feature_importance.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import tuxml \n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "\n",
    "df = tuxml.load_dataset()\n",
    "df.replace(2, 0, inplace=True) # 'm' to 'n'\n",
    "# only keep 'nbyes', \n",
    "# df.drop(columns=['nbyes', nbyesmodule', 'nbno', 'nbmodule'], inplace=True)\n",
    "df.drop(columns=['nbyesmodule'], inplace=True)\n",
    "#     \n",
    "# df['CC_OPTIMIZE'] = df['CC_OPTIMIZE_FOR_SIZE']\n",
    "# 'CC_OPTIMIZE_FOR_SIZE', \n",
    "df.drop(columns=['CC_OPTIMIZE_FOR_PERFORMANCE'], inplace=True)\n",
    "# df.query(\"cid > 110000\", inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "#\"KMEMCHECK\" in df.columns\n",
    "#df['KMEMCHECK'].value_counts()\n",
    "# df['ACPI_VIDEO'].value_counts()\n",
    "# df.query(\"CC_OPTIMIZE_FOR_SIZE == 1\")['cid'].hist()\n",
    "# & DEBUG_INFO == 0 & DEBUG_INFO_SPLIT == 0 & UBSAN_SANITIZE_ALL == 0 & GCOV_PROFILE_ALL == 0 & X86_NEED_RELOCS == 0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "#df[['CC_OPTIMIZE', 'CC_OPTIMIZE_FOR_SIZE', 'CC_OPTIMIZE_FOR_PERFORMANCE']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "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>OPENVSWITCH</th>\n",
       "      <th>TEXTSEARCH_FSM</th>\n",
       "      <th>NETFILTER_XT_MATCH_TCPMSS</th>\n",
       "      <th>MPLS</th>\n",
       "      <th>NFC_HCI</th>\n",
       "      <th>NETFILTER_XT_MATCH_TIME</th>\n",
       "      <th>NET_MPLS_GSO</th>\n",
       "      <th>NFC_SHDLC</th>\n",
       "      <th>NETFILTER_XT_MATCH_U32</th>\n",
       "      <th>OPENVSWITCH_GRE</th>\n",
       "      <th>...</th>\n",
       "      <th>LZO-bzImage</th>\n",
       "      <th>LZO-vmlinux</th>\n",
       "      <th>LZO</th>\n",
       "      <th>LZ4-bzImage</th>\n",
       "      <th>LZ4-vmlinux</th>\n",
       "      <th>LZ4</th>\n",
       "      <th>cid</th>\n",
       "      <th>nbyes</th>\n",
       "      <th>nbno</th>\n",
       "      <th>nbmodule</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>14758</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>8922064.0</td>\n",
       "      <td>11008072</td>\n",
       "      <td>8734199</td>\n",
       "      <td>9839568.0</td>\n",
       "      <td>11925896</td>\n",
       "      <td>9638560</td>\n",
       "      <td>30000</td>\n",
       "      <td>1219</td>\n",
       "      <td>7319</td>\n",
       "      <td>697</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14759</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>4519376.0</td>\n",
       "      <td>6603288</td>\n",
       "      <td>4385061</td>\n",
       "      <td>4838864.0</td>\n",
       "      <td>6923096</td>\n",
       "      <td>4693085</td>\n",
       "      <td>30001</td>\n",
       "      <td>1166</td>\n",
       "      <td>6985</td>\n",
       "      <td>1084</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14760</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>12391888.0</td>\n",
       "      <td>14475800</td>\n",
       "      <td>12256864</td>\n",
       "      <td>13362640.0</td>\n",
       "      <td>15446872</td>\n",
       "      <td>13214970</td>\n",
       "      <td>30002</td>\n",
       "      <td>1410</td>\n",
       "      <td>7825</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14761</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>8782800.0</td>\n",
       "      <td>10867576</td>\n",
       "      <td>8647251</td>\n",
       "      <td>9302992.0</td>\n",
       "      <td>11388080</td>\n",
       "      <td>9155423</td>\n",
       "      <td>30003</td>\n",
       "      <td>1924</td>\n",
       "      <td>7311</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14762</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>17933264.0</td>\n",
       "      <td>20018040</td>\n",
       "      <td>17796721</td>\n",
       "      <td>19346384.0</td>\n",
       "      <td>21431472</td>\n",
       "      <td>19197696</td>\n",
       "      <td>30004</td>\n",
       "      <td>2435</td>\n",
       "      <td>6800</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14763</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>8328144.0</td>\n",
       "      <td>10412920</td>\n",
       "      <td>8193340</td>\n",
       "      <td>8897488.0</td>\n",
       "      <td>10982576</td>\n",
       "      <td>8748171</td>\n",
       "      <td>30005</td>\n",
       "      <td>1375</td>\n",
       "      <td>7021</td>\n",
       "      <td>839</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14764</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>9491408.0</td>\n",
       "      <td>11576184</td>\n",
       "      <td>9354542</td>\n",
       "      <td>10048464.0</td>\n",
       "      <td>12133552</td>\n",
       "      <td>9901074</td>\n",
       "      <td>30006</td>\n",
       "      <td>1899</td>\n",
       "      <td>7336</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14765</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>33187600.0</td>\n",
       "      <td>35274032</td>\n",
       "      <td>33029312</td>\n",
