{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Utilities to read/write csv files\n",
    "import csv\n",
    "# Utilities to handle character encodings\n",
    "import unicodedata\n",
    "# Ordered Dicts\n",
    "from collections import OrderedDict\n",
    "\n",
    "import json\n",
    "\n",
    "\n",
    "# OPZIONAL IMPORTS\n",
    "\n",
    "# For timestamping/simple speed tests\n",
    "from datetime import datetime\n",
    "# Random number generator\n",
    "from random import *\n",
    "# System & command line utilities\n",
    "import sys\n",
    "# Json for the dictionary\n",
    "import json\n",
    "import re"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import_dir = '/Users/federicaspinelli/Google Drive/OVI-CNR/CSV/ASPO/marcovaldi/'\n",
    "export_dir = '/Users/federicaspinelli/Google Drive/OVI-CNR/RDF/ASPO/marcovaldi/'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Custom class to store URIs + related infos for the ontologies/repositories\n",
    "\n",
    "class RDFcoords:\n",
    "    def __init__(self, uri, prefix, code = None):\n",
    "        self.uri = uri\n",
    "        self.prefix = prefix\n",
    "        self.code = code\n",
    "\n",
    "\n",
    "# Repositories\n",
    "datiniCoords = RDFcoords('<http://www.archiviodistato.prato.it/accedi-e-consulta/aspoMV001/scheda/IT-ASPO-MV001->', 'dt:')\n",
    "personAuthCoords = RDFcoords('<http://www.archiviodistato.prato.it/accedi-e-consulta/aspoMV001/scheda/IT-ASPO-AU00003->', 'pa:')\n",
    "# W3/CIDOC Predicates\n",
    "hasTypeCoords = RDFcoords('<http://www.w3.org/1999/02/22-rdf-syntax-ns#type>', 'tp:')\n",
    "hasTypePCoords = RDFcoords('<http://www.cidoc-crm.org/cidoc-crm/P2_has_type>', 'te:')\n",
    "carriesCoords = RDFcoords('<http://www.cidoc-crm.org/cidoc-crm/P128_carries>', 'ca:')\n",
    "identifiedByCoords = RDFcoords('<http://www.cidoc-crm.org/cidoc-crm/P1_is_identified_by>', 'ib:')\n",
    "labelCoords = RDFcoords('<http://www.w3.org/2000/01/rdf-schema#label>', 'lb:')\n",
    "wasBroughtCoords = RDFcoords('<http://www.cidoc-crm.org/cidoc-crm/P92i_was_brought_into_existence_by>', 'wb:')\n",
    "carriedByCoords = RDFcoords('<http://www.cidoc-crm.org/cidoc-crm/P14_carried_out_by>', 'cb:')\n",
    "movedByCoords = RDFcoords('<http://www.cidoc-crm.org/cidoc-crm/P25i_moved_by>', 'mb:')\n",
    "subClassOfCoords = RDFcoords('<http://www.w3.org/2000/01/rdf-schema#subClassOf>', 'so:')\n",
    "tookPlaceCoords = RDFcoords('<http://www.cidoc-crm.org/cidoc-crm/P7_took_place_at>', 'tk:')\n",
    "carriedByCoords = RDFcoords('<http://www.cidoc-crm.org/cidoc-crm/P14_carried_out_by>', 'cb:')\n",
    "roleOfCoords = RDFcoords('<http://www.cidoc-crm.org/cidoc-crm/P14.1_in_the_role_of>', 'ro:')\n",
    "hasDomainCoords = RDFcoords('<http://www.cidoc-crm.org/cidoc-crm/P01_has_domain>', 'hd:')\n",
    "hasRangeCoords = RDFcoords('<http://www.cidoc-crm.org/cidoc-crm/P02_has_range>', 'hr:')\n",
    "hasTimeSpanCoords = RDFcoords('<http://www.cidoc-crm.org/cidoc-crm/P4_has_time-span>', 'hs:')\n",
    "movedToCoords = RDFcoords('<http://www.cidoc-crm.org/cidoc-crm/P26_moved_to>', 'mt:')\n",
    "movedFromCoords = RDFcoords('<http://www.cidoc-crm.org/cidoc-crm/P27_moved_from>', 'mf:')\n",
    "\n",
    "#CIDOC Objects\n",
    "manMadeObjectCoords = RDFcoords('<http://www.cidoc-crm.org/cidoc-crm/E22_Man-Made_Object>', 'mo:', 'E22')\n",
    "informationObjectCoords = RDFcoords('<http://www.cidoc-crm.org/cidoc-crm/E73_Information_Object>', 'io:', 'E73')\n",
    "titleCoords = RDFcoords('<http://www.cidoc-crm.org/cidoc-crm/E35_Title>', 'ti:' ,'E35')\n",
    "placeAppellationCoords = RDFcoords('<http://www.cidoc-crm.org/cidoc-crm/E44_Place_appellation>', 'pa:', 'E44')\n",
    "identifierCoords = RDFcoords('<http://www.cidoc-crm.org/cidoc-crm/E42_Identifier>', 'id:', 'E42')\n",
