1_generic_vs_ml_overview.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "f4bcec73",
   "metadata": {},
   "source": [
    "# Section 1: Generic vs ML Overview"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "afe64156-fc33-4529-80e3-688105952ab9",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import pyspark.sql.functions as F\n",
    "import pyspark.sql.types as T\n",
    "from util.read_and_print_df import *\n",
    "from util.handle_nan import *\n",
    "from util.extract_json_attributes import *\n",
    "from util.plotting import *\n",
    "from tabulate import tabulate"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "0c3b999c-61ee-4aed-88eb-975ebf8daacf",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Assigning 541 GB of memory per spark driver and executor, and use 126 cores.\n"
     ]
    }
   ],
   "source": [
    "spark = get_spark_session()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "a311e3f0-58d9-4901-a53b-c13e44f4ed8c",
   "metadata": {},
   "outputs": [],
   "source": [
    "pydf = spark.read.parquet(path_job_node_joined_dataset)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "24c3ac85-2d05-4f2c-974f-2a3f25e555b3",
   "metadata": {},
   "outputs": [],
   "source": [
    "gen_data = pydf.where(F.col(\"gpu_node\") == 0)\n",
    "ml_data = pydf.where(F.col(\"gpu_node\") == 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "1728d1dd-d74e-4341-a1b2-a45e4d5cbe3d",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "                                                                                \r"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "287 51\n"
     ]
    }
   ],
   "source": [
    "gen_count = gen_data.select(F.col(\"node\")).distinct().count()\n",
    "ml_count = ml_data.select(F.col(\"node\")).distinct().count()\n",
    "print(gen_count, ml_count)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "94227d12",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['OUT_OF_MEMORY', 'TIMEOUT', 'FAILED', 'CANCELLED', 'NODE_FAIL', 'COMPLETED']\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "                                                                                \r"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "+-----+-------------------+-------------+-------+------+---------+---------+---------+------------------+\n",
      "|node |timestamp          |OUT_OF_MEMORY|TIMEOUT|FAILED|CANCELLED|NODE_FAIL|COMPLETED|total_jobs_running|\n",
      "+-----+-------------------+-------------+-------+------+---------+---------+---------+------------------+\n",
      "|r25n4|2022-10-30 09:56:00|0            |1      |0     |0        |0        |0        |1                 |\n",
      "|r25n4|2022-10-30 13:59:30|0            |1      |0     |0        |0        |0        |1                 |\n",
      "|r25n4|2022-10-30 21:15:30|0            |0      |0     |0        |0        |1        |1                 |\n",
      "|r25n4|2022-10-30 21:32:30|0            |0      |0     |0        |0        |1        |1                 |\n",
      "|r25n4|2022-10-30 22:42:00|0            |0      |0     |0        |0        |1        |1                 |\n",
      "+-----+-------------------+-------------+-------+------+---------+---------+---------+------------------+\n",
      "only showing top 5 rows\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Stage 31:=================================================>   (178 + 11) / 189]\r"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "+-----+-------------------+--------+-----------------+-----------+----------------+---+------------------------+-----------------------+---------------------------+----------+----------+------------------------+----------------------------+-------------------------+-----------------------+-----------------+------------------------+----------------------------+---------------+-----------------------+------------------------+----------------------+-----------------------------+--------------------------+------------------+-------------------------+-------------------------+------------------+----------------+---------------------------+----------------------------+---------------------------+-------------------------------+-------------------------+-------------------------------------+--------------------------------+---------------------------------+------------------------------+--------------------------+---------------------------+--------------------------+------------------------------+--------------------------------+--------------------------------+--------------------------------+--------------------------------+--------------------------------+----------------------------------+-----------------------------------+----------------------------------+--------------------+-------------------------------+--------------------------------+-------------------------------+------------------------------+-----------------------------------+-----------------------------------+-----------------------------------+-----------------------------------+-------------------------------------+-------------------------------------+--------------------------------------+-------------------------------------+------------------------------+------------------------------------+-------------------------+--------------------------+-------------------------+---------------------------------------+-------------------+------------------------------------+--------------------------------------+----------------------------------------+--------------------+----------------------------------+-----------------------------------+-----------------+-------------------------------------------+---------------------------------------+------------------------------------+------------------------------------------+---------------------------------------------+------------------------------------------+--------------------------------------------+----------------------------------------------+----------------------------------------+-----------------------------------------+-----------------------+---------------------------+--------+--------+-------------------+-------------------+-------------------+---------+--------+--------+---------+-----------+--------------------+----+-------------+-------+------+---------+---------+---------+------------------+------------------------+\n",
      "|node |timestamp          |prom_id |node_time_seconds|node_load15|node_power_usage|up |node_netstat_Tcp_OutSegs|node_netstat_Tcp_InErrs|node_context_switches_total|node_load5|node_load1|node_memory_Active_bytes|node_netstat_Tcp_RetransSegs|node_netstat_Udp_InErrors|node_memory_Dirty_bytes|node_ambient_temp|node_netstat_Icmp_InMsgs|node_netstat_Udp_InDatagrams|node_intr_total|node_netstat_Tcp_InSegs|node_memory_Percpu_bytes|node_boot_time_seconds|node_netstat_Udp_OutDatagrams|node_netstat_Icmp_InErrors|node_procs_blocked|node_netstat_Icmp_OutMsgs|node_memory_MemFree_bytes|node_procs_running|node_forks_total|node_hwmon_temp_celsius_min|node_hwmon_temp_celsius_mean|node_hwmon_temp_celsius_max|node_filesystem_avail_bytes_sum|node_filesystem_files_sum|node_network_transmit_bytes_total_sum|node_filesystem_device_error_sum|node_disk_written_bytes_total_sum|node_filesystem_free_bytes_sum|node_thermal_zone_temp_min|node_thermal_zone_temp_mean|node_thermal_zone_temp_max|node_disk_read_bytes_total_sum|nvidia_gpu_memory_used_bytes_sum|nvidia_gpu_temperature_celsius_0|nvidia_gpu_temperature_celsius_1|nvidia_gpu_temperature_celsius_2|nvidia_gpu_temperature_celsius_3|nvidia_gpu_temperature_celsius_min|nvidia_gpu_temperature_celsius_mean|nvidia_gpu_temperature_celsius_max|node_arp_entries_sum|nvidia_gpu_fanspeed_percent_min|nvidia_gpu_fanspeed_percent_mean|nvidia_gpu_fanspeed_percent_max|node_filesystem_files_free_sum|nvidia_gpu_power_usage_milliwatts_0|nvidia_gpu_power_usage_milliwatts_1|nvidia_gpu_power_usage_milliwatts_2|nvidia_gpu_power_usage_milliwatts_3|nvidia_gpu_power_usage_milliwatts_min|nvidia_gpu_power_usage_milliwatts_sum|nvidia_gpu_power_usage_milliwatts_mean|nvidia_gpu_power_usage_milliwatts_max|node_filesystem_size_bytes_sum|node_disk_writes_completed_total_sum|nvidia_gpu_duty_cycle_min|nvidia_gpu_duty_cycle_mean|nvidia_gpu_duty_cycle_max|node_network_transmit_packets_total_sum|node_udp_queues_sum|node_network_receive_bytes_total_sum|node_network_receive_packets_total_sum|node_network_receive_multicast_total_sum|node_disk_io_now_sum|node_rapl_package_joules_total_sum|node_network_receive_drop_total_sum|timestamp_seconds|node_network_transmit_bytes_total_sum_delta|node_disk_written_bytes_total_sum_delta|node_disk_read_bytes_total_sum_delta|node_disk_writes_completed_total_sum_delta|node_network_transmit_packets_total_sum_delta|node_network_receive_bytes_total_sum_delta|node_network_receive_packets_total_sum_delta|node_network_receive_multicast_total_sum_delta|node_rapl_package_joules_total_sum_delta|node_network_receive_drop_total_sum_delta|timestamp_seconds_delta|node_rapl_package_power_sum|gpu_node|slurm_id|submit_date        |start_date         |end_date           |nodetypes|numnodes|numcores|state    |slurm_nodes|job_duration_seconds|rack|OUT_OF_MEMORY|TIMEOUT|FAILED|CANCELLED|NODE_FAIL|COMPLETED|total_jobs_running|node_power_usage_per_job|\n",
