atlas_engine 0.1.0

data/db/data/transcriber.yml

@@ -0,0 +1,760 @@
+---
+affixed_suffixes:
+  laan: "ln"
+  plantsoen: "plnts"
+  straat: "str"
+  strasse: "str"
+apostrophes:
+ - "\u0027"
+ - "\u00b4"
+ - "\u2018"
+ - "\u2019"
+directionals:
+  e: "e"
+  east: "e"
+  north: "n"
+  ne: "ne"
+  northeast: "ne"
+  "north-east": "ne"
+  nw: "nw"
+  northwest: "nw"
+  "north-west": "nw"
+  s: "s"
+  south: "s"
+  se: "se"
+  southeast: "se"
+  "south-east": "se"
+  sw: "sw"
+  southwest: "sw"
+  "south-west": "sw"
+  w: "w"
+  west: "w"
+hyphens:
+ - "\u002d"
+ - "\u058a"
+ - "\u05be"
+ - "\u1400"
+ - "\u1806"
+ - "\u2010"
+ - "\u2011"
+ - "\u2012"
+ - "\u2013"
+ - "\u2014"
+ - "\u2015"
+ - "\u2e3a"
+ - "\u2e3b"
+ - "\ufe58"
+ - "\ufe63"
+ - "\uff0d"
+months:
+  january: jan
+  february: feb
+  march: mar
+  april: apr
+  may: may
+  june: jun
+  july: jul
+  august: aug
+  september: sep
+  sept: sep
+  october: oct
+  november: nov
+  december: dec
+numbered_routes:
+  "private road": "pr"
+  "rural route": "rr"
+  "rural road": "rr"
+  "carretera": "carr"
+  "co highway": "cty hwy"
+  "co highwy": "cty hwy"
+  "co hiway": "cty hwy"
+  "co hiwy": "cty hwy"
+  "co hway": "cty hwy"
+  "co hwy": "cty hwy"
+  "county highway": "cty hwy"
+  "county highwy": "cty hwy"
+  "county hiway": "cty hwy"
+  "county hiwy": "cty hwy"
+  "county hway": "cty hwy"
+  "county hwy": "cty hwy"
+  "cty highway": "cty hwy"
+  "cty highwy": "cty hwy"
+  "cty hiway": "cty hwy"
+  "cty hiwy": "cty hwy"
+  "cty hway": "cty hwy"
+  "county road": "cr"
+  "county rd": "cr"
+  "cty road": "cr"
+  "cty rd": "cr"
+  "farm to market": "fm"
+  "farm to market rd": "fm"
+  "farm to market road": "fm"
+  "farm road": "farm road"
+  highway: "hwy"
+  highwy: "hwy"
+  hiway: "hwy"
+  hiwy: "hwy"
+  hway: "hwy"
+  loop: "loop"
+  route: "rte"
+  rt: "rte"
+  "state highway": "state hwy"
+  "state highwy": "state hwy"
+  "state hiway": "state hwy"
+  "state hiwy": "state hwy"
+  "state hway": "state hwy"
+  "state road": "sr"
+  "state rd": "sr"
+  "state route": "sr"
+  "state rte": "sr"
+  us: "us"
+  "u.s": "us"
+  "us highway": "us hwy"
+  "u.s. highway": "us hwy"
+ordinals_english:
+  "1st": "first"
+  "2nd": "second"
+  "3rd": "third"
+  "4th": "fourth"
+  "5th": "fifth"
+  "6th": "sixth"
+  "7th": "seventh"
+  "8th": "eighth"
+  "9th": "ninth"
+  "10th": "tenth"
+  "11th": "eleventh"
+  "12th": "twelfth"
+  "13th": "thirteenth"