       "      <td>35424016.0</td>\n",
       "      <td>37510768</td>\n",
       "      <td>35253064</td>\n",
       "      <td>30007</td>\n",
       "      <td>3085</td>\n",
       "      <td>6150</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14766</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>6325200.0</td>\n",
       "      <td>8409976</td>\n",
       "      <td>6188651</td>\n",
       "      <td>6734800.0</td>\n",
       "      <td>8819888</td>\n",
       "      <td>6587832</td>\n",
       "      <td>30008</td>\n",
       "      <td>1294</td>\n",
       "      <td>7941</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14767</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>9085904.0</td>\n",
       "      <td>11170328</td>\n",
       "      <td>8952807</td>\n",
       "      <td>9663440.0</td>\n",
       "      <td>11748184</td>\n",
       "      <td>9518916</td>\n",
       "      <td>30009</td>\n",
       "      <td>2291</td>\n",
       "      <td>6944</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>10 rows × 9257 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       OPENVSWITCH  TEXTSEARCH_FSM  NETFILTER_XT_MATCH_TCPMSS  MPLS  NFC_HCI  \\\n",
       "14758            0               0                          0     1        0   \n",
       "14759            0               0                          0     0        0   \n",
       "14760            0               0                          0     0        0   \n",
       "14761            0               0                          0     0        0   \n",
       "14762            0               0                          0     0        1   \n",
       "14763            0               1                          0     0        0   \n",
       "14764            0               0                          0     1        0   \n",
       "14765            0               0                          0     0        1   \n",
       "14766            0               0                          0     0        0   \n",
       "14767            0               0                          0     1        0   \n",
       "\n",
       "       NETFILTER_XT_MATCH_TIME  NET_MPLS_GSO  NFC_SHDLC  \\\n",
       "14758                        0             1          0   \n",
       "14759                        0             0          0   \n",
       "14760                        0             0          0   \n",
       "14761                        0             0          0   \n",
       "14762                        0             0          1   \n",
       "14763                        0             0          0   \n",
       "14764                        0             1          0   \n",
       "14765                        0             0          1   \n",
       "14766                        0             0          0   \n",
       "14767                        0             0          0   \n",
       "\n",
       "       NETFILTER_XT_MATCH_U32  OPENVSWITCH_GRE  ...  LZO-bzImage  LZO-vmlinux  \\\n",
       "14758                       0                0  ...    8922064.0     11008072   \n",
       "14759                       0                0  ...    4519376.0      6603288   \n",
       "14760                       0                0  ...   12391888.0     14475800   \n",
       "14761                       0                0  ...    8782800.0     10867576   \n",
       "14762                       0                0  ...   17933264.0     20018040   \n",
       "14763                       0                0  ...    8328144.0     10412920   \n",
       "14764                       0                0  ...    9491408.0     11576184   \n",
       "14765                       0                0  ...   33187600.0     35274032   \n",
       "14766                       0                0  ...    6325200.0      8409976   \n",
       "14767                       0                0  ...    9085904.0     11170328   \n",
       "\n",
       "            LZO  LZ4-bzImage  LZ4-vmlinux       LZ4    cid  nbyes  nbno  \\\n",
       "14758   8734199    9839568.0     11925896   9638560  30000   1219  7319   \n",
       "14759   4385061    4838864.0      6923096   4693085  30001   1166  6985   \n",
       "14760  12256864   13362640.0     15446872  13214970  30002   1410  7825   \n",
       "14761   8647251    9302992.0     11388080   9155423  30003   1924  7311   \n",
       "14762  17796721   19346384.0     21431472  19197696  30004   2435  6800   \n",
       "14763   8193340    8897488.0     10982576   8748171  30005   1375  7021   \n",
       "14764   9354542   10048464.0     12133552   9901074  30006   1899  7336   \n",
       "14765  33029312   35424016.0     37510768  35253064  30007   3085  6150   \n",
       "14766   6188651    6734800.0      8819888   6587832  30008   1294  7941   \n",
       "14767   8952807    9663440.0     11748184   9518916  30009   2291  6944   \n",
       "\n",
       "       nbmodule  \n",
       "14758       697  \n",
       "14759      1084  \n",
       "14760         0  \n",
       "14761         0  \n",
       "14762         0  \n",
       "14763       839  \n",