      "+-----+-------------------+--------+-----------------+-----------+----------------+---+------------------------+-----------------------+---------------------------+----------+----------+------------------------+----------------------------+-------------------------+-----------------------+-----------------+------------------------+----------------------------+---------------+-----------------------+------------------------+----------------------+-----------------------------+--------------------------+------------------+-------------------------+-------------------------+------------------+----------------+---------------------------+----------------------------+---------------------------+-------------------------------+-------------------------+-------------------------------------+--------------------------------+---------------------------------+------------------------------+--------------------------+---------------------------+--------------------------+------------------------------+--------------------------------+--------------------------------+--------------------------------+--------------------------------+--------------------------------+----------------------------------+-----------------------------------+----------------------------------+--------------------+-------------------------------+--------------------------------+-------------------------------+------------------------------+-----------------------------------+-----------------------------------+-----------------------------------+-----------------------------------+-------------------------------------+-------------------------------------+--------------------------------------+-------------------------------------+------------------------------+------------------------------------+-------------------------+--------------------------+-------------------------+---------------------------------------+-------------------+------------------------------------+--------------------------------------+----------------------------------------+--------------------+----------------------------------+-----------------------------------+-----------------+-------------------------------------------+---------------------------------------+------------------------------------+------------------------------------------+---------------------------------------------+------------------------------------------+--------------------------------------------+----------------------------------------------+----------------------------------------+-----------------------------------------+-----------------------+---------------------------+--------+--------+-------------------+-------------------+-------------------+---------+--------+--------+---------+-----------+--------------------+----+-------------+-------+------+---------+---------+---------+------------------+------------------------+\n",
      "|r11n8|2022-07-11 07:14:30|2534142 |1.65752E9        |1.0        |92.0            |1.0|2.77236E7               |0.0                    |2.19161E10                 |1.0       |1.0       |4.92327E9               |606.0                       |0.0                      |12288.0                |24.0             |57.0                    |11150.0                     |1.60376E8      |1.19344E7              |8454140.0               |1.6572E9              |21743.0                      |0.0                       |0.0               |57.0                     |9.3166E10                |14.0              |5421580.0       |32.0                       |33.24                       |37.0                       |1.883543642112E12              |2.30696998E8             |1.80337317385E11                     |0.0                             |4.6164197376E10                  |1.883597328384E12             |37.0                      |37.0                       |37.0                      |1.04135424E9                  |0.0                             |0.0                             |0.0                             |0.0                             |0.0                             |0.0                               |0.0                                |0.0                               |30.0                |0.0                            |0.0                             |0.0                            |2.30050466E8                  |0.0                                |0.0                                |0.0                                |0.0                                |0.0                                  |0.0                                  |0.0                                   |0.0                                  |1.912248328192E12             |476733.0                            |0.0                      |0.0                       |0.0                      |2.7763322E7                            |0.0                |2.25066714868E11                    |3.0323599E7                           |2.0                                     |0.0                 |50961.26                          |0.0                                |1657516470       |32797.0                                    |26112.0                                |0.0                                 |4.0                                       |30.0                                         |6939.0                                    |96.0                                        |0.0                                           |1501.8700000000026                      |0.0                                      |30.0                   |50.06233333333342          |0       |1616537 |2022-07-10 20:54:35|2022-07-10 20:55:29|2022-07-12 13:00:46|normal   |1       |16      |TIMEOUT  |[r11n8]    |144317              |r11 |0            |1      |0     |0        |0        |0        |1                 |92.0                    |\n",
      "|r11n9|2022-09-13 00:55:00|80033970|1.66303E9        |15.08      |132.0           |1.0|2.03254E9               |0.0                    |9.03307E8                  |15.07     |15.1      |3.4735E10               |1267160.0                   |0.0                      |8192.0                 |24.0             |208.0                   |39982.0                     |4.81816E9      |1.68131E9              |1.58597E7               |1.66127E9             |93615.0                      |70.0                      |0.0               |206.0                    |5.98747E10               |18.0              |4.08223E7       |42.0                       |43.06                       |44.0                       |1.881988591616E12              |2.30696824E8             |6.335930547845E12                    |0.0                             |1.0707470464E12                  |1.882042277888E12             |44.0                      |44.0                       |44.0                      |1.26521639936E11              |0.0                             |0.0                             |0.0                             |0.0                             |0.0                             |0.0                               |0.0                                |0.0                               |114.0               |0.0                            |0.0                             |0.0                            |2.30051772E8                  |0.0                                |0.0                                |0.0                                |0.0                                |0.0                                  |0.0                                  |0.0                                   |0.0                                  |1.91224825856E12              |3404689.0                           |0.0                      |0.0                       |0.0                      |2.034060221E9                          |0.0                |7.754100076435E12                   |2.141911937E9                         |2.0                                     |0.0                 |176978.75                         |0.0                                |1663030500       |68446.0                                    |17920.0                                |0.0                                 |3.0                                       |68.0                                         |13738.0                                   |93.0                                        |0.0                                           |2369.390000000014                       |0.0                                      |30.0                   |78.97966666666713          |0       |1986818 |2022-09-04 17:25:08|2022-09-09 07:20:05|2022-09-13 08:06:02|shared   |1       |3       |COMPLETED|[r11n9]    |348357              |r11 |0            |0      |0     |0        |0        |5        |5                 |26.4                    |\n",
