01-ticket-prices/ticket-pricing-generator.ipynb

   1 {
   2  "cells": [
   3   {
   4    "cell_type": "code",
   5    "execution_count": 45,
   6    "metadata": {
   7     "collapsed": true
   8    },
   9    "outputs": [],
  10    "source": [
  11     "import random\n",
  12     "import re\n",
  13     "import uuid"
  14    ]
  15   },
  16   {
  17    "cell_type": "code",
  18    "execution_count": 19,
  19    "metadata": {
  20     "collapsed": true
  21    },
  22    "outputs": [],
  23    "source": [
  24     "def cap_after_hyphen(string):\n",
  25     "    return re.sub(r'\\-(?P<first>[a-z])', \n",
  26     "                  lambda m: '-' + m.group('first').upper(), \n",
  27     "                  string)    "
  28    ]
  29   },
  30   {
  31    "cell_type": "code",
  32    "execution_count": 20,
  33    "metadata": {},
  34    "outputs": [
  35     {
  36      "data": {
  37       "text/plain": [
  38        "['Aalborg',\n",
  39        " 'Nullarbor',\n",
  40        " 'Morgantown',\n",
  41        " 'Estacada',\n",
  42        " 'Almaty',\n",
  43        " 'Tambon-Pa-Fa',\n",
  44        " 'Lemnos',\n",
  45        " 'Sofia',\n",
  46        " 'Puente-Laguna-Garzonkuala-Penyu',\n",
  47        " 'Hajnowski',\n",
  48        " 'Karlukovo',\n",
  49        " 'Mamula',\n",
  50        " 'Stonington-Island',\n",
  51        " 'Grimsey-Island',\n",
  52        " 'Uzupis',\n",
  53        " 'Sukhumi',\n",
  54        " 'Ijsseloog',\n",
  55        " 'Giessenmestia',\n",
  56        " 'Nordkapp',\n",
  57        " 'Gorongosa',\n",
  58        " 'Brorfelde',\n",
  59        " 'Parowan',\n",
  60        " 'Tubakuba',\n",
  61        " 'Geoje-Si',\n",
  62        " 'Mora',\n",
  63        " 'Holmegaard',\n",
  64        " 'Jayuya',\n",
  65        " 'Zhangye-Shi',\n",
  66        " 'Luoyang']"
  67       ]
  68      },
  69      "execution_count": 20,
  70      "metadata": {},
  71      "output_type": "execute_result"
  72     }
  73    ],
  74    "source": [
  75     "destination_names = ('Aalborg Nullarbor MORGANTOWN ESTACADA ALMATY TAMBON-PA-FA Lemnos Sofia Puente-Laguna-Garzon' +\n",
  76     "'KUALA-PENYU HAJNOWSKI KARLUKOVO Mamula STONINGTON-ISLAND Grimsey-Island Uzupis SUKHUMI IJsseloog GIESSEN' +\n",
  77     "'MESTIA NORDKAPP GORONGOSA Brorfelde PAROWAN Tubakuba GEOJE-SI MORA HOLMEGAARD JAYUYA ZHANGYE-SHI LUOYANG')\n",
  78     "destination_names = destination_names.lower().split()\n",
  79     "destination_names = [n[0].upper() + n[1:] for n in destination_names]\n",
  80     "destination_names = [cap_after_hyphen(n) for n in destination_names]\n",
  81     "\n",
  82     "destination_names"
  83    ]
  84   },
  85   {
  86    "cell_type": "code",
  87    "execution_count": 39,
  88    "metadata": {},
  89    "outputs": [
  90     {
  91      "data": {
  92       "text/plain": [
  93        "{'Aalborg': 1,\n",
  94        " 'Almaty': 1,\n",
  95        " 'Brorfelde': 1,\n",
  96        " 'Estacada': 1,\n",
  97        " 'Geoje-Si': 1,\n",
  98        " 'Giessenmestia': 1,\n",
