -f 1)
        last_num=$(echo "$last" | cut -d ' ' -f 1)
        if (( $(echo "$first_num < $last_num" | bc -l) )); then
            echo "$first_num < $last_num"
            reverse=""
        else
            echo "$first_num > $last_num: Sorting in reverse"
            reverse="r"
        fi
        sort -n$reverse "$suffix_path" > "$sorted_path"
    #fi

    # Compare the original list with the sorted list.
    echo -n "diff $suffix_path $sorted_path | wc -l: "
    diff "$suffix_path" "$sorted_path" | wc -l | xargs
    diff "$suffix_path" "$sorted_path" | head
    echo
done