      "|r11n9|2022-09-13 00:55:00|80033970|1.66303E9        |15.08      |132.0           |1.0|2.03254E9               |0.0                    |9.03307E8                  |15.07     |15.1      |3.4735E10               |1267160.0                   |0.0                      |8192.0                 |24.0             |208.0                   |39982.0                     |4.81816E9      |1.68131E9              |1.58597E7               |1.66127E9             |93615.0                      |70.0                      |0.0               |206.0                    |5.98747E10               |18.0              |4.08223E7       |42.0                       |43.06                       |44.0                       |1.881988591616E12              |2.30696824E8             |6.335930547845E12                    |0.0                             |1.0707470464E12                  |1.882042277888E12             |44.0                      |44.0                       |44.0                      |1.26521639936E11              |0.0                             |0.0                             |0.0                             |0.0                             |0.0                             |0.0                               |0.0                                |0.0                               |114.0               |0.0                            |0.0                             |0.0                            |2.30051772E8                  |0.0                                |0.0                                |0.0                                |0.0                                |0.0                                  |0.0                                  |0.0                                   |0.0                                  |1.91224825856E12              |3404689.0                           |0.0                      |0.0                       |0.0                      |2.034060221E9                          |0.0                |7.754100076435E12                   |2.141911937E9                         |2.0                                     |0.0                 |176978.75                         |0.0                                |1663030500       |68446.0                                    |17920.0                                |0.0                                 |3.0                                       |68.0                                         |13738.0                                   |93.0                                        |0.0                                           |2369.390000000014                       |0.0                                      |30.0                   |78.97966666666713          |0       |1985677 |2022-09-04 17:25:08|2022-09-09 09:55:16|2022-09-13 04:07:57|shared   |1       |3       |COMPLETED|[r11n9]    |324761              |r11 |0            |0      |0     |0        |0        |5        |5                 |26.4                    |\n",
      "|r11n9|2022-09-13 00:55:00|80033970|1.66303E9        |15.08      |132.0           |1.0|2.03254E9               |0.0                    |9.03307E8                  |15.07     |15.1      |3.4735E10               |1267160.0                   |0.0                      |8192.0                 |24.0             |208.0                   |39982.0                     |4.81816E9      |1.68131E9              |1.58597E7               |1.66127E9             |93615.0                      |70.0                      |0.0               |206.0                    |5.98747E10               |18.0              |4.08223E7       |42.0                       |43.06                       |44.0                       |1.881988591616E12              |2.30696824E8             |6.335930547845E12                    |0.0                             |1.0707470464E12                  |1.882042277888E12             |44.0                      |44.0                       |44.0                      |1.26521639936E11              |0.0                             |0.0                             |0.0                             |0.0                             |0.0                             |0.0                               |0.0                                |0.0                               |114.0               |0.0                            |0.0                             |0.0                            |2.30051772E8                  |0.0                                |0.0                                |0.0                                |0.0                                |0.0                                  |0.0                                  |0.0                                   |0.0                                  |1.91224825856E12              |3404689.0                           |0.0                      |0.0                       |0.0                      |2.034060221E9                          |0.0                |7.754100076435E12                   |2.141911937E9                         |2.0                                     |0.0                 |176978.75                         |0.0                                |1663030500       |68446.0                                    |17920.0                                |0.0                                 |3.0                                       |68.0                                         |13738.0                                   |93.0                                        |0.0                                           |2369.390000000014                       |0.0                                      |30.0                   |78.97966666666713          |0       |1988217 |2022-09-04 17:25:08|2022-09-09 10:30:15|2022-09-13 17:22:13|shared   |1       |3       |COMPLETED|[r11n9]    |370318              |r11 |0            |0      |0     |0        |0        |5        |5                 |26.4                    |\n",
      "|r11n9|2022-09-13 00:55:00|80033970|1.66303E9        |15.08      |132.0           |1.0|2.03254E9               |0.0                    |9.03307E8                  |15.07     |15.1      |3.4735E10               |1267160.0                   |0.0                      |8192.0                 |24.0             |208.0                   |39982.0                     |4.81816E9      |1.68131E9              |1.58597E7               |1.66127E9             |93615.0                      |70.0                      |0.0               |206.0                    |5.98747E10               |18.0              |4.08223E7       |42.0                       |43.06                       |44.0                       |1.881988591616E12              |2.30696824E8             |6.335930547845E12                    |0.0                             |1.0707470464E12                  |1.882042277888E12             |44.0                      |44.0                       |44.0                      |1.26521639936E11              |0.0                             |0.0                             |0.0                             |0.0                             |0.0                             |0.0                               |0.0                                |0.0                               |114.0               |0.0                            |0.0                             |0.0                            |2.30051772E8                  |0.0                                |0.0                                |0.0                                |0.0                                |0.0                                  |0.0                                  |0.0                                   |0.0                                  |1.91224825856E12              |3404689.0                           |0.0                      |0.0                       |0.0                      |2.034060221E9                          |0.0                |7.754100076435E12                   |2.141911937E9                         |2.0                                     |0.0                 |176978.75                         |0.0                                |1663030500       |68446.0                                    |17920.0                                |0.0                                 |3.0                                       |68.0                                         |13738.0                                   |93.0                                        |0.0                                           |2369.390000000014                       |0.0                                      |30.0                   |78.97966666666713          |0       |1990600 |2022-09-04 17:27:40|2022-09-12 16:51:34|2022-09-14 18:52:19|shared   |1       |3       |COMPLETED|[r11n9]    |180045              |r11 |0            |0      |0     |0        |0        |5        |5                 |26.4                    |\n",