       "14764         0  \n",
       "14765         0  \n",
       "14766         0  \n",
       "14767         0  \n",
       "\n",
       "[10 rows x 9257 columns]"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df[:10]\n",
    "# df.sort_values(by='nbyes')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "size_methods = [\"vmlinux\", \"GZIP-bzImage\", \"GZIP-vmlinux\", \"GZIP\", \"BZIP2-bzImage\", \n",
    "              \"BZIP2-vmlinux\", \"BZIP2\", \"LZMA-bzImage\", \"LZMA-vmlinux\", \"LZMA\", \"XZ-bzImage\", \"XZ-vmlinux\", \"XZ\", \n",
    "              \"LZO-bzImage\", \"LZO-vmlinux\", \"LZO\", \"LZ4-bzImage\", \"LZ4-vmlinux\", \"LZ4\"]\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "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>OPENVSWITCH</th>\n",
       "      <th>TEXTSEARCH_FSM</th>\n",
       "      <th>NETFILTER_XT_MATCH_TCPMSS</th>\n",
       "      <th>MPLS</th>\n",
       "      <th>NFC_HCI</th>\n",
       "      <th>NETFILTER_XT_MATCH_TIME</th>\n",
       "      <th>NET_MPLS_GSO</th>\n",
       "      <th>NFC_SHDLC</th>\n",
       "      <th>NETFILTER_XT_MATCH_U32</th>\n",
       "      <th>OPENVSWITCH_GRE</th>\n",
       "      <th>...</th>\n",
       "      <th>LZO-bzImage</th>\n",
       "      <th>LZO-vmlinux</th>\n",
       "      <th>LZO</th>\n",
       "      <th>LZ4-bzImage</th>\n",
       "      <th>LZ4-vmlinux</th>\n",
       "      <th>LZ4</th>\n",
       "      <th>cid</th>\n",
       "      <th>nbyes</th>\n",
       "      <th>nbno</th>\n",
       "      <th>nbmodule</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>14758</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>8922064.0</td>\n",
       "      <td>11008072</td>\n",
       "      <td>8734199</td>\n",
       "      <td>9839568.0</td>\n",
       "      <td>11925896</td>\n",
       "      <td>9638560</td>\n",
       "      <td>30000</td>\n",
       "      <td>1219</td>\n",
       "      <td>7319</td>\n",
       "      <td>697</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14759</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>4519376.0</td>\n",
       "      <td>6603288</td>\n",
       "      <td>4385061</td>\n",
       "      <td>4838864.0</td>\n",
       "      <td>6923096</td>\n",
       "      <td>4693085</td>\n",
       "      <td>30001</td>\n",
       "      <td>1166</td>\n",
       "      <td>6985</td>\n",
       "      <td>1084</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14760</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>12391888.0</td>\n",
       "      <td>14475800</td>\n",
       "      <td>12256864</td>\n",
       "      <td>13362640.0</td>\n",
       "      <td>15446872</td>\n",
       "      <td>13214970</td>\n",
       "      <td>30002</td>\n",
       "      <td>1410</td>\n",
       "      <td>7825</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14761</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>8782800.0</td>\n",
       "      <td>10867576</td>\n",
       "      <td>8647251</td>\n",
       "      <td>9302992.0</td>\n",
       "      <td>11388080</td>\n",
       "      <td>9155423</td>\n",
       "      <td>30003</td>\n",
       "      <td>1924</td>\n",
       "      <td>7311</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14762</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>17933264.0</td>\n",
       "      <td>20018040</td>\n",
       "      <td>17796721</td>\n",
       "      <td>19346384.0</td>\n",
       "      <td>21431472</td>\n",
       "      <td>19197696</td>\n",
       "      <td>30004</td>\n",
       "      <td>2435</td>\n",
       "      <td>6800</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14763</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>8328144.0</td>\n",
       "      <td>10412920</td>\n",
       "      <td>8193340</td>\n",
       "      <td>8897488.0</td>\n",
       "      <td>10982576</td>\n",
       "      <td>8748171</td>\n",
       "      <td>30005</td>\n",
       "      <td>1375</td>\n",
       "      <td>7021</td>\n",
       "      <td>839</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14764</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>9491408.0</td>\n",
       "      <td>11576184</td>\n",
       "      <td>9354542</td>\n",
       "      <td>10048464.0</td>\n",
       "      <td>12133552</td>\n",
       "      <td>9901074</td>\n",
       "      <td>30006</td>\n",
       "      <td>1899</td>\n",
       "      <td>7336</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14765</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>33187600.0</td>\n",
       "      <td>35274032</td>\n",
       "      <td>33029312</td>\n",
       "      <td>35424016.0</td>\n",
       "      <td>37510768</td>\n",
       "      <td>35253064</td>\n",
       "      <td>30007</td>\n",
       "      <td>3085</td>\n",
       "      <td>6150</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14766</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>6325200.0</td>\n",
       "      <td>8409976</td>\n",
       "      <td>6188651</td>\n",
       "      <td>6734800.0</td>\n",