+  "14th": "fourteenth"
+  "15th": "fifteenth"
+  "16th": "sixteenth"
+  "17th": "seventeenth"
+  "18th": "eighteenth"
+  "19th": "nineteenth"
+  "20th": "twentieth"
+  "30th": "thirtieth"
+  "40th": "fortieth"
+  "50th": "fiftieth"
+  "60th": "sixtieth"
+  "70th": "seventieth"
+  "80th": "eightieth"
+  "90th": "nintieth"
+  "100th": "one hundredth"
+ordinals_english_tens:
+  "20": "twenty"
+  "30": "thirty"
+  "40": "forty"
+  "50": "fifty"
+  "60": "sixty"
+  "70": "seventy"
+  "80": "eighty"
+  "90": "ninety"
+  "100": "one hundred"
+ordinals_other:
+  "1e": "eerste"
+  "1er": "premier"
+secondary_unit_designators:
+  apartment: "apt"
+  basement: "bsmt"
+  building: "bldg"
+  department: "dept"
+  floor: "fl"
+  front: "frnt"
+  hanger: "hngr"
+  key: "key"
+  lobby: "lbby"
+  lot: "lot"
+  lower: "lowr"
+  office: "ofc"
+  pier: "pier"
+  penthouse: "ph"
+  rear: "rear"
+  room: "rm"
+  side: "side"
+  slip: "slip"
+  space: "spc"
+  stop: "stop"
+  suite: "ste"
+  trailer: "trlr"
+  unit: "unit"
+  upper: "uppr"
+street_suffixes:
+  # "Suffixes" that are recognized categories of street, along with their USPS canonical abbreviation.
+  # Note that all USPS-recognized suffixes must appear in this list, even if the suffix is its own abbreviation.
+  # Thus, for example, "lndg" maps to "lndg".  We can use the suffix's presence in this list to help us parse
+  # address1 fields by identifying a token that could plausibly be a street suffix.
+  allee: "aly"
+  alley: "aly"
+  ally: "aly"
+  aly: "aly"
+  anex: "anx"
+  annex: "anx"
+  annx: "anx"
+  anx: "anx"
+  arc: "arc"
+  arcade: "arc"
+  av: "ave"
+  ave: "ave"
+  aven: "ave"
+  avenu: "ave"
+  avenue: "ave"
+  avnue: "ave"
+  bayoo: "byu"
+  bayou: "byu"
+  byu: "byu"
+  bch: "bch"
+  beach: "bch"
+  bend: "bnd"
+  bnd: "bnd"
+  blf: "blf"
+  bluf: "blf"
+  bluff: "blf"
+  blfs: "blfs"
+  bluffs: "blfs"
+  blvd: "blvd"
+  bot: "btm"
+  btm: "btm"
+  bottm: "btm"
+  bottom: "btm"
+  boul: "blvd"
+  boulevard: "blvd"
+  boulv: "blvd"
+  br: "br"
+  brnch: "br"
+  branch: "br"
+  brdge: "brg"
+  brg: "brg"
+  bridge: "brg"
+  brk: "brk"
+  brook: "brk"
+  brooks: "brks"
+  bg: "bg"
+  burg: "bg"
+  burgs: "bgs"
+  byp: "byp"
+  bypa: "byp"
+  bypas: "byp"
+  bypass: "byp"
+  camp: "cp"
+  cmp: "cp"
+  cp: "cp"
+  canyn: "cyn"
+  canyon: "cyn"
+  cnyn: "cyn"
+  cyn: "cyn"
+  cape: "cpe"
+  cpe: "cpe"
+  causeway: "cswy"
+  causewa: "cswy"
+  cswy: "cswy"
+  cen: "ctr"