      "+-----+-------------------+--------+-----------------+-----------+----------------+---+------------------------+-----------------------+---------------------------+----------+----------+------------------------+----------------------------+-------------------------+-----------------------+-----------------+------------------------+----------------------------+---------------+-----------------------+------------------------+----------------------+-----------------------------+--------------------------+------------------+-------------------------+-------------------------+------------------+----------------+---------------------------+----------------------------+---------------------------+-------------------------------+-------------------------+-------------------------------------+--------------------------------+---------------------------------+------------------------------+--------------------------+---------------------------+--------------------------+------------------------------+--------------------------------+--------------------------------+--------------------------------+--------------------------------+--------------------------------+----------------------------------+-----------------------------------+----------------------------------+--------------------+-------------------------------+--------------------------------+-------------------------------+------------------------------+-----------------------------------+-----------------------------------+-----------------------------------+-----------------------------------+-------------------------------------+-------------------------------------+--------------------------------------+-------------------------------------+------------------------------+------------------------------------+-------------------------+--------------------------+-------------------------+---------------------------------------+-------------------+------------------------------------+--------------------------------------+----------------------------------------+--------------------+----------------------------------+-----------------------------------+-----------------+-------------------------------------------+---------------------------------------+------------------------------------+------------------------------------------+---------------------------------------------+------------------------------------------+--------------------------------------------+----------------------------------------------+----------------------------------------+-----------------------------------------+-----------------------+---------------------------+--------+--------+-------------------+-------------------+-------------------+---------+--------+--------+---------+-----------+--------------------+----+-------------+-------+------+---------+---------+---------+------------------+------------------------+\n",
      "only showing top 5 rows\n",
      "\n"
     ]
    }
   ],
   "source": [
    "df_prom_slurm = pydf.withColumn('job_duration_seconds', F.unix_timestamp('end_date') - F.unix_timestamp('start_date'))\n",
    "df_prom_slurm = get_gpu_node_col(df_prom_slurm, 'node')\n",
    "df_prom_slurm = df_prom_slurm.withColumn('rack', F.regexp_extract('node', r'(r[0-9]+)n', 1))\n",
    "\n",
    "states = df_prom_slurm.select('state').distinct().rdd.map(lambda r: r[0]).collect()\n",
    "print(states)\n",
    "\n",
    "# group by node, timestamp and aggregate by summing up the counts for each state in a new column\n",
    "df_prom_slurm_job_count = df_prom_slurm.groupBy('node', 'timestamp').agg(*[F.sum(when(col('state') == state, 1).otherwise(0)).alias(state) for state in states])\n",
    "df_prom_slurm_job_count = df_prom_slurm_job_count.withColumn('total_jobs_running', F.expr(' + '.join(states)))\n",
    "df_prom_slurm_job_count.show(5, False)\n",
    "\n",
    "df_prom_slurm_2 = df_prom_slurm.join(df_prom_slurm_job_count, ['node', 'timestamp'], how='left')\n",
    "df_prom_slurm_2 = df_prom_slurm_2.withColumn('node_power_usage_per_job', col('node_power_usage') / col('total_jobs_running'))\n",
    "df_prom_slurm_2.show(5, False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "8a4b21a9",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "                                                                                \r"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "+--------+-------------+--------------------------------+\n",
      "|gpu_node|state        |job_type_node_energy_usage_total|\n",
      "+--------+-------------+--------------------------------+\n",
      "|0       |CANCELLED    |2.658925779628305E10            |\n",
      "|0       |COMPLETED    |1.377435808060837E11            |\n",
      "|0       |FAILED       |1.0661996749175968E10           |\n",
      "|0       |NODE_FAIL    |7.603104488755941E9             |\n",
      "|0       |OUT_OF_MEMORY|2.2797041191898007E9            |\n",
      "|0       |TIMEOUT      |8.517718885303772E10            |\n",
      "|1       |CANCELLED    |2.9534506993850616E10           |\n",
      "|1       |COMPLETED    |8.315266117146094E10            |\n",
      "|1       |FAILED       |7.945788024926845E9             |\n",
      "|1       |NODE_FAIL    |4.4316968978609633E8            |\n",
      "|1       |OUT_OF_MEMORY|1.7887963345175433E9            |\n",
      "|1       |TIMEOUT      |4.749171268834755E10            |\n",
      "+--------+-------------+--------------------------------+\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "                                                                                \r"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "+--------+-------------------------------------+\n",
      "|gpu_node|sum(job_type_node_energy_usage_total)|\n",
      "+--------+-------------------------------------+\n",
      "|0       |2.700548328125262E11                 |\n",
      "|1       |1.703566349028896E11                 |\n",
      "+--------+-------------------------------------+\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Stage 152:===================================>                (136 + 64) / 200]\r"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Percentage of energy generic: 61.32%     Percentage of energy ML: 38.68%\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "                                                                                \r"
     ]
    }
   ],
   "source": [
    "df_prom_slurm_energy = df_prom_slurm_2.groupBy('slurm_id', 'node').agg(\n",
    "    F.mean('node_power_usage_per_job').alias('mean_job_node_power_usage'),\n",
    "    F.first('job_duration_seconds').alias('job_duration_seconds'),\n",
    "    F.first('gpu_node').alias('gpu_node'),\n",
    "    F.first('state').alias('state')\n",
    ").withColumn('job_node_energy_usage', F.col('mean_job_node_power_usage') * F.col('job_duration_seconds')).groupBy('slurm_id').agg(\n",
    "    F.sum('job_node_energy_usage').alias('multinode_job_node_energy_usage'),\n",
    "    F.first('gpu_node').alias('gpu_node'),\n",
    "    F.first('state').alias('state')\n",
    ").groupBy('gpu_node', 'state').agg(\n",
    "    F.sum('multinode_job_node_energy_usage').alias('job_type_node_energy_usage_total')\n",
    ")\n",
    "df_prom_slurm_energy.orderBy('gpu_node', 'state').show(20, False)\n",
    "df_prom_slurm_energy_sum = df_prom_slurm_energy.groupBy('gpu_node').sum('job_type_node_energy_usage_total')\n",
    "df_prom_slurm_energy_sum.orderBy('gpu_node').show(5, False)\n",
    "\n",
    "total_energy = df_prom_slurm_energy_sum.select(F.sum('sum(job_type_node_energy_usage_total)')).collect()[0][0]\n",
    "\n",
    "print(f\"Percentage of energy generic: {df_prom_slurm_energy_sum.filter(F.col('gpu_node') == 0).collect()[0][1] / total_energy * 100:.2f}% \\\n",
    "    Percentage of energy ML: {df_prom_slurm_energy_sum.filter(F.col('gpu_node') == 1).collect()[0][1] / total_energy * 100:.2f}%\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "291522ea",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Stage 171:====================================================>(188 + 1) / 189]\r"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "+--------+-------------+------+\n",
