From d4e56cf557234d5df0a596fe74b10cf0b23382f4 Mon Sep 17 00:00:00 2001
From: Tom Ice <contact@thomasice.com>
Date: Tue, 1 Oct 2024 20:11:26 -0400
Subject: [PATCH] Adjust commitsByYear() to support multiple dates

* commitsByYear() previously could only support Default-style
  dates for git log. However, a user can use a variety of different
  date formats such as relative, default, iso, rfc, short, and raw.

  This change modifies commitsByYear() to handle multiple different
  date strings

Fixes #172
---
 git-quick-stats | 97 ++++++++++++++++++++++++++++++++++++++-----------
 1 file changed, 75 insertions(+), 22 deletions(-)

diff --git a/git-quick-stats b/git-quick-stats
index 139b16f..342a067 100755
--- a/git-quick-stats
+++ b/git-quick-stats
@@ -17,7 +17,7 @@ _since=${_GIT_SINCE:-}
 if [[ -n "${_since}" ]]; then
     _since="--since=$_since"
 else
-    _since="--since=$(git log --reverse --format='%ad' | head -n1)"
+    _since="--since=$(git log --reverse --format='%ad' --date=iso | head -n1)"
 fi
 
 # End of git log date. Respects all git datetime formats
@@ -761,6 +761,47 @@ function commitsPerDay() {
         --date=short --format='%ad' $_log_options $_pathspec | sort | uniq -c
 }
 
+################################################################################
+# DESC: Convert a timestamp to a date string to handle git's date formats
+# ARGS: $1: Timestamp
+# OUTS: Echoes a four-digit year
+################################################################################
+function parse_year() {
+    local date_str="$1"
+    local year
+    local timestamp
+    local default_git_date_regex
+
+    # Handle the raw UNIX timestamp format i.e. 1697375696 +0000
+    if [[ "$date_str" =~ ^[0-9]+(\ [+-][0-9]{4})?$ ]]; then
+        timestamp=$(echo "$date_str" | awk '{print $1}')
+        year=$(date -d "@$timestamp" '+%Y' 2>/dev/null)
+    else
+        # Default case can get funky. We need to create a clever regex to
+        # handle the default case which is like Mon Oct 15 12:34:56 2023 +0000
+        # Let's make this explicit for future devs to follow along.
+        default_git_date_regex='^' # Start from the beginning of the string
+        default_git_date_regex+='[A-Za-z]{3}\ ' # Day abbrev
+        default_git_date_regex+='[A-Za-z]{3}\ ' # Month abbrev
+        default_git_date_regex+='[0-9]{1,2}\ ' # Day of the month
+        default_git_date_regex+='[0-9]{2}:[0-9]{2}:[0-9]{2}\ ' # Time HH:MM:SS
+        default_git_date_regex+='[0-9]{4}\ ' # Year
+        default_git_date_regex+='[+-][0-9]{4}$' # Timezone offset
+
+        if [[ "$date_str" =~ $default_git_date_regex ]]; then
+            # Move the year before the time to match a format that Date can parse
+            date_str=$(echo "$date_str" | awk '{print $1, $2, $3, $5, $4, $6}')
+        elif [[ "$date_str" =~ ^[0-9]{1,2}/[0-9]{1,2}/[0-9]{2,4} ]]; then
+            # Handle DD/MM/YYYY format
+            date_str=$(echo "$date_str" | awk -F'/' '{print $2"/"$1"/"$3}')
+        fi
+        # Extract the final date
+        year=$(date -d "$date_str" '+%Y' 2>/dev/null)
+    fi
+
+    echo "$year"
+}
+
 ################################################################################
 # DESC: Displays a horizontal bar graph based on total commits per year
 # ARGS: None
@@ -769,30 +810,38 @@ function commitsPerDay() {
 function commitsByYear() {
     optionPicked "Git commits by year:"
     local year startYear endYear __since __until
-    startYear=$(echo "$_since" | sed -E 's/^.* ([0-9]{4})( .*)?$/\1/')
-    endYear=$(echo "$_until" | sed -E 's/^.* ([0-9]{4})( .*)?$/\1/')
 
+    # Extract the date strings from $_since and $_until
+    since_date="${_since#--since=}"
+    until_date="${_until#--until=}"
+
+    # Grab the four digit year from $_since and $_until
+    startYear=$(parse_year "$since_date")
+    endYear=$(parse_year "$until_date")
+    
     echo -e "\tyear\tsum"
-    for year in $(seq "$startYear" "$endYear")
-    do
-        if [ "$year" = "$startYear" ]
-        then
-          __since=$_since
-          __until="--until=$year-12-31"
-        elif [ "$year" = "$endYear" ]
-        then
-          __since="--since=$year-01-01"
-          __until=$_until
+    # Add time strings to make these a touch more robust
+    for year in $(seq "$startYear" "$endYear"); do
+        if [[ "$year" = "$startYear" ]]; then
+            __since=$_since
+            __until="--until=$year-12-31 23:59:59"
+        elif [[ "$year" = "$endYear" ]]; then
+            __since="--since=$year-01-01 00:00:00"
+            __until=$_until
         else
-          __since="--since=$year-01-01"
-          __until="--until=$year-12-31"
+            __since="--since=$year-01-01 00:00:00"
+            __until="--until=$year-12-31 23:59:59"
         fi
-
-        echo -en "\t$year\t"
-        git -c log.showSignature=false shortlog -n $_merges --format='%ad %s' \
-            "$__since" "$__until" $_log_options | grep -cE \
-              " \w\w\w [0-9]{1,2} [0-9][0-9]:[0-9][0-9]:[0-9][0-9] $year " \
-                || continue
+        
+        # Count commits directly using git rev-list instead of git log
+        commit_count=$(
+            git rev-list --count $_merges \
+            "$__since" "$__until" HEAD $_log_options
+        )
+        echo -e "\t$year\t$commit_count"
+        # TODO: The bar graph can get funky when there are only a handful of
+        # commits. We can set a max length to try to fix this, but this is a
+        # bit of a problem across all the bar graphs.
     done | awk '{
         count[$1] = $2
         total += $2
@@ -986,7 +1035,11 @@ function suggestReviewers() {
 checkUtils
 
 # Check if we are currently in a git repo.
-git rev-parse --is-inside-work-tree > /dev/null
+if ! git rev-parse --is-inside-work-tree > /dev/null; then
+    echo "ERROR: You need to be inside a git repo to parse stats!"
+    usage
+    exit 1
+fi
 
 # Parse non-interative commands
 if [[ "$#" -eq 1 ]]; then