  99        " 'Gorongosa': 1,\n",
 100        " 'Grimsey-Island': 1,\n",
 101        " 'Hajnowski': 1,\n",
 102        " 'Holmegaard': 1,\n",
 103        " 'Ijsseloog': 1,\n",
 104        " 'Jayuya': 1,\n",
 105        " 'Karlukovo': 1,\n",
 106        " 'Lemnos': 1,\n",
 107        " 'Luoyang': 1,\n",
 108        " 'Mamula': 1,\n",
 109        " 'Mora': 1,\n",
 110        " 'Morgantown': 1,\n",
 111        " 'Nordkapp': 1,\n",
 112        " 'Nullarbor': 1,\n",
 113        " 'Parowan': 1,\n",
 114        " 'Puente-Laguna-Garzonkuala-Penyu': 1,\n",
 115        " 'Sofia': 1,\n",
 116        " 'Stonington-Island': 1,\n",
 117        " 'Sukhumi': 1,\n",
 118        " 'Tambon-Pa-Fa': 1,\n",
 119        " 'Tubakuba': 1,\n",
 120        " 'Uzupis': 1,\n",
 121        " 'Zhangye-Shi': 1}"
 122       ]
 123      },
 124      "execution_count": 39,
 125      "metadata": {},
 126      "output_type": "execute_result"
 127     }
 128    ],
 129    "source": [
 130     "destination_scores = {n: 1 for n in destination_names}\n",
 131     "destination_scores"
 132    ]
 133   },
 134   {
 135    "cell_type": "code",
 136    "execution_count": 40,
 137    "metadata": {},
 138    "outputs": [
 139     {
 140      "data": {
 141       "text/plain": [
 142        "{'Aalborg': 1,\n",
 143        " 'Almaty': 2.0,\n",
 144        " 'Brorfelde': 0.9,\n",
 145        " 'Estacada': 0.4,\n",
 146        " 'Geoje-Si': 1,\n",
 147        " 'Giessenmestia': 1,\n",
 148        " 'Gorongosa': 1,\n",
 149        " 'Grimsey-Island': 1,\n",
 150        " 'Hajnowski': 1,\n",
 151        " 'Holmegaard': 1,\n",
 152        " 'Ijsseloog': 1,\n",
 153        " 'Jayuya': 0.6,\n",
 154        " 'Karlukovo': 2.2,\n",
 155        " 'Lemnos': 1,\n",
 156        " 'Luoyang': 1,\n",
 157        " 'Mamula': 1,\n",
 158        " 'Mora': 1,\n",
 159        " 'Morgantown': 2.9,\n",
 160        " 'Nordkapp': 1.5,\n",
 161        " 'Nullarbor': 2.2,\n",
 162        " 'Parowan': 1,\n",
 163        " 'Puente-Laguna-Garzonkuala-Penyu': 0.4,\n",
 164        " 'Sofia': 1,\n",
 165        " 'Stonington-Island': 1,\n",
 166        " 'Sukhumi': 1,\n",
 167        " 'Tambon-Pa-Fa': 1,\n",
 168        " 'Tubakuba': 1,\n",
 169        " 'Uzupis': 0.9,\n",
 170        " 'Zhangye-Shi': 1}"
 171       ]
 172      },
 173      "execution_count": 40,
 174      "metadata": {},
 175      "output_type": "execute_result"
 176     }
 177    ],
 178    "source": [
 179     "adjust = random.sample(destination_names, 10)\n",
 180     "inc_names = adjust[:5]\n",
 181     "dec_names = adjust[5:]\n",
 182     "for n in inc_names:\n",
 183     "    destination_scores[n] = float(random.randint(11, 30)) / 10\n",
 184     "for n in dec_names:\n",
 185     "    destination_scores[n] = float(random.randint(3, 9)) / 10 \n",
 186     "destination_scores"
 187    ]
 188   },
 189   {
 190    "cell_type": "code",
 191    "execution_count": 42,
 192    "metadata": {
 193     "collapsed": true
 194    },
 195    "outputs": [],
 196    "source": [