      "|gpu_node|state        |count |\n",
      "+--------+-------------+------+\n",
      "|1       |CANCELLED    |3669  |\n",
      "|0       |CANCELLED    |9592  |\n",
      "|0       |FAILED       |37588 |\n",
      "|1       |OUT_OF_MEMORY|130   |\n",
      "|1       |NODE_FAIL    |107   |\n",
      "|1       |COMPLETED    |17306 |\n",
      "|0       |OUT_OF_MEMORY|4499  |\n",
      "|1       |FAILED       |4664  |\n",
      "|1       |TIMEOUT      |1411  |\n",
      "|0       |COMPLETED    |202874|\n",
      "|0       |NODE_FAIL    |1426  |\n",
      "|0       |TIMEOUT      |10625 |\n",
      "+--------+-------------+------+\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "                                                                                \r"
     ]
    }
   ],
   "source": [
    "# Job Failures\n",
    "df_prom_slurm_state_count = df_prom_slurm.groupBy('slurm_id').agg(F.first('gpu_node').alias('gpu_node'), F.first('state').alias('state')).groupBy('gpu_node', 'state').count()\n",
    "df_prom_slurm_state_count.show(20, False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "d17e901c-2985-417f-99e3-17482ef881e7",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "                                                                                \r"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>type</th>\n",
       "      <th>Rack Count</th>\n",
       "      <th>Node Count</th>\n",
       "      <th>Job Energy</th>\n",
       "      <th>Job Failures</th>\n",
       "      <th>Job Submissions</th>\n",
       "      <th>Job Runtime</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Generic</td>\n",
       "      <td>0.52381</td>\n",
       "      <td>0.849112</td>\n",
       "      <td>0.613188</td>\n",
       "      <td>0.889615</td>\n",
       "      <td>0.907153</td>\n",
       "      <td>0.831832</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>ML</td>\n",
       "      <td>0.47619</td>\n",
       "      <td>0.150888</td>\n",
       "      <td>0.386812</td>\n",
       "      <td>0.110385</td>\n",
       "      <td>0.092847</td>\n",
       "      <td>0.168168</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      type  Rack Count  Node Count  Job Energy  Job Failures  Job Submissions  \\\n",
       "0  Generic     0.52381    0.849112    0.613188      0.889615         0.907153   \n",
       "1       ML     0.47619    0.150888    0.386812      0.110385         0.092847   \n",
       "\n",
       "   Job Runtime  \n",
       "0     0.831832  \n",
       "1     0.168168  "
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pltdf = pd.DataFrame([{\n",
    "    \"type\": \"Generic\",\n",
    "    \"Rack Count\": df_prom_slurm.filter(F.col(\"gpu_node\") == 0).select(F.col(\"rack\")).distinct().count(),\n",
    "    \"Node Count\": df_prom_slurm.filter(F.col(\"gpu_node\") == 0).select(F.col(\"node\")).distinct().count(),\n",
    "    \"Job Energy\": df_prom_slurm_energy_sum.filter(F.col(\"gpu_node\") == 0).collect()[0][1],\n",
    "    \"Job Failures\": df_prom_slurm_state_count.filter(F.col(\"gpu_node\") == 0).filter(F.col(\"state\") == \"FAILED\").select(F.col(\"count\")).collect()[0][0],\n",
    "    \"Job Submissions\": df_prom_slurm.filter(F.col(\"gpu_node\") == 0).select(F.col(\"slurm_id\")).distinct().count(),\n",
    "    \"Job Runtime\": df_prom_slurm.filter(F.col(\"gpu_node\") == 0).groupBy(\"slurm_id\").agg(F.first(\"job_duration_seconds\").alias(\"job_duration_seconds\")).select(F.sum(\"job_duration_seconds\")).collect()[0][0],\n",
    "},\n",
    "{\n",
    "    \"type\": \"ML\",\n",
    "    \"Rack Count\": df_prom_slurm.filter(F.col(\"gpu_node\") == 1).select(F.col(\"rack\")).distinct().count(),\n",
    "    \"Node Count\": df_prom_slurm.filter(F.col(\"gpu_node\") == 1).select(F.col(\"node\")).distinct().count(),\n",
    "    \"Job Energy\": df_prom_slurm_energy_sum.filter(F.col(\"gpu_node\") == 1).collect()[0][1],\n",
    "    \"Job Failures\": df_prom_slurm_state_count.filter(F.col(\"gpu_node\") == 1).filter(F.col(\"state\") == \"FAILED\").select(F.col(\"count\")).collect()[0][0],\n",
    "    \"Job Submissions\": df_prom_slurm.filter(F.col(\"gpu_node\") == 1).select(F.col(\"slurm_id\")).distinct().count(),\n",
    "    \"Job Runtime\": df_prom_slurm.filter(F.col(\"gpu_node\") == 1).groupBy(\"slurm_id\").agg(F.first(\"job_duration_seconds\").alias(\"job_duration_seconds\")).select(F.sum(\"job_duration_seconds\")).collect()[0][0],\n",
    "}\n",
    "])\n",
    "\n",
    "# normalize\n",
    "for col in pltdf.columns:\n",
    "    if col != \"type\":\n",
    "        pltdf[col] = pltdf[col] / pltdf[col].sum()\n",
    "\n",
    "pltdf"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "051572a5-6edc-4596-8701-277ba734053d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>type</th>\n",
       "      <th>variable</th>\n",
       "      <th>value</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Generic</td>\n",
       "      <td>Rack Count</td>\n",
       "      <td>0.523810</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>ML</td>\n",
       "      <td>Rack Count</td>\n",
       "      <td>0.476190</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Generic</td>\n",
       "      <td>Node Count</td>\n",
       "      <td>0.849112</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>ML</td>\n",
       "      <td>Node Count</td>\n",
       "      <td>0.150888</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Generic</td>\n",
       "      <td>Job Energy</td>\n",
       "      <td>0.613188</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>ML</td>\n",
       "      <td>Job Energy</td>\n",
       "      <td>0.386812</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>Generic</td>\n",
       "      <td>Job Failures</td>\n",
       "      <td>0.889615</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>ML</td>\n",
       "      <td>Job Failures</td>\n",
       "      <td>0.110385</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>Generic</td>\n",
       "      <td>Job Submissions</td>\n",
       "      <td>0.907153</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>ML</td>\n",
       "      <td>Job Submissions</td>\n",
       "      <td>0.092847</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>Generic</td>\n",
       "      <td>Job Runtime</td>\n",
       "      <td>0.831832</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>ML</td>\n",
       "      <td>Job Runtime</td>\n",
       "      <td>0.168168</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       type         variable     value\n",
       "0   Generic       Rack Count  0.523810\n",
       "1        ML       Rack Count  0.476190\n",
       "2   Generic       Node Count  0.849112\n",
       "3        ML       Node Count  0.150888\n",
       "4   Generic       Job Energy  0.613188\n",
       "5        ML       Job Energy  0.386812\n",
       "6   Generic     Job Failures  0.889615\n",
       "7        ML     Job Failures  0.110385\n",
       "8   Generic  Job Submissions  0.907153\n",
       "9        ML  Job Submissions  0.092847\n",
       "10  Generic      Job Runtime  0.831832\n",
       "11       ML      Job Runtime  0.168168"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "melted_df = pltdf.melt(id_vars=[\"type\"])\n",
    "melted_df[\"type\"] = pd.Categorical(melted_df[\"type\"], [\"ML\", \"Generic\"], ordered=True)\n",
    "melted_df[\"variable\"] = pd.Categorical(melted_df[\"variable\"], reversed([\"Rack Count\", \"Node Count\", \"Job Submissions\", \"Job Failures\", \"Job Runtime\", \"Job Energy\"]), ordered=True)\n",
    "melted_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "50b8bf04",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\\begin{tabular}{llll}\n",
      "\\hline\n",
      " type    & Job Submissions   & Job Runtime   & Job Energy   \\\\\n",
      "\\hline\n",
      " Generic & 90.72\\%            & 83.18\\%        & 61.32\\%       \\\\\n",
      " ML      & 9.28\\%             & 16.82\\%        & 38.68\\%       \\\\\n",
      "\\hline\n",
      "\\end{tabular}\n"
     ]
    }
   ],
   "source": [
    "df = pltdf.drop(columns=[\"Rack Count\", \"Node Count\", \"Job Failures\"])\n",
    "\n",
    "df = df[[\"type\", \"Job Submissions\", \"Job Runtime\", \"Job Energy\"]]\n",
    "df.iloc[:, 1:] = (df.iloc[:, 1:] * 100).round(2)\n",
    "df = df.astype(str)\n",
    "df.iloc[:, 1:] = df.iloc[:, 1:].apply(lambda x: x + \"%\")\n",
    "\n",
    "\n",
    "latex_table = tabulate(df, headers='keys', tablefmt='latex', showindex=False)\n",