+  cent: "ctr"
+  center: "ctr"
+  centr: "ctr"
+  centre: "ctr"
+  cnter: "ctr"
+  cntr: "ctr"
+  ctr: "ctr"
+  centers: "ctrs"
+  ctrs: "ctrs"
+  cir: "cir"
+  circ: "cir"
+  circl: "cir"
+  circle: "cir"
+  crcl: "cir"
+  crcle: "cir"
+  circles: "cirs"
+  clf: "clf"
+  cliff: "clf"
+  clfs: "clfs"
+  cliffs: "clfs"
+  clb: "clb"
+  club: "clb"
+  cmn: "cmn"
+  common: "cmn"
+  cmns: "cmns"
+  commons: "cmns"
+  cor: "cor"
+  corner: "cor"
+  corners: "cors"
+  cors: "cors"
+  course: "crse"
+  crse: "crse"
+  court: "ct"
+  crt: "ct"
+  ct: "ct"
+  courts: "cts"
+  crts: "cts"
+  cts: "cts"
+  cove: "cv"
+  cv: "cv"
+  coves: "cvs"
+  cvs: "cvs"
+  creek: "crk"
+  crk: "crk"
+  crescent: "cres"
+  cres: "cres"
+  crsent: "cres"
+  crsnt: "cres"
+  crest: "crst"
+  crst: "crst"
+  crossing: "xing"
+  crssing: "xing"
+  xing: "xing"
+  crossroad: "xrd"
+  xrd: "xrd"
+  crossroads: "xrds"
+  xrds: "xrds"
+  curve: "curv"
+  curv: "curv"
+  dale: "dl"
+  dl: "dl"
+  dam: "dm"
+  dm: "dm"
+  div: "dv"
+  divide: "dv"
+  dv: "dv"
+  dvd: "dv"
+  dr: "dr"
+  driv: "dr"
+  drive: "dr"
+  drv: "dr"
+  drives: "drs"
+  drs: "drs"
+  estate: "est"
+  est: "est"
+  estates: "ests"
+  ests: "ests"
+  exp: "expy"
+  expr: "expy"
+  express: "expy"
+  expressway: "expy"
+  expw: "expy"
+  expy: "expy"
+  ext: "ext"
+  extension: "ext"
+  extn: "ext"
+  extnsn: "ext"
+  fall: "fall"
+  falls: "fls"
+  ferry: "fry"
+  frry: "fry"
+  fry: "fry"
+  field: "fld"
+  fld: "fld"
+  fields: "flds"
+  flds: "flds"
+  flat: "flt"
+  flt: "flt"
+  flats: "flts"
+  flts: "flts"
+  ford: "frd"
+  frd: "frd"
+  fords: "frds"
+  frds: "frds"
+  forest: "frst"
+  forests: "frst"
+  frst: "frst"
+  forg: "frg"
+  forge: "frg"
+  frg: "frg"
+  forges: "frgs"
+  frgs: "frgs"
+  fork: "frk"
+  frk: "frk"
+  forks: "frks"
+  frks: "frks"
+  fort: "ft"
+  frt: "ft"
+  ft: "ft"
+  freeway: "fwy"
+  freewy: "fwy"
+  frway: "fwy"
+  frwy: "fwy"
+  garden: "gdn"
+  gardn: "gdn"
+  gdn: "gdn"
+  grden: "gdn"
+  gardens: "gdns"
+  gdns: "gdns"
+  grdns: "gdns"
+  gateway: "gtwy"
+  gatewy: "gtwy"
+  gatway: "gtwy"
+  gtway: "gtwy"
+  gtwy: "gtwy"
+  glen: "gln"
+  gln: "gln"
+  glens: "glns"
+  glns: "glns"
+  green: "grn"
+  grn: "grn"
+  greens: "grns"
+  grns: "grns"
+  grov: "grv"
+  grove: "grv"
+  grv: "grv"
+  groves: "grvs"
+  grvs: "grvs"
+  harb: "hbr"
+  harbor: "hbr"
+  harbr: "hbr"
+  hbr: "hbr"
+  hrbor: "hbr"
+  harbors: "hbrs"