 197     "destination_scores = {'Aalborg': 1, 'Almaty': 2.0, 'Brorfelde': 0.9, 'Estacada': 0.4, 'Geoje-Si': 1,\n",
 198     " 'Giessenmestia': 1, 'Gorongosa': 1, 'Grimsey-Island': 1, 'Hajnowski': 1, 'Holmegaard': 1, 'Ijsseloog': 1,\n",
 199     " 'Jayuya': 0.6, 'Karlukovo': 2.2, 'Lemnos': 1, 'Luoyang': 1, 'Mamula': 1, 'Mora': 1, 'Morgantown': 2.9,\n",
 200     " 'Nordkapp': 1.5, 'Nullarbor': 2.2, 'Parowan': 1, 'Puente-Laguna-Garzonkuala-Penyu': 0.4, 'Sofia': 1,\n",
 201     " 'Stonington-Island': 1, 'Sukhumi': 1, 'Tambon-Pa-Fa': 1, 'Tubakuba': 1, 'Uzupis': 0.9, 'Zhangye-Shi': 1}"
 202    ]
 203   },
 204   {
 205    "cell_type": "code",
 206    "execution_count": 66,
 207    "metadata": {
 208     "collapsed": true
 209    },
 210    "outputs": [],
 211    "source": [
 212     "destination_score_exceptions = {'Almaty': 2.0, 'Brorfelde': 0.9, 'Estacada': 0.4, 'Jayuya': 0.6, 'Karlukovo': 2.2,  'Morgantown': 2.9,\n",
 213     " 'Nordkapp': 1.5, 'Nullarbor': 2.2,  'Puente-Laguna-Garzonkuala-Penyu': 0.4, 'Uzupis': 0.9}"
 214    ]
 215   },
 216   {
 217    "cell_type": "code",
 218    "execution_count": 44,
 219    "metadata": {},
 220    "outputs": [
 221     {
 222      "name": "stdout",
 223      "output_type": "stream",
 224      "text": [
 225       "| Destination | Score |\n",
 226       "|-------------|-------|\n",
 227       "| Almaty | 2.0 |\n",
 228       "| Brorfelde | 0.9 |\n",
 229       "| Estacada | 0.4 |\n",
 230       "| Jayuya | 0.6 |\n",
 231       "| Karlukovo | 2.2 |\n",
 232       "| Morgantown | 2.9 |\n",
 233       "| Nordkapp | 1.5 |\n",
 234       "| Nullarbor | 2.2 |\n",
 235       "| Puente-Laguna-Garzonkuala-Penyu | 0.4 |\n",
 236       "| Uzupis | 0.9 |\n"
 237      ]
 238     }
 239    ],
 240    "source": [
 241     "print('| Destination | Score |')\n",
 242     "print('|-------------|-------|')\n",
 243     "for n in sorted(destination_scores):\n",
 244     "    if destination_scores[n] != 1:\n",
 245     "        print('|', n, '|', destination_scores[n], '|')"
 246    ]
 247   },
 248   {
 249    "cell_type": "code",
 250    "execution_count": 48,
 251    "metadata": {
 252     "collapsed": true
 253    },
 254    "outputs": [],
 255    "source": [
 256     "def make_holiday(name=None):\n",
 257     "    if not name:\n",
 258     "        name = random.choice(destination_names)\n",
 259     "    days = random.choice([3, 7, 7, 7, 10, 14, 14, 14, 14, 21, 21])\n",
 260     "    price = 0\n",
 261     "    for _ in range(4):\n",
 262     "        price += random.randint(100, 500)\n",
 263     "    hol_id = str(uuid.uuid4())\n",
 264     "    return hol_id, price, name, days"
 265    ]
 266   },
 267   {
 268    "cell_type": "code",
 269    "execution_count": 54,
 270    "metadata": {},
 271    "outputs": [
 272     {
 273      "data": {
 274       "text/plain": [
 275        "('4615dad7-f5ab-4c9b-8028-b3e5c676ff24', 1046, 'Almaty', 14)"
 276       ]
 277      },
 278      "execution_count": 54,
 279      "metadata": {},
 280      "output_type": "execute_result"
 281     }