    "print(latex_table)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "9e8f1b9a",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/tmp/ipykernel_1765088/3627473518.py:12: FutureWarning: The default value of observed=False is deprecated and will change to observed=True in a future version of pandas. Specify observed=False to silence this warning and retain the current behavior\n",
      "  pivot_df = melted_df_2.pivot_table(index='variable', columns='type', values='value', fill_value=0)\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA1wAAAD9CAYAAACoTXYSAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjkuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8hTgPZAAAACXBIWXMAAA9hAAAPYQGoP6dpAABEk0lEQVR4nO3dT2zj2J0v+q8bwXiqb6eKLftuCqnCFBX0mxvMAwqUvO4BTG2z6EfK+wFMbrK6DxDh1Sw9JJDd25C1zcYm04u3JZ2ZXlsiehDgpV8jYiNdgx4MYsunHCS53kR3UTlsySL1x5asP/5+AKHbEnX4kyip+OM553e2+v1+H0RERERERDR3Hyw7ACIiIiIiok3FhIuIiIiIiGhBmHAREREREREtCBMuIiIiIiKiBWHCRUREREREtCBMuIiIiIiIiBaECRcREREREdGCMOEiIiIiIiJakB8sOwCiRfnLX/6C7777Dj/84Q+xtbW17HCIiIiIaEP0+3384Q9/wPPnz/HBB+P7sJhw0cb67rvv8OLFi2WHQUREREQb6u3bt/jRj340dhsmXLSxfvjDHwJ4/0V4+vTpkqMhIiIiok1xfX2NFy9e5Oeb4zDhoo0lhxE+ffqUCRcRERERzd0001ZYNIOIiIiIiGhBmHAREREREREtCBMuIiIiIiKiBeEcLtp4X375JT766KNlh0FEtNJubm6wvb297DCIiKa2u7uLly9fLjuMiZhw0cb79NNPlx0CEdHK+2BrC3/p95cdBhHR1D588gS/+eqrlU+6mHDRxvvJT3+Gp89/vOwwiIhW1sXXbfz2V79A8Nln+GR3d9nhEBFN9PXFBazPP8fFxQUTrmXyPA9xHCNJEgCArutQFGVoG1VVcXR0NHL/omRZBtM0kWUZ6vU64jie+z48z8PJyQmA968PAPb29tBqtfIY0jSFYRhz3/cq+nD3R0y4iIjG+OPv3wIAPtndxevnz5ccDRHRZtnoohmtVgtxHEPTNCiKgjiOEYbh0K3RaODVq1fwPO9BYlJVFZ1OB/V6fe5tR1GEjz/+GABwdnaGTqeTv07DMPJEz3Ec9Hq9ue9/3tI0XXYIRERERET3stEJl1SpVFCpVAof03UdYRjCcRwEQfBgMWmaNtf2giCAaZo4OztDq9Uq7MmTrzOKornue1F83192CERERERE9/IoEq5JdF0HALiuu+RI7ibLMti2jTAMJyZy6/QasyxbdghERERERPfChOuvFEVZ2xN80zShadpUc7JUVYVlWQ8Q1f2sSy8cEREREdE4TLgACCEghMh7utZJFEVI0xQHBwdTP6fRaCwwovvLsgyHh4fLDoOIiIiI6N42ukrhtBzHgaIopXOGhBAIgiCfF9XpdGDbdunwvSiKcH5+jp2dHVxeXqJarU7Vq5QkyVChCMMw8iqDZWSVw1mSRV3XC3vzoihClmVQFAVCCADIKxsC79+H/f19CCGgKAo6nU7+mOd58H0fvV4PYRjm8RRVZZRz5YQQOD8/h+u6+esMggBxHKNSqaDdbsM0zXwfg9sREREREa2DR5twCSHQbrfh+z4qlQq++eab0tLwx8fHQ3OfhBB49erVUGIhOY6DLMsQhuHQ9o7jTJw/ValU4DgODMOAbdtTJRftdhsAZkpEFEUZSRZN00Sj0RhKsLIsQ61WQxiGUFU1T7JkAjWo1WrBMAxUq9Wh+2VVRtM088S12Wzm73UURajVari6ugIAWJYFy7Jg2/bI+zjJzc0Nbm5u8r+vr6+nfi4RERER0SI8miGFvV4PQRDkt+PjYziOg729Pfi+X5pspWma9/xIiqLAsiw4jjO0bZIk8DwPb968Gbo/y7KRNoq02+28jPu0PVayJ+o+64gFQYAsy0Z64VRVhW3bsG176P69vb3CdsYlfaqqot1uo16vD8Wq6zqEEHMpAX98fIxnz57ltxcvXty7TSIiIiKi+3g0CVelUsl7TyzLguu66HQ6+ZA/mbjcpigKer3eSLJUrVZH7pO9U7eTn16vN3HdqyAIUK/XZ16MWO6rLP5pOI5TOges2WwiSZJ88ei72tnZgRBipGdNxj+PdcGOjo7w7t27/Pb27dt7t0lEREREdB+PJuEq47ouer3eSG+VpKoqrq6uhuYkpWk6NH9JStO0sJdH13VcXV2V9gDZto1ut3untbnkAsp3rbCYZVlhIiTJhGgePVCLnn+1vb2Np0+fDt2IiIiIiJbp0SdcwPuk5fT0tPRxOQfLtm0kSQJFUVCr1Ya2kQnPzs7O1PuViV6tVoPneXdKmmTFwVl6oOQQR/n/kyiKgvPz85ljK2qHiIiIiOgxYcL1V7I0/G1ZluHVq1eoVqvwfR+WZUFVVVQqlaHtZO/N5eXlTPs8OjrKhzkOVuSblqxkeHJyMvVzkiTJe+xk3OOGJAohVqI6INfmIiIiIqJ1w4QL3/e8FPX2mKZZuFjw7TlHSZJAVdWxPUa3kxpN0/J9+76PLMtKhzaO4/t+XtxjGt1uN9+vTKTK4pb3lxXKGHSfeWRFbr/H67owNRERERE9Xky4UJx0yCF6aZoWVgzsdDpDCUaWZXBdF1EUlfaUTRr2F4YhPM+beb6UrutwXRemaU58rud5I1UHXdctXYMsiiJomjZUzGNwna5B9y2sMWhcIRMiIiIionWx8QmXEAJZluW3IjIBkYsIA98XidB1fSSRyLIsnzslS5rLCoNlQwOjKBpKWoqGMOq6DsMw8jWrZtFqtRCGIfb39+F53sjjch6apmkjwwNbrRY0TRvpXUvTFL7vj6yF1Ww282Ibg+3LHqmi9/ny8nLisMVBcnFmeX+apncqKkJEREREtExb/X6/v+wgFiUIgpFqgoqiFC5AnCQJXNdFo9GAoijQdT1PTGzbRq/Xy5MsVVWh6zo8z0McxzBNc2jIYRRFiOMY1WoVqqqi1+vlj8thgzKJ03UdR0dH0DQtX2hYCJGv9TVpseTb5OLCck6XfA2qquLo6Ghs4YogCNDtdvPCH5eXl6XPSZIEvu9jb28vf9yyLGxtbeXvXxiGEELg8PAQSZJACAHDMHBwcADDMPL3Tw7H1DRtKLmT76MsUHJ7WOck19fXePbsGer/9C+o/N0/zPRcIqLH5D///V/x61/+HP9mWXj9/PmywyEimujL777DP/71XH8ZF+Xleea7d+8mVsbe6ISLHjcmXERE02HCRUTrZp0Srh88UExES/Oni//AD/7mb5cdBhHRyvrz1X8BAL6+uFhyJERE01mn3yv2cNHGklceiIhosg+2tvAXnhIQ0Rr58MkT/Oarr/Dy5csH3zd7uIgGfPHFF/joo4+WHQYR0Uq7ubnB9vb2ssMgIpra7u7uUpKtWTHhoo33+vXriVceiIiIiIgWYePLwhMRERERES0Le7ho43355ZccUkhEREQLweG4y8MhhUQr4tNPP112CERERLShWHBmeZZZNGMWTLho4/3kpz/D0+c/XnYYREREtGEuvm7jt7/6BYLPPsMnu7vLDudR+friAtbnn+Pi4oIJ1zJ4noc4jpEkCQCULogWBAFc10WWZVAUBbquIwzDucbiOA6iKEKv18PZ2dmDLMzmeR5OTk4AAKqqAgD29vbQarUAAFmWIU1TGIax8FhWwYe7P2LCRURERHP3x9+/BQB8srvLRcOp1EYWzWi1WojjGJZlQVEUHB4eFm5nWRa63S40TUOn05l7sgUArusiDEMIIebe9m1RFOHjjz8GAJydneWvKQxDGIYB0zSRZRkcx0Gv11t4PPeVpumyQyAiIiIiupeNTLikarWKN2/eIE1TBEFQup2u63lP0CIssm0pCAKYpomzszO0Wi0oijISQxiGeY/bOvB9f9khEBERERHdy0YnXABgGAY0TYPjOKW9TDs7OwuN4XbyM29ZlsG2bYRhOHHIouu6C41lnrIsW3YIRERERET3svEJF4B8SF/Z0MJ1Z5omNE2bak6WqqqwLOsBorqfdemFIyIiIiIa51EkXKqqotVqIYqivJDGpoiiCGma4uDgYOrnNBqNBUZ0f1mWbWxyTERERESPy0ZWKSziui6CIIBt2+h2u1M/L4qivIqhHJIoq/0VSZIEcRyjWq3m9zWbzbH78DwvH3bY7XZRrVan7oWK4xjA+3lo09J1vXC43qTXKoTA/v4+hBBQFAWdTmfoNfi+j16vhzAM83iyLMuLddTrdcRxnM+nE0Lg/Pwcruvm89yCIEAcx6hUKmi32zBNM9/H4HZEREREROvg0SRcwPuhhY1GA57njU2aJNM00Wg0hrbNsgy1Wg1hGI6c/Mt5YoPFHoQQcByndB+1Wg1HR0dDwwFN00S3251qvlW73QYwW2EORVFG5npN81plkiUTqEGtVguGYQwlmjIu+RwhBIIgQLPZzBPMKIpQq9VwdXUF4H3lSMuyYNs2siybqXLkzc0Nbm5u8r+vr6+nfi4RERER0SI8iiGFkq7r0HUdjuNMLMgQBAGyLBvpaVJVFbZtw7btofuTJEEQBCOV9RRFGdlW8jwPAEbmXrmuC8/zpioaIXui7lOYY9bXure3V9jOuKRPVVW0223U6/WhWHVdhxBiLiXgj4+P8ezZs/z24sWLe7dJRERERHQfjyrhAr4vNV6WBEmO45TOi2o2m0iSZGg+mOM4pcP6yhKRsn3I3qRp5pvJ5OU+63zN+lrvYmdnB0KIkZ41Gf881gU7OjrCu3fv8tvbt2/v3SYRERER0X08uoRLVVW4rjs2iciyrDA5kGSSMNgrk6Zpac9P2T6A93O2oigaucnHJqnX60Ptzeour/WuFj3/ant7G0+fPh26EREREREt06OawyW1Wi34vg/TNPO5Q4OmSV4URcH5+fnQ9rMM65PPMU2zsGdsmhLvwPuKg0EQIEmSiWtwDe47TVMYhjHza72PRa9HRkRERES0ah5dD5fk+35pQQvZEzNumJ4QIt9umu3vso9pGIYBVVVxcnIy9XOSJMmTvFlf6zJxbS4iIiIiWjePNuHSdR2GYcDzvJHeG5lclPX+yPsHhxCqqjpTuXk5T2tcz9G0yZjv+0jTdOqEpNvt5r1Nd3mtZe6bPN52e17XXYdNEhEREREty0YnXJeXl2Mff/PmDQAUzuVyXXek4qAURRE0TRsa9ue6Lk5PTwu3l6Xbi/Yh16S6LU3T0ufdpus6XNeFaZoT51p5njdSMGTW1zq4TtegeS4qXa1W557AERERERE9tI1OuMqSGUlRlNJEo9VqQdO0kSGHaZrC9/2R9aEMw0Cz2RzZXgiRb3u7x8ayLOi6PpIACSGGhv1No9VqIQxD7O/v5+Xmb7fpOA40TRsZHjjra202m3mxjcH25esr6om6vLycOGxxkFycWd6fpunUc9SIiIiIiFbFVr/f7y87iHlzHAdRFCHLMqiqCsMwxi4i3Gg0EMdx4WNBEKDb7WJnZwfA+8Th6OiotABEFEU4Pz8fWgC42Wzi448/hqIoaDabI0leEATodDqoVqtQVRW9Xm9kTaxpycWF5ZyuwXlm4+Ke9bUmSQLf97G3t5c/blkWtra2oCgKdF1HGIYQQuDw8BBJkkAIAcMwcHBwkA/njOMYSZJAVVVomjaU3EVRhDiOUavV8vZncX19jWfPnqH+T/+Cyt/9w0zPJSIiIprkP//9X/HrX/4c/2ZZeP38+bLDeVS+/O47/ONfz6GXcVFenme+e/duYmXsjUy4iIDvvwg/+enP8PT5j5cdDhEREW2Yi6/b+O2vfoHgs8/wye7ussN5VL6+uID1+edMuIiWSX4RiIiIiBblg60t/IWn00vx4ZMn+M1XX+Hly5cPvu9ZEq5HuQ4XPS5ffPEFPvroo2WHQURERBvo5uYG29vbyw7jUdrd3V1KsjUrJly08V6/fj3xygMRERER0SJsdJVCIiIiIiKiZWIPF228L7/8kkMKiYiI6FF4TEMcOaSQaEV8+umnyw6BiIiI6EE8piIeyyyaMQsmXLTxWBaeiIiIHoPHVKZeloW/uLjYzIQryzI4jgMhBJIkAQBomgbXdaHr+sxtmaaJLMtQr9dLFyC+iyAIEMcxKpUKAEBRFNi2DVVV4XkeWq3WndteZNw0Xx/u/ogJFxEREW28P/7+LQDgk91dLsS8Qu6UcKmqijAMAQC1Wg1ZlqHT6dwpAFVV0el00Gg07vT8MqZpYm9vL49Tsm0btVoNvu/fK+FaVNyzqlar0HUdvu+vZftERERERJvs3lUKVVWdRxxzXSE6iiIIIQoTKt/3R5Kw+1jGyta3KYqy1u0TEREREW2qjZzDdXJyAtM0Sx/3fX/pPVPz0u1217p9IiIiIqJNtpHrcKVpCiFE6eOqqs6tZ46IiIiIiKjMRiZcmqZNnHNk2/YDRUNERERERI/VwocURlGELMugKEre6zSuWIUQAqenp/nfnU4HjuPM1CN1dHSEWq2GWq2GMAwLn2sYxtA+9/f3IYSAoihDBUA8z4Pv++j1egjDsLQK4zRxF1U2DIIAwPuhe2mawnVdaJqGNE3RbrfztqrV6sj71mg08vd2MOYsyxBFEVRVRa/XgxACqqri/PwcrutOvU1Z+9KkYzvu9Qoh8n3dPj7TxEZEREREtA4WmnCZpolGozFyEl6WCGVZhtPTU1iWNXR/rVabqeS8LFHvOA6q1SpUVYWu62g0GkOJliQTCpkcDGq1WjAMA9VqtXR/08YtKxuapgkhBIIgQLPZzItSRFGE/f19nJ2dodfrDbVXrVahKMrQfXEcw7btPDED3icyjuOMFAYJgiB/bdNsU9a+NM2xneb11mo1XF1dzRR/mZubG9zc3OR/X19fj92eiIiIiGjRFjakUJ4g305CVFWFbdulQ/pubw8AruuOLYJRpNVqodvt5glBEAQwTRNbW1twHKfwOXt7e4X3T9O7Nkvcqqqi3W6jXq8PVQDUdR1CCPi+P5Jc6rpeWF3xdiKYJEm+7tjt+OT902xT1j4w+7Gd9HrTNJ0p/jLHx8d49uxZfnvx4sXY7YmIiIiIFm1hCZfjODg4OCh8rNlsIkmSfNFkqaz8uDwx9zxvphhUVYXruuh2u7i6usqHBHqeN9c5XLPGvbOzAyHESEl52U5RkqMoCnq93sRYVFXF6enpUBIjyeRvmm3GmfXYTnq9g6/rPrEdHR3h3bt3+e3t27cTXwsRERER0SItJOHKsqzwBFuSJ9pFJ9Vl5Byeu1IUBYZhII5jtFqtqYaozUNZ3ON6ze6ztpemadB1HbVaDdVqFbZtI4oiAMh7zabZpsxdj+20c/DuE9v29jaePn06dCMiIiIiWqaFJVyTKIoyUwKlKMrUCZI8QS/jui4URRnpYVuEsrgXuZhwGIaI4xi6riNJEpimiWq1OpQETbNNkbse21le711jIyIiIiJaNQtJuGRvxri1sGTluWnNsv3JycnEber1+tj45mXW13lfMiHSdR2+7+fDKTVNw/7+/tTblFnEsZ01fiIiIiKidTF1wuV5Hra2tkZ6GYpOvOXJdllviLy/rEhF2XOm3X7SwsfA+3lD0w7du09iNkvc81A2Ny4MQ1QqFWRZNtU2ZRZxbGeNn4iIiIhoXUydcHW7XSiKMjI0rN1uo9lsjmzvum7p4sNRFEHTtJES7WWJTRRFUBRl7Ppdt40rsCDnId2eEzS4ntSgSUMP5xn3PBRVMwSG54ZNs02ZuxzbWdwnNiIiIiKiVTJ1wuU4zsgitXIR26LFaFutFjRNGynBnqYpfN8vPKk2DGNk/pUQAsfHxzg7O5s2VADvEy7btkd6RORivEUJQ7PZzJOxwf3LKnplvSuzxn15eTlxSN4sbm9fVOVPbiOP3zTblLU/67Gd9fXOEhsRERER0SqbeuFjVVVRr9fheV7eE9TtdvHNN9+UFkQIwxBBEMBxHOzs7AB4f/Ld6XRGnlOtVvOeMpnIAUCn08HZ2dlMRRcMw4BlWXnSM5g0ybiKTtwVRUEcxzg8PMTe3l6+T8uyYNs2HMdBHMdDCcUscQshcHh4iCRJIISAaZo4ODiAYRjwPA9xHAN4n9yenJwgDENkWQbHcYaeY9s26vX6UFuNRgOu66JSqeDs7AxZliHLsvx1CyHyuKfZ5nassn3ZyzTNsb3L650mNiIiIiKidbHV7/f7yw6CaBGur6/x7Nkz/OSnP8PT5z9edjhEREREC3XxdRu//