+  hbrs: "hbrs"
+  haven: "hvn"
+  hvn: "hvn"
+  heights: "hts"
+  ht: "hts"
+  hts: "hts"
+  highway: "hwy"
+  highwy: "hwy"
+  hiway: "hwy"
+  hiwy: "hwy"
+  hway: "hwy"
+  hwy: "hwy"
+  hill: "hl"
+  hl: "hl"
+  hills: "hls"
+  hls: "hls"
+  hllw: "holw"
+  hollow: "holw"
+  hollows: "holw"
+  holw: "holw"
+  holws: "holw"
+  inlet: "inlt"
+  inlt: "inlt"
+  is: "is"
+  island: "is"
+  islnd: "is"
+  islands: "iss"
+  islnds: "iss"
+  iss: "iss"
+  isle: "isle"
+  isles: "isle"
+  jct: "jct"
+  jction: "jct"
+  jctn: "jct"
+  junction: "jct"
+  junctn: "jct"
+  juncton: "jct"
+  jctns: "jcts"
+  jcts: "jcts"
+  junctions: "jcts"
+  key: "ky"
+  ky: "ky"
+  keys: "kys"
+  kys: "kys"
+  knl: "knl"
+  knol: "knl"
+  knoll: "knl"
+  knls: "knls"
+  knolls: "knls"
+  lake: "lk"
+  lk: "lk"
+  lakes: "lks"
+  lks: "lks"
+  land: "land"
+  landing: "lndg"
+  lndg: "lndg"
+  lndng: "lndg"
+  lane: "ln"
+  ln: "ln"
+  lgt: "lgt"
+  light: "lgt"
+  lgts: "lgts"
+  lights: "lgts"
+  lf: "lf"
+  loaf: "lf"
+  lck: "lck"
+  lock: "lck"
+  lcks: "lcks"
+  locks: "lcks"
+  ldg: "ldg"
+  ldge: "ldg"
+  lodg: "ldg"
+  lodge: "ldg"
+  loop: "loop"
+  loops: "loop"
+  mall: "mall"
+  manor: "mnr"
+  manors: "mnrs"
+  meadow: "mdw"
+  mdw: "mdw"
+  mdws: "mdws"
+  meadows: "mdws"
+  medows: "mdws"
+  ml: "ml"
+  mill: "ml"
+  mills: "mls"
+  missn: "msn"
+  msn: "msn"
+  mssn: "msn"
+  motorway: "mtwy"
+  mtwy: "mtwy"
+  mnt: "mt"
+  mount: "mt"
+  mt: "mt"
+  mntain: "mtn"
+  mntn: "mtn"
+  mountain: "mtn"
+  mountin: "mtn"
+  mtin: "mtn"
+  mtn: "mtn"
+  mntns: "mtns"
+  mountains: "mtns"
+  nck: "nck"
+  neck: "nck"
+  orch: "orch"
+  orchard: "orch"
+  orchrd: "orch"
+  oval: "oval"
+  ovl: "oval"
+  overpass: "opas"
+  park: "park"
+  prk: "park"
+  parks: "park"
+  parkway: "pkwy"
+  parkwy: "pkwy"
+  pkway: "pkwy"
+  pkwy: "pkwy"
+  pky: "pkwy"
+  parkways: "pkwy"
+  pkwys: "pkwy"
+  pass: "pass"
+  passage: "psge"
+  path: "path"
+  paths: "path"
+  pike: "pike"
+  pikes: "pike"
+  pine: "pne"
+  pne: "pne"
+  pines: "pnes"
+  pnes: "pnes"
+  pl: "pl"
+  place: "pl"
+  plain: "pln"
+  pln: "pln"
+  plains: "plns"
+  plns: "plns"
+  plaza: "plz"
+  plz: "plz"
+  plza: "plz"
+  point: "pt"
+  pt: "pt"
+  points: "pts"
+  pts: "pts"
+  port: "prt"
+  prt: "prt"
+  ports: "prts"
+  prts: "prts"
+  pr: "pr"
+  prairie: "pr"
+  prr: "pr"
+  rad: "radl"
+  radial: "radl"
+  radiel: "radl"
+  radl: "radl"
+  ramp: "ramp"