 282    ],
 283    "source": [
 284     "make_holiday()"
 285    ]
 286   },
 287   {
 288    "cell_type": "code",
 289    "execution_count": 59,
 290    "metadata": {},
 291    "outputs": [
 292     {
 293      "data": {
 294       "text/plain": [
 295        "[('3c425201-5538-4a3f-bfd7-1219137c9d63', 1147, 'Aalborg', 21),\n",
 296        " ('5913e357-dc18-49c5-850d-aa0172041cc4', 1398, 'Nullarbor', 3),\n",
 297        " ('bc966859-49ad-4c26-b36e-94251b0a29b5', 1064, 'Morgantown', 7),\n",
 298        " ('ab1d0a54-41bc-4351-beb8-f6e5fd4f07f0', 1722, 'Estacada', 14),\n",
 299        " ('b5efba4e-373d-4a15-b048-8fcb3169c441', 1132, 'Almaty', 7),\n",
 300        " ('bd1ff2b9-5b97-4a13-9d8a-601faea40e9c', 1085, 'Tambon-Pa-Fa', 14),\n",
 301        " ('5b9df004-a8b7-4c28-a27e-21e24cd566b5', 1264, 'Lemnos', 7),\n",
 302        " ('5951de60-9e98-471f-a917-74e3c8a3457a', 938, 'Sofia', 14),\n",
 303        " ('c6a4cad7-cd9a-4e05-bd16-92305c933269',\n",
 304        "  922,\n",
 305        "  'Puente-Laguna-Garzonkuala-Penyu',\n",
 306        "  14),\n",
 307        " ('9527ee0b-4bcd-4a1f-95fd-b3a627eacf02', 1071, 'Hajnowski', 14),\n",
 308        " ('82b6adfb-f09d-420f-bbf9-ea90450d21f6', 734, 'Karlukovo', 3),\n",
 309        " ('7b08192c-5e30-45dc-bb31-2da61861c0ac', 1692, 'Mamula', 21),\n",
 310        " ('5abbe3e4-7d2e-404c-9ce2-27083f95c140', 1277, 'Stonington-Island', 7),\n",
 311        " ('b3f517ca-2a07-427d-8531-81227468b611', 1280, 'Grimsey-Island', 3),\n",
 312        " ('f1fa90d2-0f58-480f-80c0-0fddb845b71a', 1225, 'Uzupis', 7),\n",
 313        " ('330e4f58-31d2-474f-8e23-654224689a18', 1016, 'Sukhumi', 14),\n",
 314        " ('a6837ba1-42aa-488d-b39a-5b49422ce165', 966, 'Ijsseloog', 21),\n",
 315        " ('a7e4161a-0435-4915-961e-8c204e527a4b', 1465, 'Giessenmestia', 21),\n",
 316        " ('633be2e4-a719-496e-9987-d5ea8b6ab428', 1256, 'Nordkapp', 3),\n",
 317        " ('bd83d3e3-6134-4511-918c-a94338ffb0c8', 1508, 'Gorongosa', 7),\n",
 318        " ('05ed4593-3582-4fde-8958-a63274507344', 1099, 'Brorfelde', 7),\n",
 319        " ('b8361d0a-7f23-4491-9450-c28917807a68', 1364, 'Parowan', 7),\n",
 320        " ('dc7f49c2-672e-4abc-a58a-4e8b45e786d7', 1491, 'Tubakuba', 14),\n",
 321        " ('b709a89e-e5c5-4a60-97ca-55ad350fa7f1', 1373, 'Geoje-Si', 7),\n",
 322        " ('18888ae2-e410-4430-87af-1d07ba9b25bb', 1317, 'Mora', 3),\n",
 323        " ('b08ebc25-7d02-4973-8e28-fa3b75ed4e29', 1483, 'Holmegaard', 14),\n",
 324        " ('d3519dde-9a15-4982-b1ed-1dd6dbecc2c1', 653, 'Jayuya', 21),\n",
 325        " ('fa882848-1787-424a-90c3-8593bb91f174', 1027, 'Zhangye-Shi', 10),\n",
 326        " ('21d2fcc8-f528-467d-b0ba-a6748933a4a8', 1261, 'Luoyang', 14)]"
 327       ]
 328      },
 329      "execution_count": 59,
 330      "metadata": {},
 331      "output_type": "execute_result"
 332     }
 333    ],
 334    "source": [
 335     "holidays = []\n",
 336     "for n in destination_names:\n",
 337     "    holidays += [make_holiday(n)]\n",
 338     "holidays"
 339    ]
 340   },