    "typeCoords = RDFcoords('<http://www.cidoc-crm.org/cidoc-crm/E55_Type>', 'ty:', 'E55')\n",
    "creationCoords = RDFcoords('<http://www.cidoc-crm.org/cidoc-crm/E65_Creation>', 'cr:', 'E65')\n",
    "personCoords = RDFcoords('<http://www.cidoc-crm.org/cidoc-crm/E21_Person>', 'ps:', 'E21')\n",
    "pcarriedByCoords = RDFcoords('<http://www.cidoc-crm.org/cidoc-crm/PC14_carried_out_by>', 'cy:', 'PC14')\n",
    "timeSpanCoords = RDFcoords('<http://www.cidoc-crm.org/cidoc-crm/E52_Time-Span>', 'ts:', 'E52')\n",
    "placeCoords = RDFcoords('<http://www.cidoc-crm.org/cidoc-crm/E53_Place>', 'pl:', 'E53')\n",
    "\n",
    "\n",
    "# New classes (subclasses of E7 Activity) - Exchange, Sending, Recive Letters\n",
    "exchangeLettersCoords = RDFcoords('<http://www.cidoc-crm.org/cidoc-crm/EL1_Exchange_Letters>', 'el:', 'EL1')\n",
    "sendLetterCoords = RDFcoords('<http://www.cidoc-crm.org/cidoc-crm/EL2_Send_Letter>', 'sl:', 'EL2')\n",
    "receiveLetterCoords = RDFcoords('<http://www.cidoc-crm.org/cidoc-crm/EL3_Receive_Letter>', 'rl:', 'EL3')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Basic functions for triples / shortened triples in TTL format\n",
    "\n",
    "def triple(subject, predicate, object1):\n",
    "    line = subject + ' ' + predicate + ' ' + object1\n",
    "    return line\n",
    "\n",
    "def doublet(predicate, object1):\n",
    "    line = '    ' + predicate + ' ' + object1\n",
    "    return line\n",
    "\n",
    "def singlet(object1):\n",
    "    line = '        ' + object1\n",
    "    return line\n",
    "\n",
    "# Line endings in TTL format\n",
    "continueLine1 = ' ;\\n'\n",
    "continueLine2 = ' ,\\n'\n",
    "closeLine = ' .\\n'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "def writeTTLHeader(output):\n",
    "    output.write('@prefix ' + datiniCoords.prefix + ' ' + datiniCoords.uri + closeLine)\n",
    "    output.write('@prefix ' + personAuthCoords.prefix + ' ' + personAuthCoords.uri + closeLine)\n",
    "    output.write('@prefix ' + hasTypeCoords.prefix + ' ' + hasTypeCoords.uri + closeLine)\n",
    "    output.write('@prefix ' + hasTypePCoords.prefix + ' ' + hasTypePCoords.uri + closeLine)\n",
    "    output.write('@prefix ' + manMadeObjectCoords.prefix + ' ' + manMadeObjectCoords.uri + closeLine)\n",
    "    output.write('@prefix ' + carriesCoords.prefix + ' ' + carriesCoords.uri + closeLine)\n",
    "    output.write('@prefix ' + informationObjectCoords.prefix + ' ' + informationObjectCoords.uri + closeLine)\n",
    "    output.write('@prefix ' + identifiedByCoords.prefix + ' ' + identifiedByCoords.uri + closeLine)\n",
    "    output.write('@prefix ' + titleCoords.prefix + ' ' + titleCoords.uri + closeLine)\n",
    "    output.write('@prefix ' + labelCoords.prefix + ' ' + labelCoords.uri + closeLine)\n",
    "    output.write('@prefix ' + identifierCoords.prefix + ' ' + identifierCoords.uri + closeLine)\n",
    "    output.write('@prefix ' + wasBroughtCoords.prefix + ' ' + wasBroughtCoords.uri + closeLine)\n",
    "    output.write('@prefix ' + typeCoords.prefix + ' ' + typeCoords.uri + closeLine)\n",
    "    output.write('@prefix ' + carriedByCoords.prefix + ' ' + carriedByCoords.uri + closeLine)\n",
    "    output.write('@prefix ' + personCoords.prefix + ' ' + personCoords.uri + closeLine)\n",
    "    output.write('@prefix ' + exchangeLettersCoords.prefix + ' ' + exchangeLettersCoords.uri + closeLine)\n",
    "    output.write('@prefix ' + sendLetterCoords.prefix + ' ' + sendLetterCoords.uri + closeLine)\n",
    "    output.write('@prefix ' + receiveLetterCoords.prefix + ' ' + receiveLetterCoords.uri + closeLine)\n",
    "    output.write('@prefix ' + subClassOfCoords.prefix + ' ' + subClassOfCoords.uri + closeLine)\n",