       "      <td>8819888</td>\n",
       "      <td>6587832</td>\n",
       "      <td>30008</td>\n",
       "      <td>1294</td>\n",
       "      <td>7941</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14767</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>9085904.0</td>\n",
       "      <td>11170328</td>\n",
       "      <td>8952807</td>\n",
       "      <td>9663440.0</td>\n",
       "      <td>11748184</td>\n",
       "      <td>9518916</td>\n",
       "      <td>30009</td>\n",
       "      <td>2291</td>\n",
       "      <td>6944</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>10 rows × 9257 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       OPENVSWITCH  TEXTSEARCH_FSM  NETFILTER_XT_MATCH_TCPMSS  MPLS  NFC_HCI  \\\n",
       "14758            0               0                          0     1        0   \n",
       "14759            0               0                          0     0        0   \n",
       "14760            0               0                          0     0        0   \n",
       "14761            0               0                          0     0        0   \n",
       "14762            0               0                          0     0        1   \n",
       "14763            0               1                          0     0        0   \n",
       "14764            0               0                          0     1        0   \n",
       "14765            0               0                          0     0        1   \n",
       "14766            0               0                          0     0        0   \n",
       "14767            0               0                          0     1        0   \n",
       "\n",
       "       NETFILTER_XT_MATCH_TIME  NET_MPLS_GSO  NFC_SHDLC  \\\n",
       "14758                        0             1          0   \n",
       "14759                        0             0          0   \n",
       "14760                        0             0          0   \n",
       "14761                        0             0          0   \n",
       "14762                        0             0          1   \n",
       "14763                        0             0          0   \n",
       "14764                        0             1          0   \n",
       "14765                        0             0          1   \n",
       "14766                        0             0          0   \n",
       "14767                        0             0          0   \n",
       "\n",
       "       NETFILTER_XT_MATCH_U32  OPENVSWITCH_GRE  ...  LZO-bzImage  LZO-vmlinux  \\\n",
       "14758                       0                0  ...    8922064.0     11008072   \n",
       "14759                       0                0  ...    4519376.0      6603288   \n",
       "14760                       0                0  ...   12391888.0     14475800   \n",
       "14761                       0                0  ...    8782800.0     10867576   \n",
       "14762                       0                0  ...   17933264.0     20018040   \n",
       "14763                       0                0  ...    8328144.0     10412920   \n",
       "14764                       0                0  ...    9491408.0     11576184   \n",
       "14765                       0                0  ...   33187600.0     35274032   \n",
       "14766                       0                0  ...    6325200.0      8409976   \n",
       "14767                       0                0  ...    9085904.0     11170328   \n",
       "\n",
       "            LZO  LZ4-bzImage  LZ4-vmlinux       LZ4    cid  nbyes  nbno  \\\n",
       "14758   8734199    9839568.0     11925896   9638560  30000   1219  7319   \n",
       "14759   4385061    4838864.0      6923096   4693085  30001   1166  6985   \n",
       "14760  12256864   13362640.0     15446872  13214970  30002   1410  7825   \n",
       "14761   8647251    9302992.0     11388080   9155423  30003   1924  7311   \n",
       "14762  17796721   19346384.0     21431472  19197696  30004   2435  6800   \n",
       "14763   8193340    8897488.0     10982576   8748171  30005   1375  7021   \n",
       "14764   9354542   10048464.0     12133552   9901074  30006   1899  7336   \n",
       "14765  33029312   35424016.0     37510768  35253064  30007   3085  6150   \n",
       "14766   6188651    6734800.0      8819888   6587832  30008   1294  7941   \n",
       "14767   8952807    9663440.0     11748184   9518916  30009   2291  6944   \n",
       "\n",
       "       nbmodule  \n",
       "14758       697  \n",
       "14759      1084  \n",
       "14760         0  \n",
       "14761         0  \n",
       "14762         0  \n",
       "14763       839  \n",
       "14764         0  \n",
       "14765         0  \n",
       "14766         0  \n",
       "14767         0  \n",
       "\n",
       "[10 rows x 9257 columns]"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df[:10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# df.sort_values(by='vmlinux', ascending=True)[['vmlinux', 'nbno', 'nbyes', 'nbmodule', 'nbyesmodule']][:10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "HAVE_ALIGNED_STRUCT_PAGE            1\n",
