Merge pull request #195 from oss-slu/feature/issue_186_remove_inactive_users

Feature/issue 186 remove inactive users

Author: hcaballero2

Backend/dataCollection/formatJSON.py

@@ -1,6 +1,7 @@
 
 import numpy as np
 import pandas as pd
+from Backend.streakCalculation import calculate_current_streak
 
 
 def format_json_data(raw_data, sprint = -1):
@@ -44,11 +45,17 @@ def format_json_data(raw_data, sprint = -1):
             avg_time_to_close = group['lead_time'].mean()
             total_issues_opened = len(group)
             total_issues_closed = group['state'].value_counts().get('closed', 0)
+            if 'closed_at' in group.columns:
+                closed_issue_dates = group[group['state'] == 'closed']['closed_at'].dropna().tolist()
+            else:
+                closed_issue_dates = []
+            current_streak = calculate_current_streak(closed_issue_dates)
 
             formatted_data['issues'][user] = {
                 "average_time_to_close": avg_time_to_close if not np.isnan(avg_time_to_close) else 0,
                 "total_issues_opened": total_issues_opened,
-                "total_issues_closed": total_issues_closed
+                "total_issues_closed": total_issues_closed,
+                "currentStreak": current_streak
             }
 
     #Process Pull Requests

Backend/streakCalculation.py

@@ -0,0 +1,87 @@
+from datetime import datetime, timedelta
+
+ISSUE_THRESHOLD = 3
+
+def get_week_start(date_obj):
+    """
+    Returns the Monday of the week for a given data
+    """
+    return (date_obj - timedelta(days=date_obj.weekday())).date()
+
+def parse_closed_dates(closed_issue_dates):
+    """
+    Converts a list of date strings or datetime objects into datetime objects
+    Skips invalid or empty values
+    """
+    parsed_dates = []
+
+    for value in closed_issue_dates:
+        if not value:
+            continue
+
+        if isinstance(value, datetime):
+            parsed_dates.append(value)
+            continue
+
+        try:
+            parsed_dates.append(datetime.fromisoformat(value.replace("Z", "+00:00")))
+        except (ValueError, AttributeError):
+            continue
+
+    return parsed_dates
+
+def group_issues_by_week(closed_issue_dates):
+    """
+    Counts how many issues were closed in each week
+    Returns a dictionary like:
+    {
+        week_start_date: count
+    }
+    """
+    weekly_counts = {}
+
+    parsed_dates = parse_closed_dates(closed_issue_dates)
+
+    for closed_date in parsed_dates:
+        week_start = get_week_start(closed_date)
+        weekly_counts[week_start] = weekly_counts.get(week_start, 0) + 1
+
+    return weekly_counts
+
+def calculate_current_streak(closed_issue_dates, threshold=ISSUE_THRESHOLD, reference_date=None):
+    """
+    Calculates the active streak based on weekly issue closure threshold
+
+    A streak increases by 1 for every consecutive week where
+    the number of closed issues meets or exceeds the threshold
+
+    Args:
+        closed_issue_dates: list of datetime objects or ISO date strings
+        threshold: minimum number of issues closed in a week
+        reference_date: optional datetime to use instead of current time
+
+    Returns:
+        int: current streak count
+    """
+    if reference_date is None:
+        reference_date = datetime.now()
+
+    weekly_counts = group_issues_by_week(closed_issue_dates)
+
+    if not weekly_counts:
+        return 0
+    
+    current_week_start = get_week_start(reference_date)
+    streak = 0
+    
+    while weekly_counts.get(current_week_start, 0) >= threshold:
+        streak += 1
+        current_week_start -= timedelta(days=7)
+
+    return streak
+
+def count_total_closed_issues(closed_issue_dates):
+    """
+    Returns the total number of valid closed issue dates
+    """
+    return len(parse_closed_dates(closed_issue_dates))