    "    output.write('@prefix ' + movedByCoords.prefix + ' ' + movedByCoords.uri + closeLine)\n",
    "    output.write('@prefix ' + roleOfCoords.prefix + ' ' + roleOfCoords.uri + closeLine)\n",
    "    output.write('@prefix ' + hasDomainCoords.prefix + ' ' + hasDomainCoords.uri + closeLine)\n",
    "    output.write('@prefix ' + pcarriedByCoords.prefix + ' ' + pcarriedByCoords.uri + closeLine)\n",
    "    output.write('@prefix ' + hasRangeCoords.prefix + ' ' + hasRangeCoords.uri + closeLine)\n",
    "    output.write('@prefix ' + timeSpanCoords.prefix + ' ' + timeSpanCoords.uri + closeLine)\n",
    "    output.write('@prefix ' + placeCoords.prefix + ' ' + placeCoords.uri + closeLine)\n",
    "    output.write('@prefix ' + hasTimeSpanCoords.prefix + ' ' + hasTimeSpanCoords.uri + closeLine)\n",
    "    output.write('@prefix ' + movedToCoords.prefix + ' ' + movedToCoords.uri + closeLine)\n",
    "    output.write('@prefix ' + movedFromCoords.prefix + ' ' + movedFromCoords.uri + closeLine)\n",
    "    \n",
    "    output.write('\\n')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "filePrefix = 'data_'\n",
    "fileType = 'item'\n",
    "max_entries = 1000000000\n",
    "\n",
    "with open(import_dir + filePrefix + fileType + '.csv', newline=\"\") as csv_file, open(export_dir + filePrefix + fileType + '_date.ttl', 'w') as output:\n",
    "    reader = csv.DictReader(csv_file)\n",
    "    writeTTLHeader(output)\n",
    "    first = True\n",
    "    ii = 0\n",
    "    for row in reader:\n",
    "        # The index ii is used to process a limited number of entries for testing purposes\n",
    "        ii = ii+1\n",
    "        # Skip the first line as it carries info we don't want to triplify\n",
    "        if(first):\n",
    "            first = False\n",
    "            continue\n",
    "        el1placeHolder = \"<http://www.archiviodistato.prato.it/accedi-e-consulta/aspoMV001/scheda/\" + row['id'] + \"/\" + exchangeLettersCoords.code + \">\"\n",
    "        el2placeHolder = \"<http://www.archiviodistato.prato.it/accedi-e-consulta/aspoMV001/scheda/\" + row['id'] + \"/\" + sendLetterCoords.code + \">\"\n",
    "        el3placeHolder = \"<http://www.archiviodistato.prato.it/accedi-e-consulta/aspoMV001/scheda/\" + row['id'] + \"/\" + receiveLetterCoords.code + \">\"\n",
    "        \n",
    "        # Data invio\n",
    "        if(row['data_inizio'] != ''):\n",
    "            e52PplaceHolder = \"<http://www.archiviodistato.prato.it/accedi-e-consulta/aspoMV001/scheda/\" + row['id'] + \"/\" + sendLetterCoords.code + \"_\" + timeSpanCoords.code + \">\"\n",
    "            line = triple(el2placeHolder, e52PplaceHolder, '\\\"'+ row['data_inizio'] + '\\\"') + closeLine\n",
    "            output.write(line)\n",
    "       \n",
    "        if(row['data_fine'] != ''):\n",
    "            e52AplaceHolder = \"<http://www.archiviodistato.prato.it/accedi-e-consulta/aspoMV001/scheda/\" + row['id'] + \"/\" + receiveLetterCoords.code + \"_\" + timeSpanCoords.code + \">\"\n",
    "            line = triple(el3placeHolder, e52AplaceHolder, '\\\"' + row['data_fine'] + '\\\"') + closeLine\n",
    "            output.write(line)\n",
    "             \n",
    "        output.write('\\n')\n",
    "        #\n",
    "        #\n",
    "        # Limit number of entries processed (if desired)\n",
    "        if(ii>max_entries):\n",
    "            break\n",
    "        "
   ]
  }
 ],
 "metadata": {
  "interpreter": {
   "hash": "31f2aee4e71d21fbe5cf8b01ff0e069b9275f58929596ceb00d14d90e3e16cd6"
  },
  "kernelspec": {
   "display_name": "Python 3.7.3 64-bit",
   "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"
  },
  "metadata": {
   "interpreter": {
    "hash": "31f2aee4e71d21fbe5cf8b01ff0e069b9275f58929596ceb00d14d90e3e16cd6"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}