From e983cd5c56946daa8d744c08ab7a7a93572c4555 Mon Sep 17 00:00:00 2001
From: Peter Howell <peter.howell@gmail.com>
Date: Thu, 24 Jul 2025 14:20:25 -0700
Subject: [PATCH] user tracking improved slightly

---
 users.py | 29 +++++++++++++++++++++++++----
 1 file changed, 25 insertions(+), 4 deletions(-)

diff --git a/users.py b/users.py
index 3c996b1..d7ffd63 100644
--- a/users.py
+++ b/users.py
@@ -6,6 +6,8 @@ import pytz, time
 import pandas as pd
 import matplotlib.pyplot as plt
 
+
+
 #from pandas import TimeGrouper
 from PIL import Image
 from collections import defaultdict
@@ -51,7 +53,7 @@ lds_stamp = parser.parse(log_default_startdate)
 
 recvd_date = '2023-01-01T00:00:00Z'
 num_threads = 25
-max_log_count = 250000
+max_log_count = 500000
 
 
 ##########
@@ -1763,6 +1765,23 @@ def track_user_q(id, q):
         q.task_done()
 
 
+
+def count_courses(csv_file="cache/users/logs/45268.csv"):
+    course_counts = defaultdict(int)
+    course_regex = r'courses/(\d+)/'
+
+    with open(csv_file, mode='r') as file:
+        reader = csv.reader(file)
+        for row in reader:
+            match = re.search(course_regex, row[1])
+            if match:
+                course_id = match.group(1)
+                course_counts[course_id] += 1
+
+    for course_id, count in course_counts.items():
+        print(f"Course ID {course_id}: {count} occurrences")
+
+
 # honestly it doesn't make much sense to get full histories this way if they're
 # already in the canvas data tables....
 
@@ -1801,8 +1820,9 @@ def track_user(id=0,qid=0):
         url_addition = ""
         
         if 1:    # hard code dates
-            start_date = "2024-01-01T00:00:00-07:00"
-            end_date   = dt.now().strftime("%Y-%m-%dT%H:%M:%S-07:00")                               # "2024-10-01T00:00:00-07:00"
+            start_date = "2022-01-01T00:00:00-07:00"
+            end_date   = dt.now().strftime("%Y-%m-%dT%H:%M:%S-07:00")                               # 
+            end_date = "2022-07-01T00:00:00-07:00"
             url_addition = f"?start_time={start_date}&end_time={end_date}"
         elif 'last_days_log' in info:
             print("There's existing log data for %s (%s)" % (info['name'] , info['sis_user_id']))
@@ -2639,7 +2659,8 @@ if __name__ == "__main__":
                 8: ['Gather teacher history, a variety of stats.', getTeachersInfoMain],
                 9: ['test rtr.', read_training_records],
                 10: ['Get a users logs', track_user],
-                11: ['test: oneYearSchedule', oneYearSchedule],
+                11: ['tally a users logs from csv file', count_courses],
+                #11: ['test: oneYearSchedule', oneYearSchedule],
                 12: ['summarize hit activity', activity_summary],
                 13: ['Get all users logs in a class', track_users_in_class],
                 14: ['Get logs for 1 user', track_a_user ],