dUvEHz2GT7Z3V12OAv19cUFrM8/R6fTWcq0E3me+e7du4lLETHhoo0lvwhEREREj8UHW1v4yyM5vf/wyRP85quv8PLlywff9ywJ19RDConW1RdffIGPPvpo2WEQERERLdzNzQ22t7eXHcaD2N3dXUqyNSsmXLTxXr9+PfHKAxERERHRIixk4WMiIiIiIiJiDxc9Al9++SWHFBIREdHaekzDBGfBIYVEK+LTTz9ddghEREREd/aYCmHMYplFM2bBhIs2HsvCExER0bp6TKXeZyHLwl9cXDDhWja5cLAQAkmSAAA0TYPrutB1fea2TNNElmWo1+v5wr33IRcAbrfbAIB6vT6yyLNcBBh4vwiz67r32qdchDjLMiiKgk6nc6/2Vt2Huz9iwkVERERr6Y+/fwsA+GR3F6+fP19yNHQXG59wqaqKMAwBALVaDVmW3TnBUFUVnU4HjUZjbvG1Wi20Wq08trIkbjDZm1W1WoWu6/B9HwCgKAriOL5ze0RERERENJ1HVaVQVdW5tLOI1awrlcrYx1VVxdnZGYQQd2r/dq8ZAOzt7d2pLSIiIiIims7G93BtEkVR7pQ0drvdBURDRERERESTPKoerk1QrVaXHQIREREREU2JCdeKS9N0aBjhIoYzEhERERHRYnBI4YAoivLKfTLJabVapdsLIXB6epr/3el04DjO3OaKAUC73R5qT1ZWFEIgCIJ8blan04Ft2yMJWaPRmLoaoRAC+/v7EEKMbO95HnzfR6/XQxiGeRyDxTyazSZc10UQBACAOI5HioB4npfH3O12Ua1WYVlW/niWZYiiCKqqotfrQQgBVVVxfn5+7+qMREREREQPjQnXX5mmiUajMZRgZVmGWq2GMAxHkqgsy3B6ejqULADvKyHepeR8mTAM0Ww2R+4/Pj4eSkCEEHj16tVQMgS8T3ps287Lzo8jk6yi6oWtVguGYYwMaRys3Njr9RAEAVqtFpIkgeM4yLIsf+9qtRqOjo5gGEb+fNM00e124bouhBBwHCevKikFQcBqikRERES0ljikEN+f0N9OnlRVhW3bsG278Hm3twcA13Vhmuad4pDrY8lbtVrN1w4blKZp3hsnKYoCy7LgOM7I9rPO+yqrXjiu507TNKRpmid7uq7j6uoqf47neQAwlGwB798vz/OQZRmSJCms1mhZ1sQqjgBwc3OD6+vroRsRERER0TIx4QLgOA4ODg4KH2s2m0iSZCTxKSqzDrxPNIQQeYIxC0VREIbh0K0oyVEUBb1eb6TXp1qtLrUnqNfrDQ1pHHyPyt5jVVWhKAqSJIGqqjg9PUWapiPbTZPEHh8f49mzZ/ntxYsXd3shRERERERz8uiHFGZZBiFEaTEKmTQM9t5MIucc3ZemaYW9a6qq4urqKv9bvoa7Lug8L2U9YDIJ7Ha7iKKocJtutwvLsqDrOmq1GlRVha7raDQaMAxjqvf+6OgI//N//s/87+vrayZdRERERLRUTLim6BFSFGWmBEpRlLn1NJUlMUIIHB8fQwiBWq2WJyqDRTweWtmwP/lemKZZmDgNDjMMwxBJkuT/DYIAqqoiDMOJFRq3t7exvb19j1dARERERDRfjz7hkgnNYOn122SlvGmN6zGb1e05T8D3xTxc1x2aRzbNPKdlmOY9BpAX2NB1faga4+HhIfb394d69YiIiIiI1sFGzuHyPA9bW1sjc4GKTvhlMlDWIyXvLyskUfacWbaflWmaUFV1pGhHr9cb+ruo4MZ9TEqYysh5WuN6CYUQpXPlwjBEpVJhpUIiIiIiWjsbmXB1u10oijJS2KLdbheWWHddF77vF7YVRRE0TRvpaSpLPqIogqIoY9fvuq+y+WSdTmcorrsmKIPrkA26TwI3uD7XbWma5mXrb5eEl7jgMxERERGto41MuBzHgeu6Q8MA5cl+0eK5rVYLmqaNlFRP0xS+7xcmAYZhjBSAkPOqzs7OZo5ZLvI7DV3XR5KfLMvQaDTyONI0Rb1eH4qtSNH9zWYzL8QxuJ3sQStK5AYfLyILYtwuAiJ7tmQCWVSlUMYxzwWliYiIiIgewkbO4VJVFfV6HZ7n5b013W4X33zzTWk59zAMEQQBHMfBzs4OAODy8hKdTmfkOdVqNe8pG+y16XQ6ODs7K91HEc/zcH5+nicZskLf0dFRaa+OXMxYLtYsX7NhGPk6YKZpotlswjRNJEkCIQQajUaeiB4eHub3m6YJ27bzpEdRFMRxjMPDQ+zt7eWvx7Is2LYNx3EQxzHCMESWZXAcZ6gtVVULE1v5Htu2jWq1ClVV0ev18t7ASqWCs7MzZFmGLMvyBE4IUdrzRURERES0yrb6/X5/2UEQLcL19TWePXuG+j/9Cyp/9w/LDoeIiIhoZv/57/+KX//y5/g3y8Lr58+XHc7K+PK77/CPQYBOp7OUqSfyPPPdu3d4+vTp2G03soeLaNCfLv4DP/ibv112GEREREQz+/PVfwEAvr64WHIkq2Wd3g/2cNHGklceiIiIiNbZB1tb+AtP2Ud8+OQJfvPVV3j58uWD75s9XEQDvvjiC3z00UfLDoOIiIjoTm5ubrC9vb3sMFbO7u7uUpKtWTHhoo33+vXriVceiIiIiIgWYSPLwhMREREREa0C9nDRxvvyyy85pJCIiIimxiF864FDColWxKeffrrsEIiIiGiNsEjFelhm0YxZMOGijfeTn/4MT5//eNlhEBER0Rq4+LqN3/7qFwg++wyf7O4uOxwq8fXFBazPP8fFxQUTrkXLsgyO40AIgSRJAACapsF1Xei6PnNbpmkiyzLU63XEcXzv+DzPQxzHeWy6rkNRlKFtVFXF0dHRyP00Hx/u/ogJFxEREU3lj79/CwD4ZHeXCw3TXKx9wqWqKsIwBADUajVkWYZOp3PntjqdDhqNxtzia7VaaLVaeWxFSVySJHj16hWOjo7QarXmtu+7SNN07Grd1WoVuq7D9/0HjIqIiIiIaD1tVJVCVVXn0s64hOOuKpUKKpVK4WO6riMMQziOgyAI5r7vWUyTSLEnjoiIiIhoOmvfw7Up5PBH13VhWdbS4siybOzj3W73gSIhIiIiIlp/G9XDte4URZmY8CxSFEVL2zcRERER0SZiwrUihBAQQsxc6GNesizD4eHhUvZNRERERLSpHtWQwiiKkGUZFEWBEAIAxhapEELg9PQ0/7vT6cBxnLnNFRvkOA4URRmZQyWEwP7+PoQQUBRlqCCI53nwfR+9Xg9hGObJWlG1RTk3TAiB8/NzuK6bv44gCBDHMSqVCtrtNkzTzPcxuF2j0cjfPxnHuH11u12kaQrXdaFpGtI0Rbvdzt/LarVa+v57npfPFet2u6hWq0sdaklEREREdBePJuEyTRONRmPoBD/LMtRqNYRhOJJEZVmG09PTkZP8Wq12p5LzRYQQaLfb8H0flUoF33zzzUhBCpncyKRmUKvVgmEYqFarQ/fLaoumaUIIgSAI0Gw287ajKEKtVsPV1RUAwLIsWJYF27aRZVle9fG2OI5h23aeNE27r/39fZydnaHX6w29n9VqFYqiFL7HR0dHMAwjv880TXS7XbiuW/p+3tzc4ObmJv/7+vq6dFsiIiIioofwKIYUBkGALMtGTuxVVYVt27Btu/B5RT0qrusO9QDNotfrIQiC/HZ8fAzHcbC3twff98dW/9vb2yu8f1xvm6qqaLfbqNfrQ23rug4hBNI0nfk13E7upt2X7/sjSaqszjjI8zwAGEq2gPfvu+d5Y+e4HR8f49mzZ/ntxYsXs7w0IiIiIqK5exQJl+M4ODg4KHys2WwiSZJ8YWKpLPmRCYRMDGZRqVTy3iTLsuC6LjqdDi4vL1GtVvNhjvOys7MDIcRImXv52nq93oPtqyhRUxRlJIayY6WqKhRFGTlOg46OjvDu3bv89vbt2zu8EiIiIiKi+dn4IYVZlhUmApJMCNI0nXqYoKqqOD8/n1eIcF0XQRDAcZy5Lyi8iPlmd9nXNGubyd6rbrdbWjFxXFn67e1tbG9vT9wPEREREdFDeRQJ1ySKosyUQC2ifHu9Xsfp6encE66HXKT4vvuS76lpmoXJ7+1hhkREREREq27jhxTKXpdxw/WEEDP1BM26/SztzntY4X095Npc0xwrIiIiIqJ1spYJl+d52NraGin6UHSiLk/iy3qk5P1lRSnKnjPL9tOQvUOz9pzNOzm5PafqIRdilvO0xvU2MhkjIiIionWylglXt9uFoigjQ9ja7TaazebI9q7rlg7Vi6IImqaNDFcrO7GPogiKooxdv+suihLDwQIRg2uHDRpXRGJWiyjcMSs5n63I4DpeRERERETrYC0TLsdxhhbkBZCfpBet09RqtaBpGhzHGbo/TVP4vl+47pRhGCPD6YQQOD4+xtnZ2UzxCiGQZVl+KyJL08dxPBSf1Gw28wIgg+3KHqmidi8vLycOpRyk6/rQPtI0LSx2UdTmrPsqY1kWdF0fKdUvhECSJHNZ/4yIiIiI6KFs9fv9/rKDuIs0TZEkSd7zIxfFHVe4IQgCdLtd7OzsAHifJBwdHY08Ry7eCwCnp6f5/Z1OZ+I+ivbZ6XSG7lMUpTAxTJIEruui0WhAURTouj6UVCZJAt/3sbe3l8dgWRa2trby7cMwhBACh4eHSJIEQggYhoGDgwMYhgHP8xDHMZIkgaqq0DRtKOGMoghxHKNWq+XtAxhpU9f1POmddV9ZlsFxnKHn2LY9lEzJ961arUJV1ZFFk6dxfX2NZ8+eof5P/4LK3/3DTM8lIiKix+k///1f8etf/hz/Zll4/fz5ssOhEl9+9x3+8a/ni9NUw543eZ757t07PH36dOy2a5twEU0ivwg/+enP8PT5j5cdDhEREa2Bi6/b+O2vfoHgs8/wye7ussOhEl9fXMD6/HMmXETLJL8IRERERLP4YGsLf+Ep8sr78MkT/Oarr/Dy5csH3/csCdfGr8NF9MUXX+Cjjz5adhhERES0Jm5ubrC9vb3sMGiC3d3dpSRbs2LCRRvv9evXE688EBEREREtwlpWKSQiIiIiIloH7OGijffll19ySCEREXGYGNGG4ZBCohXx6aefLjsEIiJaASyEQLRZllk0YxZMuGjjsSw8ERGx1DfRZpFl4S8uLphwrTK5AK8QAkmSAAA0TYPrukOL8E7blmmayLIM9XodcRzfOz65cHC73QYA1Ov1kUWXhRD545VKBd1u99773TQf7v6ICRcR0SP3x9+/BQB8srvLxWyJ6EE96oRLVVWEYQgAqNVqyLIMnU7nzm11Oh00Go25xddqtdBqtfLYxiVxtm0jCIK57ZuIiIiIiO6PVQr/SlXVubSziJWuK5XKxG1834emaRBCzH3/RERERER0N0y4Noht28iybNlhEBERERHRXzHh2iD1eh29Xm/ZYRARERER0V8x4VpjaZoiTdP870UMZyQiIiIiort71EUzZhFFEbIsg6Io+TypVqtVur0QAqenp/nfnU4HjuPMba4Y8L4y4u32ZHXFoqqJsqiGEALn5+dwXbc0Hs/z8oqI3W4X1WoVlmWNtN1sNuG6bt52HMcjxT2CIIAQApeXl3k7cvswDKFp2tD7++bNGxiGMRSL4zhQFAWWZcF13Xu+c0RERERED4MJ1xRM00Sj0RhKsLIsQ61WQxiGI0lLlmU4PT3NExSpVqvdqeR8kTRN4ThOXmXxNlk10TRNCCEQBAGazWaeREVRhFqthqurq5Hn1mo1HB0dDSU9pmmi2+3mSZqsyNjr9RAEAVqtFpIkgeM4Q4mgaZrY29vL37sgCPDxxx8P7deyLBwdHeHjjz8e2S/wPrGN4xi+749NWG9ubnBzc5P/fX19PeFdJCIiIiJaLA4pnCAIAmRZNpI8qaoK27Zh23bh825vDwCu68I0zTvFIYSAaZp58id7mCZRVRXtdntkDS9d1yGEGBqSCLzvTQIwkvS4rgvP84b2qWka0jTNE0hd13F1dZUnRUmSIIqioUTVsiwIIRBFUf43gLz36uTkpPB1NBqNib2Dx8fHePbsWX578eLF2O2JiIiIiBaNCdcEjuPg4OCg8LFms4kkSfJFk6XbixNLMsmRSc0sFEVBGIYIwxBxHKPb7RYmdbft7OxACDEyv0vGeLvIRtnrVVUViqKMvNZerzfU9uBrj+O4MEnSNK0wsbJte2ReGvA+cbudABY5OjrCu3fv8tvbt28nPoeIiIiIaJE4pHCMLMsKkxVJJheDvTyTqKqK8/PzucRX1FsmhBhJ+KadNyZ7r7rdbt4DdVu325267Wq1Wlg1UQhRmohpmgbf9+H7fn7/tO/v9vY2tre3J25HRERERPRQmHCNMc2QPUVRZkqgFEWZ21pZ9Xp95D45n+r2Pqch4zJNszDBKeplGrcos+zRG5zTJf8u6zWUwzRlwlWWnBERERERrQMOKRxjMEkoM2tCMM8EQlGUkWTq8vLyzu1N83pnbc+yrHxB5izLcHh4CN/3S3sN5TBJWcUwCIKphhMSEREREa2iR5dweZ6Hra2tkXlCRUmGTEDKeqTk/Xt7e1PvP8uymbaf1X2SJTlPa1yP3SztCyHQaDQQhmE+N+vNmzcT555ZljU0pJCIiIiIaF09uoSr2+0W9gy12200m82R7V3XLT35j6IImqaN9MCUJSVRFEFRlLHrd91HEARTDx8sM7im1m1pmqLdbk/dVrvdRq/Xg6IoMAwDhmFMFZ8snuF5Hnu3iIiIiGitPbqEy3GckQV/ZYJRtKBuq9WCpmlwHGfo/jRN4ft+4TpYhmGMFJ0QQuD4+BhnZ2czx1xUeOK2IAhg2zaq1erQ/ZeXlxOHRA6yLAu6ro+UuxdCIEmSobldQoixsdXrdbiuizRNIYSYundMFs84OTnh/C0iIiIiWmuPrmiGqqqo1+vwPA+KokAIgW63i2+++aa09yUMQwRBAMdxsLOzA+B9ItPpdEaeU61W856ywZ6iTqeDs7OzmXqgPM9DHMf58MdGozHyfCEE2u12nszIQhpCCBweHiJJknwNr4ODAxiGkbcLvE9AT05OhhJH+XplAqeqKnq9Xt4zl2UZHMcZaltV1ZGEVVEU2LaNWq02cr9MxsrmcpWtb0ZEREREtE62+v1+f9lB0GZyHAfVanVkzlaWZUiSBI7j4M2bN4XDBj3Pu/fQy+vrazx79gz1f/oXVP7uH+7VFhERrbf//Pd/xa9/+XP8m2Xh9fPnyw6HiO7py+++wz8GATqdTukF/EWS55nv3r3D06dPx2776Hq46GFEUYQsywqHacrqhZVKBcfHxwufp/Wni//AD/7mbxe6DyIiWm1/vvovAMDXFxdLjoSI5mGdvstMuGgher3e2DW6gPdDCyuVCtI0RZZleeKVJMlck7D/7//9f+bWFhERra8PtrZgff75ssMgojn58MkT7O7uLjuMiZhw0UJYloVGo4EoigqTJ9n7JatAttvtfLssywoXXr6rL774Ah999NHc2iMiovV0c3OD7e3tZYdBRHOyu7uLly9fLjuMiTiHixYqiiKcn59jZ2cnL/ghC3xYlgVFUfI5XfKxeZXNn2VsLRERERHRtGY5z2TCRRuLCRcRERERLcIs55mPbh0uIiIiIiKih8KEi4iIiIiIaEFYNIM2lhwte319veRIiIiIiGiTyPPLaWZnMeGijXV5eQkAePHixZIjISIiIqJN9Ic//AHPnj0buw0TLtpYch2wb7/9duIXgTbL9fU1Xrx4gbdv37JgyiPDY/948dg/Tjzuj9eyj32/38cf/vAHPH/+fOK2TLhoY33wwfspis+ePeOP8CP19OlTHvtHisf+8eKxf5x43B+vZR77aS/os2gGERERERHRgjDhIiIiIiIiWhAmXLSxtre38c///M/Y3t5edij0wHjsHy8e+8eLx/5x4nF/vNbp2G/1p6llSERERERERDNjDxcREREREdGCMOEiIiIiIiJaECZcREREREREC8J1uGitJEmCOI5RrVYhhAAAtFqtpbVDD2dexywIAnS7XaRpil6vB13X4brunKOleVrU99XzPGiaBl3X790WLcY8j32WZXAcBwBQqVSgKAq/+ytsXsc+iiLEcTx0n+u6UBRlDlHSPAkh4DjOvb6bK3t+1ydaE2EY9g3DGLovjuO+rutLaYcezryOWavV6ne73fzvq6urvq7rfUVR+ldXV/MIleZsUd/Xq6urPoB+GIb3aocWZ57HPgzDvqZpI9//Vqt17zhp/ub5mx/H8dB93W63r2kaf/NXSKvV6huG0Xddt6+qat+yrDu1s8rnd0y4aC1cXV2VnhRrmtb3ff9B26GHM69jFoZhv9PpFLYPYCV+kGnYIr+vrusy4Vph8zz2nU6nsC3DMPqqqt4zUpq3eR37TqdTmlB3Op2RE3NaDZqm3SnhWvXzO87horVwenoKVVULhwAcHBzA9/0HbYcezryO2fn5OTRNG7lfURRYloUkSfLhB7QaFvV9TZKEwwhX3DyPveM4ODo6Gmmr0WjAtu17RkrzNq9jnyQJ9vb2Ch/TNA1pmt4nTFoxq35+x4SL1kIYhqhUKoWPqaqKNE2nOlmeVzv0cOZ1zIIgQKPRKHysVqsBANrt9p3jpPlb1Pc1TdPC5JtWx7yOfZqmSJIElmWNPGZZ1mrM7aAh8/zel51kZ1kGVVXvGiKtoFU/v2PCRWuh3W6X/jjK+7Mse7B26OHM65jV6/XSx+SPcNmPNS3HIr6vQRDwJHsNzOvY+75fetWbVtO