 341   {
 342    "cell_type": "code",
 343    "execution_count": 60,
 344    "metadata": {
 345     "collapsed": true
 346    },
 347    "outputs": [],
 348    "source": [
 349     "for _ in range(95):\n",
 350     "    holidays += [make_holiday()]"
 351    ]
 352   },
 353   {
 354    "cell_type": "code",
 355    "execution_count": 62,
 356    "metadata": {},
 357    "outputs": [],
 358    "source": [
 359     "random.shuffle(holidays)\n",
 360     "with open('01-holidays.txt', 'w') as f:\n",
 361     "    for hid, price, name, days in holidays:\n",
 362     "        f.write('{} {} {} {}\\n'.format(hid[-10:], price, name, days))"
 363    ]
 364   },
 365   {
 366    "cell_type": "code",
 367    "execution_count": 63,
 368    "metadata": {},
 369    "outputs": [
 370     {
 371      "data": {
 372       "text/plain": [
 373        "[['5f12ce1b86', '1192', 'Sukhumi', '14'],\n",
 374        " ['0279c8a91b', '1008', 'Estacada', '14'],\n",
 375        " ['1faea40e9c', '1085', 'Tambon-Pa-Fa', '14']]"
 376       ]
 377      },
 378      "execution_count": 63,
 379      "metadata": {},
 380      "output_type": "execute_result"
 381     }
 382    ],
 383    "source": [
 384     "with open('01-holidays.txt') as f:\n",
 385     "    hols = [h.split() for h in f.readlines()]\n",
 386     "hols[:3]"
 387    ]
 388   },
 389   {
 390    "cell_type": "code",
 391    "execution_count": 64,
 392    "metadata": {
 393     "collapsed": true
 394    },
 395    "outputs": [],
 396    "source": [
 397     "assert len(set(h[0] for h in hols)) == len(hols)"
 398    ]
 399   },
 400   {
 401    "cell_type": "code",
 402    "execution_count": 78,
 403    "metadata": {},
 404    "outputs": [
 405     {
 406      "data": {
 407       "text/plain": [
 408        "[('6138068a-3c16-4f7f-8dc6-a1199608abc5', 1209, 'Nordkapp', 21),\n",
 409        " ('eafb15a4-1839-4fb5-be9b-d8389018bd07', 1052, 'Estacada', 21),\n",
 410        " ('ab659b71-347d-45dc-9416-34def36ffc7d', 1514, 'Giessenmestia', 21),\n",
 411        " ('bb8d5034-49ce-483d-ab6c-d16caf2584a5', 724, 'Stonington-Island', 14),\n",
 412        " ('7476b214-8b55-47f6-833e-dfa487c4270a', 782, 'Geoje-Si', 14),\n",
 413        " ('f2ffe169-7cc5-42e1-9b46-8cdb61bb906d', 769, 'Morgantown', 3),\n",
 414        " ('2e6a6d6d-6c08-4021-b99f-05202c898b5f', 1184, 'Morgantown', 21)]"
 415       ]
 416      },
 417      "execution_count": 78,
 418      "metadata": {},
 419      "output_type": "execute_result"
 420     }
 421    ],
 422    "source": [
 423     "example_holidays = []\n",
 424     "names = random.sample(destination_score_exceptions.keys(), 2) + random.sample(destination_names, 4)\n",
 425     "names += [names[-1]]\n",
 426     "example_holidays = [make_holiday(n) for n in names]\n",
 427     "example_holidays"
 428    ]
 429   },
 430   {
 431    "cell_type": "code",
 432    "execution_count": 79,
 433    "metadata": {},
 434    "outputs": [
 435     {
 436      "data": {
 437       "text/plain": [
 438        "[('f2ffe169-7cc5-42e1-9b46-8cdb61bb906d', 769, 'Morgantown', 3),\n",