       "SLUB                                1\n",
       "X86_DIRECT_GBPAGES                  1\n",
       "BROKEN_ON_SMP                       1\n",
       "EXPERT                              1\n",
       "HZ_PERIODIC                         1\n",
       "NEED_PER_CPU_KM                     1\n",
       "HAVE_HARDENED_USERCOPY_ALLOCATOR    1\n",
       "TINY_RCU                            1\n",
       "TINY_SRCU                           1\n",
       "INLINE_SPIN_UNLOCK_IRQ              1\n",
       "INLINE_READ_UNLOCK                  1\n",
       "INLINE_READ_UNLOCK_IRQ              1\n",
       "INLINE_WRITE_UNLOCK                 1\n",
       "INLINE_WRITE_UNLOCK_IRQ             1\n",
       "X86_MPPARSE                         1\n",
       "CPU_SUP_INTEL                       1\n",
       "DEBUG_KERNEL                        1\n",
       "DEFAULT_SECURITY_DAC                1\n",
       "EMBEDDED                            1\n",
       "CPU_SUP_AMD                         1\n",
       "CPU_SUP_CENTAUR                     1\n",
       "UP_LATE_INIT                        1\n",
       "SCSI_MOD                            1\n",
       "Name: 109097, dtype: int64"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# modified tiny\n",
    "mtinyconfig = df.sort_values(by='vmlinux', ascending=True).drop(columns=[\"cid\"]).drop(columns=size_methods).iloc[0]\n",
    "mtinyconfig[mtinyconfig == 1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "X86_DIRECT_GBPAGES         1\n",
       "BROKEN_ON_SMP              1\n",
       "EXPERT                     1\n",
       "HZ_PERIODIC                1\n",
       "NEED_PER_CPU_KM            1\n",
       "OPTIMIZE_INLINING          1\n",
       "TINY_RCU                   1\n",
       "TINY_SRCU                  1\n",
       "INLINE_SPIN_UNLOCK_IRQ     1\n",
       "INLINE_READ_UNLOCK         1\n",
       "INLINE_READ_UNLOCK_IRQ     1\n",
       "INLINE_WRITE_UNLOCK        1\n",
       "INLINE_WRITE_UNLOCK_IRQ    1\n",
       "X86_MPPARSE                1\n",
       "CPU_SUP_INTEL              1\n",
       "DEBUG_KERNEL               1\n",
       "CC_OPTIMIZE_FOR_SIZE       1\n",
       "DEFAULT_SECURITY_DAC       1\n",
       "EMBEDDED                   1\n",
       "SLOB                       1\n",
       "CPU_SUP_AMD                1\n",
       "CPU_SUP_CENTAUR            1\n",
       "UP_LATE_INIT               1\n",
       "SCSI_MOD                   1\n",
       "Name: 58621, dtype: int64"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# the real tiny\n",
    "tinyconfig = df.sort_values(by='vmlinux', ascending=True).drop(columns=[\"cid\"]).drop(columns=size_methods).iloc[1]\n",
    "tinyconfig[tinyconfig == 1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "HAVE_ALIGNED_STRUCT_PAGE differs\n",
      "real tiny config: 0 vs modified tiny config: 1\n",
      "SLUB differs\n",
      "real tiny config: 0 vs modified tiny config: 1\n",
      "HAVE_HARDENED_USERCOPY_ALLOCATOR differs\n",
      "real tiny config: 0 vs modified tiny config: 1\n",
      "OPTIMIZE_INLINING differs\n",
      "real tiny config: 1 vs modified tiny config: 0\n",
      "CC_OPTIMIZE_FOR_SIZE differs\n",
      "real tiny config: 1 vs modified tiny config: 0\n",
      "SLOB differs\n",
      "real tiny config: 1 vs modified tiny config: 0\n",
      "nbyes differs\n",
      "real tiny config: 24 vs modified tiny config: 25\n",
      "nbno differs\n",
      "real tiny config: 9211 vs modified tiny config: 9210\n"
     ]
    }
   ],
   "source": [
    "for o in tinyconfig.index:\n",
    "    if (tinyconfig[o] != mtinyconfig[o]):\n",
    "        print(o, \"differs\")\n",
    "        print(\"real tiny config:\", tinyconfig[o], \"vs modified tiny config:\", mtinyconfig[o])\n",
    "#for o in tinyconfig:\n",
    "#    print(o)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.5745765013612033\n",
      "-0.11588853956517173\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "''"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#df[df.columns[0:110]].corr()\n",
    "# df[['DEBUG_INFO', 'ATH9K_COMMON_DEBUG']].corr()\n",
    "# df[df.columns[6900:7000]].corr()['MODULES']\n",
    "from scipy.stats import pearsonr \n",
    "# df.to_pickle(\"all_size_withyes.pkl\")\n",
    "print(pearsonr(df['DEBUG_INFO'], df['vmlinux'])[0])\n",
    "print(pearsonr(df['CC_OPTIMIZE_FOR_SIZE'], df['vmlinux'])[0])\n",
    "# print(pearsonr(df['CC_OPTIMIZE_FOR_PERFORMANCE'], df['vmlinux'])[0])\n",
    "# pd.DataFrame(columns=[\"MODULES\"], data=[pearsonr(df['MODULES'], df[col])[0] for col in df]).sort_values(by=\"MODULES\", na_position='last', ascending=False)\n",
    "# df.columns[12403]\n",
    "# pearsonr(df['MODULES'], df['STRICT_MODULE_RWX'])[0] # pvalue 0\n",
    "def _filter(df):\n",
    "    return df.drop(columns=[\"cid\"]).drop(columns=size_methods)\n",
    "#correlation_vmlinux = _filter(df)[_filter(df).columns].corrwith(df['vmlinux']).dropna().sort_values(ascending=False)\n",
    "#correlation_vmlinux.to_csv(\"correlations_vmlinux.csv\")\n",
    "#correlation_vmlinux\n",