8jr1hGEiSBKZpjpxou66bF1ChzbDq53dMuGgtCCEm/oM5zRdpXu3Qw5nXMYvjeKRSldTtdgGAvR4rZt7f1yzLmFSviXkd+yRJ8pMtz/PgeR6CIIDjOBzNsKLmdexVVYXruoiiCK9evUKSJAC+H+3AYcWbZdXP71gWntae/IL1er2VaIcezryOWRAEhUOOaHXd5dhHUcTerQ0wy7HPsgyapsHzvKFjn2UZarUaOp0Oe7/WyKzf+1arBVVVYZomGo0GVFVFHMccTvjIrML5HXu4iOhRcxwnvxJKmyuKIhiGsewwaAnSNB059qqqQtd1HB4eLikqeiiKoqDVakHXdWRZhkajwYIZ9OCYcNHam9f8G87jWT/3PWZpmiIIAsRxzKvca2aWYy+EQK/X41XtDXGX733Rsa/VaoiiiEML18isx14OHXVdF3Ecw/f9vHdTDjGkzbcK53dMuIjo0TJNE2dnZzwR33AcMvq4TZpIz+qkmykIAgAY6t20LAvdbjcfZshkmx4KEy5aC6qqlk52lGNypzlpnlc79HAWdcwajQZ832ehjBU2j2PPEvDraZ6/+as8kZ5GzevYu65bOFRcVVV0Oh0AYC/XBln18zsWzaC1oGla6ZUo+QWb5qRqXu3Qw1nEMbNtG47jsErVipvHsc+yDCcnJyPr8ch2j4+PcXJygkqlsvSFMel78/zNn5RQjVsygh7ePI79pIp1iqLg6OiIyfYGWfXzO/Zw0VpoNBqlP4zdbnfqE+d5tUMPZ97HzPM8mKY58rwsy3i1c8XM49gbhoEwDAtvAHB0dIQwDJlsrZh5fe8PDg5KCyTI9jmqYbXM49grijLVcEEe+82x6ud3TLhoLTSbTfR6vcIvUxRFsG176D4hROHJ86zt0PLN69jL7TVNK/zhTdOU//iumHkee1ov8zr2hmFAURREUTTyWBiGsCyLBXNWzLyOvaqqY38P4jhe+kk4zW5tz+/6RGsiDMO+YRgj9+m6PrKtrut9AH3f9+/VDq2GeRz7TqfT13W97/v+0M113b7run1N0xb6Guhu5vW9v63T6Uy9LS3HvI59HMd9VVX7V1dX+X2+74/cR6tjHsf+6uqqr2laP47jkfstyxq5n1aDqqojx37Qup7fcQ4XrQ15pdJxHFSr1Xy4QBzHI9s2Gg202+3CsfmztEOrYR7Hfn9/f2wPCHu3VtO8vvdSmqY4Pj7Or4I6joM4jtFoNFjJcMXM69jrug7f93F4eIhKpZIvEdDtdhf9EuiO5nHsFUVBp9OB4zj5EGLJdV32bK4Qz/Nwfn6OLMvyW6PRgKIoODg4GKo0ua7nd1v9fr+/7CCIiIiIiIg2EedwERERERERLQgTLiIiIiIiogVhwkVERERERLQgTLiIiIiIiIgWhAkXERERERHRgjDhIiIiIiIiWhAmXERERERERAvChIuIiIiIiGhBmHAREREREREtCBMuIiKiDZBlGaIoghBi2aEQEdGAHyw7gEm+/fZbXFxcLDuMe9vd3cXLly+XHQatOX4faF48z8PJyQmyLMvvq9frI9tlWYZer5f/fXV19SDx0Ww8z8Pl5SWOjo5gmiZ834eqqmOf02g00G63AQCVSiXfvtfr5Z+LsvsB4JtvvoGiKAt4NXfTaDTyz+vZ2Rk0TZv4HMdxkCQJ0jQFACiKAlVV0el0xj7Ptm0EQZA/p9lswvf9hcdLRGuqv8J+97vf9f/2yYd9AGt/+9snH/Z/97vfLfT9cl23r2laX9O0vmEYfcMw+q7r5o93u91+GIYLjYEW53e/+13/wydPlv5ZnsftwydP5vp9cF23r+t63n6n05nqeVdXV31FUfoA+rqu91ut1lCbhmHkj8vv1bRtr4tWq9UHMPRbcVu32+1rmtYH0L+6unq44ObEMIy+qqrLDmNh4jjua5rW7/e//0wPfpbLqKra13W93+12Rx6Tx7voMfmZKXps2VzXnek3QDIMow+g7/v+TPvSNO1e34m7xktE62Wle7guLi7wv/78J/yf/9f/jf/2318sO5w7++Pv3+LXv/w5Li4uFnJVP4oiHB4ewrIsnJ2dDV1xzLIMpmnizZs3cBwHjUZj7vt/KNVqFbquz3wVcVNcXFzgT3/+M4LPPsMnu7vLDufOvr64gPX553P9PrRaLbRaLTiOgyAI4Pv+VJ+T09NT1Ot1ZFmGOI5H2gSAWq2GNE0nXvFeV9VqFQDG9lTIK/7VahVZlq3UlfggCKDr+tjenDRNkWUZhBAr1SMzL7Ztw3Gc/G9VVaf+rb/9uZcqlcrQfwe5rgshBLIsm9iLVmaa43YXuq7f6XlHR0eIoghhGMKyrKmeoygK3rx5c6/P1F3jJaL1stIJl/Tf/vsLPH3+42WHsZKCIIBt2+h0OoUnQaqqIgxDmKaJKIrWOuECxp8UPhaf7O7i9fPnyw5jJe3s7MCyrDzpmta4z1WlUuHn7q8cxxkaXrgKut3uxJPWbre7scmWTHzkcFBFUaa+OHCfk33TNO81V2ya4/aQNE2DpmlIkmTqz0qn05k6OSOix41FM9ZYlmWwbRthGE684uy67gNFtTjdbncjXgctlm3bEEIgiqKx26VpulInfOtA1/WhOTyrIEmSqbbbxGQLQH48Zn19QgjUarU777der98r+Z72uD0k27YBIJ+bNU6WZfd6/4jocWHCtcZM04SqqjAMY+K2025HtO5UVYWmaTg5ORm7Xbvdnvtwpk2nqupKVcDLsiwvdkCz6fV6hcMFp6Uoyp0/C6t63JrNJgBM1Tvu+36+PRHRJEy41pSsqiSvyE3j4OBggRERrY6Dg4OJ5bHvc7L5WERRNNKjtSoXboQQaz9EepnmcRHuLsPpVvm4KYoCwzCmSgg3dYgqES3GWszholFhGAKYbQy+rut5CeDbPM/L//HodruoVqv5P6ay8IacJxDHcT7kQgiB8/NzuK5b2lswbdvNZhOu6+Ztx3GcT+iW5XPL5idEUYTz83Ps7Ozg8vJyaB/0+FiWBcdxcHp6Wvg5SJKEwwmncHvImBACtm0jyzJkWYZ+v59f/Ol2uwBGeweCIMgT38vLSwgh4DjO2N7FLMvguu7QCe3gdzoIAoRhmD9umuZQAj1YCGKa0ttCCBwfH6NarUIIUfobIoTIf696vV7+WySHr3a7XWRZNhTbrKaNJUkSuK6bH6PB98B13QcpbCJf47Qxz3LcBp8z6+fnPmzbRhRFYwvvRFEE0zQLH5tHvLc/Z7c/t6ZpIk1T9Ho9uK5b+BuXpilOTk7yY9LtdmHb9sjnYvDfZxm3pmnIsoz/hhLNEROuNSUTp1l+xBVFKTzJrNVqODo6GrraaZpmPmdKViiTk6SDIECz2cx/pKMoQq1WK1yfZ9q2G40Ger0egiBAq9VCkiRwHCevghXHMWzbLkwY5XYyCQWQ/yPHOV+Pk/ys+75feNKQZRkTrgmEEPB9f+h9UhQFcRzDcRx4npcXGGi1Wvm6RIMnqZ7nwbKsoeQjSRJUq1XEcVx4DKIoguM4iON46PdNCAHP89BqtWBZFizLQpIkaDQaCMOw9LcwjmN4njdUxW+Q/K25XeE1CIK8bXn/7dcve0JkRUvg/e9RrVbLE9BZzBKLruvQdR1pmqJWq+HNmzdLqR45S8yzHDfgbp+f+5KVE8cV3onjuPCxecWrKArCMMTx8TE8zxt5PAzDiUmf7/tDFyflnD3XdfN/j28fH6lWq3FEDNGccUjhmpJXou47pEH+mN8eWuK6bn5CIamqina7jXq9PrRfXdchhBgZgjFL25qmDRUx0HUdV1dXQ/8Yy/LVg5Ikged5ePPmzdD9WZYVDoeix8O27bwc+CAhBOduFXBdF41GA41GA7VaDa9evSodViWHhMVxnH+/Xdcd6aFwHAfHx8dD9+m6Dk3TCi+GyB7voh7zIAjyiyuzKjvRlcPbik46LcuCpmk4PDwceZ48GfV9fyjZApD3AM5aFOKusSzTomOe9fMzL+OKZ2RZVvhvETDfeBVFGZv0lH2m5VSDwQuQsj3HcfLjMbjQ8228UEk0f0y41tTtIQB35ThO4Y+6qqpQFGXopGFnZwdCiJGrqDKW28OPZmlbPn+w7WmSScdxYBjGyLa9Xm/lylfTw5Kfi9tXojmcsJjsVYrjGGdnZ1MNixs88SzqQS8rsiHXPiuKQc6jKaIoylzn3h0eHkLTtNIEXA4vm5R4DpLxzfrbfN9YlmHRMc/6+ZkX2Ste1ItV1msOPGy8Zd/Nw8PD0vXNms3mUAXXJEkK4+LvI9H8cUjhmlJVNb96XzaMRAiBw8PDwhMU3/fzH9put1taQvv2sJhpewYW2fagstLesoeMHrdms4kgCIau2C4qEd+kSfQyeTo6Ohq7nVz7qcy4YXVFxyFN09I25eLW8xRF0dh5KvI36eTkpHSdw9vKLkAtOpZlWHTMs35+5kV+/mVCcntoa9n3fFnxDkrTtPSYyLjPz89hGAZUVUWtVkOz2USj0Ri60DHv7xrRY8eEa03JKmxJkpT+QybHgQPI5zBYlpVftZNJkWmahUlL0VXmaU8o79L2rFeu5T52dnZmeh49HnJekezVGlwgdp6EEKUFOtbZpIsg0/4eJEmCMAxRrVahKEpp8Z5xF5Dmbdr1qxRFKe2hmVdv2zxieWgPGfO0n595sm0bSZLA9/38go2cezbJMuIFvh8mmGVZ6Vpivu/nv4GdTgeHh4cIgiDfXtO0pc0HJNpkTLjWlBwudXJyMtWVKJn0DC7UKE+mFrGuziLbvr2Py8vLhe2D1psc7hSGYX7FehFJ0e2r4JtC1/WxJ9STEo4kSWCaJizLGqo62Ol0SodYzWPo1Tx7G4UQKzM8eZVimdYsMd8+bnf5/MyL/Dd2sIc8DMOxa3QtM17g++RX07SpfucGL8rKaqMnJyeo1Wpjh04S0ew4h2uNvXnzBmmalg7Zm0TOpTo/Py/d5q4J0yLbvr2fcf+QrdIirbQcspdrkU5OTjYy4bpP0iJ7A1zXHSnxXkaWo76v09PTidvMclFoEb2iqxrLtBYR8+Bxu8vnZ94sy8rnPE1K4h863qL3XR6Tab5DMsGSdF1Hq9VCp9NBq9UqrepJRHfDhGuNGYYB13XzNTnGKRvSMLju1W1pmt5rKMQi2x7cR9kCt3epFEbrrai3Uw5fLRveel9yqQQupDxM9goUXSW/3eMhK5rKiqfjTuInfadnucgiez0n7aus/PY8rVIs05pnzLeP210+P/MmqxX6vo8gCPK/iywi3nFJW1lSZRjGxPMBuZzDyclJ4eOu6078HhLRbJhwrblWqwXf97G/vw/HcQp/IGVCUjRvyrIs6Lo+8g+JEGKkmptcxLHM7cdmaXvaYSe392EYBizLKvwHPYqi0mpntJkGFx2VVFWFpmml5eB7vd7Yz/W4z6Vc22aTCmbMy7hhlmmaFr7nsmeg7Op6EARDbRZd0Z9leKcsHlR2girXIXuIqm2rFMu07hrzNMftLp+feZO/HUmS4Pz8fOJi3fOOV17EKfoNKkt037x5M/ZiYxRFebvjev7lKBUimo+1mMP1x9+/XXYI97Lo+C3Lyqux7e/vA3j/Yyl/VE3TRKvVghCi8KpYGIb51btqtQpVVdHr9fK5YbLaobwqZpomDg4OYBgGPM/L195xHAcnJydD639MajvLMjiOM9S2qqpDVeVu718O25CTen3fRxRFpfvYNF9fXCw7hHtZRPye58H3fQgh8OrVq3wOhVRUbc/zPJyfn+cni6ZpYm9vL//cyM+2fLzRaAydgAyeaK7ziYk8EZzlhFBu2+v1Sl+7bdtwHAdpmg5NwA+CAGEY5okqMFyc4+zsDPv7+yNV0+Rv1+2ES9O0fI4e8P5E9PZ3vyxplvP7TNMcWWhZruN3dnY28jzZXlG79xmGPe9Y7kO+3+OO8V1jnua43fXzM+/3wbZt2LY9sVjGIuKV1RJvL5qcpmnexuDixvI5cRzDNE2cnZ0NxZKmKXq9HgzDyI+vXEx80KTePCKa3Va/3+8vO4gy3377Lf6Pv/8f+F9//tOyQ7m3v33yIf7/r36Dly9fLjsUWlPffvst/sff/z3+9Oc/LzuUe/vwyRP85quv+H1YInki2G638xNBeVJXtgaXvCgin6OqKlRVheM4hT0vch+apqFareY97aqq5glto9EYOeETQuS9XPIiClBc3XTwQk21Ws0XmpUajQayLMt7IHRdHyl8kGUZXNfN1xW7vLzEzs5OYVymaQ61J5MORVHy4d1ZlkFRFNTrdfi+P9P8vmljuf3a5P5s275zz75ctFneAOTtAih9LbPELE06bsDsn5/bx1pV1ZHFuO+iWq2OLfm+yHhvfxcURck/x1tbW/nftysLjnse8P2wQk3TEEURFEUZSgg5OoRovlY64QLen2RerPkVfQDY3d3lySXdG78PREREROtl5RMuIiIiIiKidcWiGURERERERAvChIuIiIiIiGhBmHAREREREREtCBMuIiIiIiKiBWHCRUREREREtCBMuIiIiIiIiBaECRcREREREdGCMOEiIiIiIiJaECZcREREREREC8KEi4iIiIiIaEGYcBERERERES0IEy4iIiIiIqIFYcJFRERERES0IEy4iIiIiIiIFoQJFxERERER0YIw4SIiIiIiIloQJlxEREREREQLwoSLiIiIiIhoQZhwERERERERLQgTLiIiIiIiogVhwkVERERERLQgTLiIiIiIiIgW5H8DuK/mnrU0RNQAAAAASUVORK5CYII=",
      "text/plain": [
       "<Figure size 900x300 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "if \"seaborn-whitegrid\" in plt.style.available:\n",
    "    plt.style.use(\"seaborn-whitegrid\")\n",
    "elif \"seaborn-v0_8-whitegrid\" in plt.style.available:\n",
    "    plt.style.use('seaborn-v0_8-whitegrid')\n",
    "else:\n",
    "    print(\"Check available styles with plt.style.available\")\n",
    "\n",
    "set_one_column_figure_font(height_multiplier=0.50)\n",
    "\n",
    "melted_df_2 = melted_df.copy()\n",
    "melted_df_2['type'] = pd.Categorical(melted_df_2['type'], categories=['Generic', 'ML'], ordered=True)\n",
    "pivot_df = melted_df_2.pivot_table(index='variable', columns='type', values='value', fill_value=0)\n",
    "\n",
    "fig, ax = plt.subplots()\n",
    "colors = [non_gpu_color, gpu_color]\n",
    "\n",
    "bottom = None\n",
    "for i, col in enumerate(pivot_df.columns):\n",
    "    ax.barh(pivot_df.index, pivot_df[col], left=bottom, color=colors[i], edgecolor='black', label=col)\n",
    "    if bottom is None:\n",
    "        bottom = pivot_df[col]\n",
    "    else:\n",
    "        bottom += pivot_df[col]\n",
    "\n",
    "legend = ax.legend(title='', ncol=2, loc=\"upper left\", bbox_to_anchor=(-0.25, -0.10))\n",
    "legend.get_frame().set_alpha(0)\n",
    "loc='upper left'\n",
    "ax.set_xlabel('Fraction of Total Values')\n",
    "ax.set_ylabel('')\n",
    "\n",
    "\n",
    "plt.tight_layout()\n",
    "plt.rcParams['text.usetex'] = True\n",
    "plt.rcParams['font.family'] = 'libertine'\n",
    "\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "f1515fc9",
   "metadata": {},
   "outputs": [],
   "source": [
    "fig.savefig(\"plots/section_1/page1_gen_vs_ml.png\", dpi=900, bbox_inches='tight')\n",
    "fig.savefig(\"plots/section_1/page1_gen_vs_ml.pdf\", bbox_inches='tight')"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.12.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}