Backend/test_streakCalculation.py

@@ -0,0 +1,72 @@
+from datetime import datetime
+
+from Backend.streakCalculation import (
+    calculate_current_streak,
+    group_issues_by_week,
+    count_total_closed_issues
+)
+
+def test_calculate_current_streak_consecutive_weeks():
+    closed_issue_dates = [
+        "2026-04-01T10:00:00",
+        "2026-04-02T11:00:00",
+        "2026-04-03T12:00:00",
+        "2026-03-24T10:00:00",
+        "2026-03-25T11:00:00",
+        "2026-03-26T12:00:00",
+    ]
+
+    reference_date = datetime(2026, 4, 3)
+    streak = calculate_current_streak(closed_issue_dates, threshold=3, reference_date=reference_date)
+
+    assert streak == 2
+
+def test_calculate_current_streak_resets_when_week_missed():
+    closed_issue_dates = [
+        "2026-04-01T10:00:00",
+        "2026-04-02T11:00:00",
+        "2026-04-03T12:00:00",
+        "2026-03-10T10:00:00",
+        "2026-03-11T11:00:00",
+        "2026-03-12T12:00:00",
+    ]
+
+    reference_date = datetime(2026, 4, 3)
+    streak = calculate_current_streak(closed_issue_dates, threshold=3, reference_date=reference_date)
+
+    assert streak == 1
+
+def test_calculate_current_streak_returns_zero_for_no_data():
+    closed_issue_dates = []
+
+    reference_date = datetime(2026, 4, 3)
+    streak = calculate_current_streak(closed_issue_dates, threshold=3, reference_date=reference_date)
+
+    assert streak == 0
+
+
+def test_group_issues_by_week_counts_correctly():
+    closed_issue_dates = [
+        "2026-04-01T10:00:00",
+        "2026-04-02T11:00:00",
+        "2026-04-08T12:00:00",
+    ]
+
+    weekly_counts = group_issues_by_week(closed_issue_dates)
+
+    assert len(weekly_counts) == 2
+    assert sum(weekly_counts.values()) == 3
+
+
+def test_count_total_closed_issues_counts_valid_dates():
+    closed_issue_dates = [
+        "2026-04-01T10:00:00",
+        "2026-04-02T11:00:00",
+        None,
+        "",
+        "invalid-date"
+    ]
+
+    total_closed = count_total_closed_issues(closed_issue_dates)
+
+    assert total_closed == 2

Frontend/src/components/TopContributorsRepos.jsx

@@ -17,17 +17,28 @@ const TOP_N = 5;
 
 const TopContributorsRepos = () => {
   // Converting repo JSON structure into event-like objects
-  const {topContributors, topRepos} = getTopContributorsAndRepos(lifetimeData, TOP_N);
+  const selectedRepo = "core_desk";
+  const {topContributors, topRepos} = getTopContributorsAndRepos(lifetimeData, TOP_N, selectedRepo);
 
   return (
     <div style={{ display: "flex", gap: "30px" }}>
       {/* Render list of top contributors by activity count */}
       <div>
-        <h3>Top Contributors</h3>
+        <h3>
+          Top Contributors{" "}
+          <span
+            tabIndex="0"
+            aria-label="Leaderboard streak info"
+            title="A streak increases for every consecutive 7-day period where a user closes at least 3 issues. Streak resets if threshold is not met."
+            style={{ cursor: "pointer", border: "1px solid black", borderRadius: "50%", padding: "2px 6px", fontSize: "12px" }}
+          >
+            i
+          </span>
+        </h3>
         <ul>
           {topContributors.map((user) => (
             <li key={user.name}>
-              {user.name} ({user.count})
+              {user.name} 🔥 {user.currentStreak} {user.currentStreak === 1 ? "week" : "weeks"} streak ({user.count})
             </li>
           ))}
         </ul>

Frontend/src/components/charts/__tests__/test_getTopContributorsAndRepos.test.js

@@ -24,11 +24,13 @@ const mockJSON = {
   it("counts contributor activity and sorts correctly", () => {
     const {topContributors} = getTopContributorsAndRepos(mockJSON, 5);
     // charlie: 4, alicce: 2, bob: 1
-    expect(topContributors).toEqual([
-      { name: "charlie", count: 4 },
-      { name: "alice", count: 2 },
-      { name: "bob", count: 1 },
-    ]);
+    expect(topContributors).toEqual(
+      expect.arrayContaining([
+        expect.objectContaining({ name: "charlie", count: 4 }),
+        expect.objectContaining({ name: "alice", count: 2 }),
+        expect.objectContaining({ name: "bob", count: 1 }),
+      ])
+    );
   });
 
   it("counts repository activity and sorts correctly", () => {