    "#df.columns.get_loc('MODULES')\n",
    "\"\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "# correlation_vmlinux['nbno']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "#from sklearn.feature_selection import SelectKBest\n",
    "#from sklearn.feature_selection import f_regression, mutual_info_regression\n",
    "#transformer = SelectKBest(f_regression, k=100)\n",
    "#X_new = transformer.fit_transform(_filter(df), df['vmlinux'])\n",
    "#transf_support = pd.DataFrame(transformer.get_support())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "#X_new.shape\n",
    "# df.columns[23116]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "#for i in transf_support.rename(columns={0 : \"isit\"}).query(\"isit == True\").index: # [transf_support == True]\n",
    "#    print(_filter(df).columns[i])\n",
    "# _filter(df).columns[5829]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "fdf = _filter(df)\n",
    "\n",
    "dict_corr = []\n",
    "def _samecolumns_after_ithcolumn(fd, i):\n",
    "    col_i = fd[fd.columns[i]] # the column we want to compare with\n",
    "    for j in range(i+1, len(fd.columns) - 1):\n",
    "        if (col_i.equals(fd[fd.columns[j]])):\n",
    "            dict_corr.append({fd.columns[i] : fd.columns[j]})\n",
    "            print(fd.columns[i], \"correlated to\", fd.columns[j])\n",
    "\n",
    "\n",
    "#for i in range(0, len(fdf)):\n",
    "#    print(i)\n",
    "#    _samecolumns_after_ithcolumn(fdf, i)\n",
    "    #if (i == 10):\n",
    "    #    break\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "# fdf.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "# dict_corr"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "# import scipy.stats\n",
    "# def ttest_option(ft_name):\n",
    "    # return scipy.stats.ttest_ind(df.query(ft_name + \" == 1\")['vmlinux'], df.query(ft_name + \" != 1\")['vmlinux'])\n",
    "#    return scipy.stats.ttest_ind(df[df[ft_name] == 1]['vmlinux'], df[df[ft_name] != 1]['vmlinux'])\n",
    "#ttest_option(\"CC_OPTIMIZE_FOR_PERFORMANCE\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn import ensemble\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn import tree\n",
    "from sklearn.inspection import partial_dependence\n",
    "# from sklearn.ensemble.partial_dependence import partial_dependence, plot_partial_dependence\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.ensemble import GradientBoostingRegressor\n",
    "import numpy as np\n",
    "\n",
    "\n",
    "def run_regressorML(reg, test_size, size_target=\"vmlinux\", ft_importances=None):\n",
    "    assert(size_target in size_methods)\n",
    "    if (ft_importances is None):\n",
    "        ft_importances = df.drop(columns=[\"cid\"]).drop(columns=size_methods).columns # all\n",
    "    X_train, X_test, y_train, y_test = train_test_split(df.drop(columns=[\"cid\"]).drop(columns=size_methods)[ft_importances], df[size_target], test_size=test_size, random_state=42)  \n",
    "    reg.fit(X_train, y_train)\n",
    "    y_pred = reg.predict(X_test)\n",
    "    dfErrors = pd.DataFrame({\"error\":(y_pred - y_test).abs(), \"% error\":((y_pred - y_test) / y_test).abs()*100})\n",
    "#    print(partial_dependence(reg, X_test, ftsDep)) # , method='brute')\n",
    "    return dfErrors[\"% error\"].describe()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Index(['nbyes', 'UBSAN_SANITIZE_ALL', 'GCOV_PROFILE_ALL', 'KASAN',\n",
      "       'KASAN_OUTLINE', 'UBSAN_ALIGNMENT', 'BLK_MQ_PCI', 'KCOV_INSTRUMENT_ALL',\n",
      "       'nbno', 'MDIO',\n",
      "       ...\n",
      "       'HAVE_ALIGNED_STRUCT_PAGE', 'GENERIC_NET_UTILS', 'SCSI_FC_ATTRS',\n",
      "       'MODULE_SIG_FORCE', 'SERIAL_ALTERA_JTAGUART', 'LWTUNNEL',\n",
      "       'NET_VENDOR_OKI', 'TRACER_MAX_TRACE', 'IPV6_SEG6_LWTUNNEL',\n",
      "       'DRM_GMA600'],\n",
      "      dtype='object', length=200)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "count    9248.000000\n",
       "mean        5.315073\n",
       "std         4.287579\n",
       "min         0.000152\n",
       "25%         2.123000\n",
       "50%         4.366777\n",
       "75%         7.436246\n",
       "max        42.533826\n",
       "Name: % error, dtype: float64"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.pipeline import Pipeline\n",
    "from sklearn.feature_selection import SelectFromModel\n",
    "from sklearn import linear_model\n",
    "   \n",
    "hyperparams_DT = {\n",
    "    \"criterion\": \"mse\",\n",
    "    \"max_depth\": 25,\n",
    "    \"min_samples_leaf\": 5,\n",
    "    \"max_leaf_nodes\": None,\n",
    "}\n",
    "\n",
    "# reg_dt = tree.DecisionTreeRegressor(**hyperparams_DT)\n",
    "# res_dt = run_regressorML(reg_dt, test_size=0.9)\n",
    "# res_dt\n",
    "\n",
    "hyperparams_RF = {\n",
    "    \"max_depth\": 25,\n",
    "    \"n_estimators\": 10,\n",
    "    \"n_jobs\" : -1\n",
    "}\n",
    "\n",
    "#reg_rf = Pipeline([\n",
    "#   ('feature_selection', SelectFromModel(ensemble.RandomForestRegressor(**hyperparams_RF))), # threshold=-np.inf, max_features=1000)),\n",
