From 2577d105f50bcefca09c49dd302503a29717275b Mon Sep 17 00:00:00 2001
From: Peter Howell <peter.howell@gmail.com>
Date: Thu, 6 Nov 2025 19:14:00 +0000
Subject: [PATCH] k

---
 users.py | 121 +++++++++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 121 insertions(+)

diff --git a/users.py b/users.py
index 9d39189..80d7b46 100644
--- a/users.py
+++ b/users.py
@@ -2813,6 +2813,126 @@ def summarize_student_logs(id=0):
     print(f"Done. Output written to cache/users/logs/{id}_summary.json")
 
 
+def readable_user_summary():
+    import re
+    import ast
+    import math
+    from datetime import timedelta
+    from zoneinfo import ZoneInfo
+
+    import pandas as pd
+
+
+    COURSE_RE = re.compile(r"/courses/(\d+)(?:/|$)")
+
+    def extract_course_id(url, links_str) :
+        """Return Canvas course id from URL (/courses/<id>) or from links.context."""
+        if url:
+            m = COURSE_RE.search(url)
+            if m:
+                return int(m.group(1))
+        if links_str:
+            try:
+                d = ast.literal_eval(links_str)
+                ctx = d.get("context")
+                if isinstance(ctx, int):
+                    return ctx
+            except Exception:
+                pass
+        return None
+
+
+    id = input("user id? ")
+    csv_path = f"cache/users/logs/{id}.csv"
+    gap_minutes = 10
+    tz_name = "America/Los_Angeles"
+    # Load
+    df = pd.read_csv(csv_path)
+
+    # Extract course_id
+    df["course_id"] = [
+        extract_course_id(u, l) for u, l in zip(df.get("url"), df.get("links"))
+    ]
+
+    # Keep only rows with a course_id
+    df = df.dropna(subset=["course_id"]).copy()
+    df["course_id"] = df["course_id"].astype(int)
+
+    # Parse times (UTC) and convert to local Pacific time (user asked for “Pacific standard time”;
+    # we’ll convert to America/Los_Angeles which accounts for DST automatically).
+    df["ts_utc"] = pd.to_datetime(df["created_at"], utc=True, errors="coerce")
+    df = df.dropna(subset=["ts_utc"]).copy()
+    local_tz = ZoneInfo(tz_name)
+    df["ts_local"] = df["ts_utc"].dt.tz_convert(local_tz)
+
+    # Sort for gap detection
+    df = df.sort_values(["course_id", "session_id", "ts_local"])
+
+    # Session splitting: new session if gap > gap_minutes within same course+session_id
+    gap = pd.Timedelta(minutes=gap_minutes)
+    # Identify gap starts within each course+session_id stream
+    df["gap_new"] = (
+        df.groupby(["course_id", "session_id"])["ts_local"]
+        .diff()
+        .gt(gap)
+        .fillna(True)  # first row starts a session
+    )
+    # Session index within each course+session_id
+    df["session_idx"] = df.groupby(["course_id", "session_id"])["gap_new"].cumsum()
+
+    # Session key across all data
+    df["session_key"] = list(zip(df["course_id"], df["session_id"], df["session_idx"]))
+
+    # Aggregate to sessions
+    agg = (
+        df.groupby("session_key")
+        .agg(course_id=("course_id", "first"),
+            start=("ts_local", "min"),
+            end=("ts_local", "max"),
+            hits=("ts_local", "size"))
+        .reset_index(drop=True)
+    )
+
+    # Duration in minutes (ceil; minimum 1 minute)
+    dur_secs = (agg["end"] - agg["start"]).dt.total_seconds().clip(lower=0)
+    agg["mins"] = [max(1, math.ceil(s / 60)) for s in dur_secs]
+
+    # Group by local calendar day for reporting
+    agg["day"] = agg["start"].dt.strftime("%b %-d")  # e.g., "Oct 27"
+    # Format start time like "5:05pm"
+    agg["start_str"] = agg["start"].dt.strftime("%-I:%M%p").str.lower()
+
+    # Order: day, then chronological within day
+    agg = agg.sort_values(["start"])
+
+    # Group by calendar date only
+    agg["day_key"] = agg["start"].dt.date
+
+    # Build text report with nice formatting
+    lines: list[str] = []
+    for day_key, day_df in agg.groupby("day_key", sort=False):
+        # Pretty date for the section header
+        # Using the first session time for that day to include time
+        first_time = day_df["start"].min()
+        # linux
+        # pretty_day = first_time.strftime("%b %-d %-I:%M%p").lower()  # e.g., "Sep 26 8:02pm"
+        
+        # windows
+        pretty_day = first_time.strftime("%b %#d %#I:%M%p").lower()
+        lines.append(f"{pretty_day}:")
+        for _, row in day_df.iterrows():
+            lines.append(
+                f" - course {row.course_id}, {row.start_str}: "
+                f"{row.hits} hits over {row.mins} minutes"
+            )
+    report_out = codecs.open(f"cache/users/logs/{id}_report.txt", "w", "utf-8")
+    report_out.write( "\n".join(lines) )
+
+    return "\n".join(lines) if lines else "No course activity found."
+
+
+    # Example usage:
+    # print(generate_course_report("canvas_logs.csv", gap_minutes=10))
 
 
 if __name__ == "__main__":
@@ -2851,6 +2971,7 @@ if __name__ == "__main__":
 
                 40: ['reset user email without confirmation', set_email_skip_confirm],
                 41: ['summarize users logs', summarize_student_logs],
+                50: ['summarize users logs 2', readable_user_summary], 
 
 
                 #3: ['Main index, 1 year, teachers and their classes', getAllTeachersInTerm],