 439        " ('2e6a6d6d-6c08-4021-b99f-05202c898b5f', 1184, 'Morgantown', 21),\n",
 440        " ('ab659b71-347d-45dc-9416-34def36ffc7d', 1514, 'Giessenmestia', 21),\n",
 441        " ('eafb15a4-1839-4fb5-be9b-d8389018bd07', 1052, 'Estacada', 21),\n",
 442        " ('7476b214-8b55-47f6-833e-dfa487c4270a', 782, 'Geoje-Si', 14),\n",
 443        " ('bb8d5034-49ce-483d-ab6c-d16caf2584a5', 724, 'Stonington-Island', 14),\n",
 444        " ('6138068a-3c16-4f7f-8dc6-a1199608abc5', 1209, 'Nordkapp', 21)]"
 445       ]
 446      },
 447      "execution_count": 79,
 448      "metadata": {},
 449      "output_type": "execute_result"
 450     }
 451    ],
 452    "source": [
 453     "random.shuffle(example_holidays)\n",
 454     "example_holidays"
 455    ]
 456   },
 457   {
 458    "cell_type": "code",
 459    "execution_count": 82,
 460    "metadata": {
 461     "collapsed": true
 462    },
 463    "outputs": [],
 464    "source": [
 465     "example_holidays = [('f2ffe169-7cc5-42e1-9b46-8cdb61bb906d', 769, 'Morgantown', 3),\n",
 466     " ('2e6a6d6d-6c08-4021-b99f-05202c898b5f', 1284, 'Morgantown', 21),\n",
 467     " ('ab659b71-347d-45dc-9416-34def36ffc7d', 1514, 'Giessenmestia', 21),\n",
 468     " ('eafb15a4-1839-4fb5-be9b-d8389018bd07', 1052, 'Estacada', 21),\n",
 469     " ('7476b214-8b55-47f6-833e-dfa487c4270a', 782, 'Geoje-Si', 14),\n",
 470     " ('bb8d5034-49ce-483d-ab6c-d16caf2584a5', 724, 'Stonington-Island', 14),\n",
 471     " ('6138068a-3c16-4f7f-8dc6-a1199608abc5', 1209, 'Nordkapp', 21)]"
 472    ]
 473   },
 474   {
 475    "cell_type": "code",
 476    "execution_count": 83,
 477    "metadata": {},
 478    "outputs": [
 479     {
 480      "name": "stdout",
 481      "output_type": "stream",
 482      "text": [
 483       "db61bb906d 769 Morgantown 3\n",
 484       "202c898b5f 1284 Morgantown 21\n",
 485       "def36ffc7d 1514 Giessenmestia 21\n",
 486       "389018bd07 1052 Estacada 21\n",
 487       "a487c4270a 782 Geoje-Si 14\n",
 488       "6caf2584a5 724 Stonington-Island 14\n",
 489       "199608abc5 1209 Nordkapp 21\n"
 490      ]
 491     }
 492    ],
 493    "source": [
 494     "for hid, price, name, days in example_holidays:\n",
 495     "        print('{} {} {} {}'.format(hid[-10:], price, name, days))"
 496    ]
 497   },
 498   {
 499    "cell_type": "code",
 500    "execution_count": null,
 501    "metadata": {
 502     "collapsed": true
 503    },
 504    "outputs": [],
 505    "source": []
 506   }
 507  ],
 508  "metadata": {
 509   "kernelspec": {
 510    "display_name": "Python 3",
 511    "language": "python",
 512    "name": "python3"
 513   },
 514   "language_info": {
 515    "codemirror_mode": {
 516     "name": "ipython",
 517     "version": 3
 518    },
 519    "file_extension": ".py",
 520    "mimetype": "text/x-python",
 521    "name": "python",
 522    "nbconvert_exporter": "python",
 523    "pygments_lexer": "ipython3",
 524    "version": "3.5.2+"
 525   }
 526  },
 527  "nbformat": 4,
 528  "nbformat_minor": 2
 529 }