    "#  ('feature_selection', SelectFromModel(GradientBoostingRegressor())), # linear_model.Lasso(alpha=0.1))),\n",
    "#  ('regression', ensemble.RandomForestRegressor(**hyperparams_RF))\n",
    "#])\n",
    "\n",
    "#reg_rf = ensemble.RandomForestRegressor(**hyperparams_RF)\n",
    "#res_rf = run_regressorML(reg_rf, test_size=0.7, size_target=\"vmlinux\")\n",
    "#res_rf\n",
    "\n",
    "#hyperparams_GB = {\n",
    "#    'learning_rate': 0.1, \n",
    "#    'max_depth': 25, \n",
    "#    'n_estimators': 10    \n",
    "# }\n",
    "\n",
    "# number of features in feature selection\n",
    "compressed_size = \"GZIP\" # None # \n",
    "nft=200\n",
    "\n",
    "if (compressed_size == \"GZIP\"):\n",
    "    df_importance = pd.read_csv(\"feature_importanceRF-GZIP.csv\")\n",
    "    df_importance.rename(columns={\"Unnamed: 0\" : 'option', 1 : 'importance' }, inplace=True)\n",
    "else:\n",
    "    df_importance = pd.read_csv(\"feature_importanceRF.csv\", header=None)\n",
    "    df_importance.rename(columns={0 : 'option', 1 : 'importance' }, inplace=True)\n",
    "    ### for Lasso\n",
    "    # df_importance = pd.read_csv(\"feature_importanceLasso09.csv\")\n",
    "    # df_importance = df_importance.append({'option' : 'nbyes', 'importance' : 1}, ignore_index=True)\n",
    "    # nft = len(df_importance)\n",
    "df_importance.query(\"option != 'nbyesmodule' and option != 'CC_OPTIMIZE_FOR_PERFORMANCE'\", inplace=True)\n",
    "# df_importance = df[df_importance[:100][\"Unnamed: 0\"].values]\n",
    "df_importance = df[df_importance[:nft][\"option\"]].columns\n",
    "\n",
    "print(df_importance)\n",
    "# df_importance.head()\n",
    "\n",
    "\n",
    "#reg_gb = GradientBoostingRegressor(**hyperparams_GB)\n",
    "#res_gb = run_regressorML(reg_gb, test_size=0.9, size_target=\"vmlinux\", ft_importances=None)\n",
    "#res_gb = run_regressorML(reg_gb, test_size=0.9, size_target=\"vmlinux\", ft_importances=df_importance)\n",
    "#res_gb\n",
    "\n",
    "\n",
    "hyperparams_ET = {\n",
    "   \"max_depth\": 25,\n",
    "    \"n_estimators\": 10,\n",
    "    \"n_jobs\" : -1\n",
    " }\n",
    "\n",
    "exp_size_target = \"vmlinux\" # default \n",
    "if (compressed_size == \"GZIP\"):\n",
    "    exp_size_target = \"GZIP-vmlinux\"\n",
    "\n",
    "reg_et = ensemble.ExtraTreesRegressor(**hyperparams_ET)\n",
    "res_et = run_regressorML(reg_et, test_size=0.1, size_target=exp_size_target, ft_importances=df_importance)\n",
    "# res_et = run_regressorML(reg_et, test_size=0.9, size_target=exp_size_target, ft_importances=None)\n",
    "res_et\n",
    "\n",
    "\n",
    "\n",
    "# run_regressorML(ensemble.AdaBoostRegressor(), test_size=0.98, size_target=\"vmlinux\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'res_gb' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m-----------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m                       Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-22-822f35d5ac82>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mres_gb\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m: name 'res_gb' is not defined"
     ]
    }
   ],
   "source": [
    "res_gb"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "UBSAN_SANITIZE_ALL                  0.176364\n",
       "nbyes                               0.163695\n",
       "MODULES                             0.092045\n",
       "BLK_MQ_PCI                          0.085223\n",
       "STRICT_MODULE_RWX                   0.068285\n",
       "GCOV_PROFILE_ALL                    0.066845\n",
       "KASAN_OUTLINE                       0.052527\n",
       "KASAN                               0.049133\n",
       "UBSAN_ALIGNMENT                     0.029723\n",
       "nbno                                0.024367\n",
       "KCOV_INSTRUMENT_ALL                 0.015497\n",
       "MII                                 0.015148\n",
       "MDIO                                0.009573\n",
       "SCSI_ISCSI_ATTRS                    0.008758\n",
       "PHYLIB                              0.006063\n",
       "UBSAN_NULL                          0.005017\n",
       "DRM_I915                            0.004096\n",
       "NET_RX_BUSY_POLL                    0.004060\n",
       "IOSCHED_NOOP                        0.003658\n",
       "DQL                                 0.003081\n",
       "NET                                 0.002969\n",
       "DRM_AMDGPU                          0.002902\n",
       "MAY_USE_DEVLINK                     0.002556\n",
       "DRM_TTM                             0.002487\n",
       "SBITMAP                             0.002459\n",
       "DRM_NOUVEAU                         0.002438\n",
       "nbmodule                            0.002359\n",
       "BLOCK                               0.002274\n",
       "GRO_CELLS                           0.002273\n",
       "DRM_RADEON                          0.002104\n",
       "                                      ...   \n",
       "COMPILE_TEST                        0.000223\n",
       "TIMER_OF                            0.000219\n",
