From 976dafc351de9d70da9826fc57215ff3711ca81a Mon Sep 17 00:00:00 2001
From: Nate-Wessel <nate.wessel@toronto.ca>
Date: Tue, 23 Apr 2024 17:51:51 +0000
Subject: [PATCH] compute a bootstrap sample dist

---
 backend/app/get_travel_time.py | 101 ++++++++++-----------------------
 1 file changed, 30 insertions(+), 71 deletions(-)

diff --git a/backend/app/get_travel_time.py b/backend/app/get_travel_time.py
index 5f893db..9b273b2 100644
--- a/backend/app/get_travel_time.py
+++ b/backend/app/get_travel_time.py
@@ -1,5 +1,7 @@
 from app.db import getConnection
 from app.get_links import get_links
+import numpy
+import random
 
 def get_travel_time(start_node, end_node, start_time, end_time, start_date, end_date, include_holidays, dow_list):
     
@@ -13,72 +15,24 @@ def get_travel_time(start_node, end_node, start_time, end_time, start_date, end_
             SELECT 1 FROM ref.holiday WHERE ta_path.dt = holiday.dt
         )'''
 
-    
-    agg_tt_query = f''' 
-        -- Aggregate segments to corridor on a daily, hourly basis
-        WITH corridor_hourly_daily_agg AS (
-            SELECT
-                cn.dt,
-                cn.hr,
-                SUM(cn.unadjusted_tt) AS corr_hourly_daily_tt
-            FROM congestion.network_segments_daily AS cn
-            WHERE
-                cn.segment_id::integer = ANY(%(seglist)s)
-                AND cn.hr <@ %(time_range)s::numrange
-                AND date_part('ISODOW', cn.dt)::integer = ANY(%(dow_list)s)
-                AND cn.dt <@ %(date_range)s::daterange
-                {tt_holiday_clause}
-            GROUP BY
-                cn.dt,
-                cn.hr
-            -- where corridor has at least 80pct of links with data
-            HAVING SUM(cn.length_w_data) >= %(length_m)s::numeric * 0.8
-        ),
-
-        -- Average the hours selected into daily period level data
-        corridor_period_daily_avg_tt AS (
-            SELECT
-                dt,
-                AVG(corr_hourly_daily_tt) AS avg_corr_period_daily_tt
-            FROM corridor_hourly_daily_agg
-            GROUP BY dt
-        )
-
-        -- Average all the days with data to get period level data for each date range
-        SELECT 
-            ROUND(AVG(avg_corr_period_daily_tt) / 60, 2) AS average_tt_min
-        FROM corridor_period_daily_avg_tt
-    '''
-
-    sample_size_query = f"""
-        SELECT SUM( ((length / 1000.0) / mean) * sample_size ) AS probe_hours
-        FROM here.ta_path
-        WHERE
-            link_dir = ANY(%(link_dir_list)s)
-            AND dt <@ %(date_range)s::daterange
-            AND EXTRACT(ISODOW FROM dt)::integer = ANY(%(dow_list)s)
-            AND EXTRACT(HOUR FROM tod)::numeric <@ %(time_range)s::numrange
-            {sample_holiday_clause}
-    """
-
-    sample_hour_query = f"""
+    hourly_tt_query = f'''
         SELECT
-            cn.dt,
-            cn.hr,
-            SUM(cn.unadjusted_tt) AS corr_hourly_daily_tt
+            dt,
+            hr,
+            SUM(cn.unadjusted_tt) * %(length_m)s::real / SUM(cn.length_w_data) AS tt
         FROM congestion.network_segments_daily AS cn
-        WHERE   
-            cn.segment_id::integer IN %(seglist)s
+        WHERE
+            cn.segment_id::integer = ANY(%(seglist)s)
             AND cn.hr <@ %(time_range)s::numrange
-            AND date_part('ISODOW', cn.dt)::integer IN %(dow_list)s
-            AND cn.dt <@ %(date_range)s::daterange 
+            AND date_part('ISODOW', cn.dt)::integer = ANY(%(dow_list)s)
+            AND cn.dt <@ %(date_range)s::daterange
             {tt_holiday_clause}
         GROUP BY
             cn.dt,
             cn.hr
         -- where corridor has at least 80pct of links with data
-        HAVING SUM(cn.length_w_data) >= %(length_m)s::numeric * 0.8
-    """
+        HAVING SUM(cn.length_w_data) >= %(length_m)s::numeric * 0.8;
+    '''
 
     links = get_links(start_node, end_node)
 
@@ -97,20 +51,25 @@ def get_travel_time(start_node, end_node, start_time, end_time, start_date, end_
     connection = getConnection()
     with connection:
         with connection.cursor() as cursor:
-            cursor.execute(agg_tt_query, query_params)
-            # travel_time may be null if there's insufficient data
-            travel_time, = cursor.fetchone()
-            cursor.execute(sample_size_query, query_params)
-            probe_hours, = cursor.fetchone()
+            # get the hourly travel times
+            cursor.execute(hourly_tt_query, query_params)
+            tt_hourly = [ tt for (dt, hr, tt) in cursor.fetchall() ]
+    connection.close()
 
-            cursor.execute(sample_hour_query, query_params)
-            hour_bins_used = cursor.rowcount
+    # bootstrap for synthetic sample distribution
+    sample_distribution = []
+    for i in range(0,100):
+        bootstrap_sample = random.choices(
+            tt_hourly,
+            k = len(tt_hourly)
+        )
+        sample_distribution.append( numpy.mean(bootstrap_sample) )
 
-    connection.close()
     return {
-        'bins': hour_bins_used,
-        'travel_time': None if travel_time is None else float(travel_time),
-        'links': links,
-        'estimated_vehicle_count': None if travel_time is None else float((probe_hours * 60) / travel_time),
-        'query_params': query_params
+        'average_travel_time': numpy.mean(tt_hourly),
+        'upper': numpy.percentile(sample_distribution,95),
+        'lower': numpy.percentile(sample_distribution,5),
+        'hourly_travel_times': tt_hourly
+        #'links': links,
+        #'query_params': query_params
     }
\ No newline at end of file