       "FUJITSU_ES                          0.000209\n",
       "BNX2                                0.000206\n",
       "IP_SCTP                             0.000203\n",
       "BLK_DEV_TRIFLEX                     0.000195\n",
       "CLKSRC_MMIO                         0.000188\n",
       "MDIO_CAVIUM                         0.000176\n",
       "HAVE_ALIGNED_STRUCT_PAGE            0.000175\n",
       "SLOB                                0.000173\n",
       "SLUB                                0.000173\n",
       "TIMER_PROBE                         0.000169\n",
       "SCSI_FC_ATTRS                       0.000154\n",
       "HAVE_HARDENED_USERCOPY_ALLOCATOR    0.000152\n",
       "FEALNX                              0.000141\n",
       "SCSI_PMCRAID                        0.000127\n",
       "FIXED_PHY                           0.000123\n",
       "NET_VENDOR_TEHUTI                   0.000120\n",
       "SCSI_BNX2_ISCSI                     0.000119\n",
       "SWPHY                               0.000112\n",
       "MODULE_SIG_FORCE                    0.000108\n",
       "RTLWIFI_USB                         0.000106\n",
       "LIVEPATCH                           0.000086\n",
       "KRETPROBES                          0.000077\n",
       "KPROBES                             0.000074\n",
       "NFP                                 0.000053\n",
       "VIDEO_SAA7134_GO7007                0.000021\n",
       "NET_SCTPPROBE                       0.000010\n",
       "VIDEO_COBALT                        0.000007\n",
       "XEN_BLKDEV_BACKEND                  0.000005\n",
       "Length: 199, dtype: float64"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def mk_ftimportances(reg, ft_importances=None):\n",
    "    col = df.drop(columns=[\"cid\"]).drop(columns=size_methods).columns\n",
    "    if isinstance(reg, Pipeline):\n",
    "        col = col[reg.named_steps['feature_selection'].get_support()]\n",
    "        imps = reg.named_steps['regression'].feature_importances_        \n",
    "    else:\n",
    "        imps = reg.feature_importances_\n",
    "    if (ft_importances is None):\n",
    "        ftimportance = pd.Series(imps, index=col.values)\n",
    "    else:\n",
    "        ftimportance = pd.Series(imps, index=ft_importances)\n",
    "    return ftimportance[ftimportance > 0].sort_values(ascending=False)\n",
    "\n",
    "#ft_dt = mk_ftimportances(reg_dt)\n",
    "#ft_dt.to_csv(\"feature_importanceDT.csv\", header=True)\n",
    "#ft_dt\n",
    "\n",
    "#ft_rf = mk_ftimportances(reg_rf)\n",
    "#ft_rf.to_csv(\"feature_importanceRF-ccoptimize.csv\", header=True)\n",
    "#ft_rf\n",
    "#ft_rf.rank(ascending=False)[['CC_OPTIMIZE_FOR_SIZE', 'MODULES']]\n",
    "# ft_rf[ft_rf.index.get_loc(\"CC_OPTIMIZE_FOR_SIZE\")]\n",
    "# df['CC_OPTIMIZE_FOR_SIZE'].value_counts()\n",
    "\n",
    "#ft_gb = mk_ftimportances(reg_gb, ft_importances=df_importance)\n",
    "#ft_gb.to_csv(\"feature_importanceGB-ccoptimize10.csv\", header=True)\n",
    "#ft_gb\n",
    "\n",
    "ft_et = mk_ftimportances(reg_et, ft_importances=df_importance)\n",
    "if (compressed_size == \"GZIP\"):\n",
    "       ft_et.to_csv(\"feature_importanceET-withftimportance-GZIP-\" + str(nft) + \".csv\", header=True)\n",
    "else:\n",
    "    ft_et.to_csv(\"feature_importanceET-withftimportance\" + str(nft) + \".csv\", header=True)\n",
    "ft_et\n",
    "#'nbyes' in ft_et.index"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "df['DELL_WMI'].value_counts(), df['ACPI_VIDEO'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "ft_et.rank(ascending=False)['CC_OPTIMIZE_FOR_SIZE']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "####### column' indices of feature importances\n",
    "# TODO: enhance!\n",
    "if (False):\n",
    "    col = df.drop(columns=[\"cid\"]).drop(columns=size_methods).columns\n",
    "    ftIndiceImportanceGB = []\n",
    "    for ft_name in ft_rf.index:\n",
    "        i = 0\n",
    "        for vl in col.values:\n",
    "            if (ft_name == vl):\n",
    "                ftIndiceImportanceGB.append(i)\n",
    "    #            print(ft_name, \"at\", i)\n",
    "                continue\n",
    "            i = i + 1\n",
    "    ftIndiceImportanceGB\n",
    "\n",
    "    # TODO: instead of plotting, retrieving the coefficients\n",
    "    col = df.drop(columns=[\"cid\"]).drop(columns=size_methods).columns\n",
    "    import matplotlib.pyplot as plt\n",
    "    # TODO: pair-wise feature interactions\n",
    "    ftsDep = ftIndiceImportanceGB[:20] # importanceGB[importanceGB > 0].sort_values(ascending=False)[:10].values\n",
    "    plt.figure(figsize=(20, 20))\n",
    "    # TODO: X_train is an input! (it works only if X_train is accessible/visible)\n",
    "    fig, ax = plot_partial_dependence(reg_rf, X_train, ftsDep, feature_names=col.values, grid_resolution=100)\n",
    "    fig.set_figwidth(8)\n",
    "    fig.set_figheight(15)\n",
    "    fig.tight_layout()\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import sklearn\n",
    "sklearn.__version__"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}