+  ranch: "rnch"
+  ranches: "rnch"
+  rnch: "rnch"
+  rnchs: "rnchs"
+  rapid: "rpd"
+  rpd: "rpd"
+  rapids: "rpds"
+  rpds: "rpds"
+  rest: "rst"
+  rst: "rst"
+  rdg: "rdg"
+  rdge: "rdg"
+  ridge: "rdg"
+  rdgs: "rdgs"
+  ridges: "rdgs"
+  riv: "riv"
+  river: "riv"
+  rvr: "riv"
+  rivr: "riv"
+  rd: "rd"
+  road: "rd"
+  rds: "rds"
+  roads: "rds"
+  route: "rte"
+  rte: "rte"
+  row: "row"
+  rue: "rue"
+  run: "run"
+  shoal: "shl"
+  shl: "shl"
+  shls: "shls"
+  shoals: "shls"
+  shr: "shr"
+  shoar: "shr"
+  shore: "shr"
+  shrs: "shrs"
+  shoars: "shrs"
+  shores: "shrs"
+  skwy: "skwy"
+  skyway: "skwy"
+  slip: "slip"
+  spg: "spg"
+  spng: "spg"
+  spring: "spg"
+  spgs: "spgs"
+  spngs: "spgs"
+  springs: "spgs"
+  spur: "spur"
+  sq: "sq"
+  sqr: "sq"
+  sqre: "sq"
+  squ: "sq"
+  square: "sq"
+  sqrs: "sqs"
+  sqs: "sqs"
+  squares: "sqs"
+  sta: "sta"
+  station: "sta"
+  statn: "sta"
+  stn: "sta"
+  stra: "stra"
+  strav: "stra"
+  straven: "stra"
+  stravenue: "stra"
+  stravn: "stra"
+  strvn: "stra"
+  strvnue: "stra"
+  stream: "strm"
+  streme: "strm"
+  strm: "strm"
+  st: "st"
+  street: "st"
+  strt: "st"
+  str: "st"
+  sts: "sts"
+  strts: "sts"
+  streets: "sts"
+  smt: "smt"
+  sumit: "smt"
+  summit: "smt"
+  ter: "ter"
+  terr: "ter"
+  terrace: "ter"
+  throughway: "trwy"
+  trwy: "trwy"
+  trace: "trce"
+  trce: "trce"
+  traces: "trce"
+  track: "trak"
+  tracks: "trak"
+  trk: "trak"
+  trks: "trak"
+  trafficway: "trfy"
+  trfy: "trfy"
+  trail: "trl"
+  trails: "trl"
+  trl: "trl"
+  trls: "trl"
+  trailer: "trlr"
+  trlr: "trlr"
+  trlrs: "trlr"
+  tunel: "tunl"
+  tunl: "tunl"
+  tunls: "tunl"
+  tunnel: "tunl"
+  tunnels: "tunl"
+  tunnl: "tunl"
+  tpke: "tpke"
+  trnpk: "tpke"
+  turnpike: "tpke"
+  turnpk: "tpke"
+  underpass: "upas"
+  upas: "upas"
+  un: "un"
+  union: "un"
+  unions: "uns"
+  uns: "uns"
+  valley: "vly"
+  vally: "vly"
+  vlly: "vly"
+  vly: "vly"
+  valleys: "vlys"
+  vlys: "vlys"
+  vdct: "via"
+  via: "via"
+  viadct: "via"
+  viaduct: "via"
+  view: "vw"
+  vw: "vw"
+  views: "vws"
+  vws: "vws"
+  vill: "vlg"
+  villag: "vlg"
+  village: "vlg"
+  villg: "vlg"
+  villiage: "vlg"
+  vlg: "vlg"
+  villages: "vlgs"
+  vlgs: "vlgs"
+  ville: "vl"
+  vis: "vis"
+  vist: "vis"
+  vista: "vis"
+  vst: "vis"
+  vsta: "vis"
+  walk: "walk"
+  walks: "walk"
+  way: "way"
+  wy: "way"
+  ways: "ways"
+  wells: "wls"
+  wls: "wls"
+substitutions:
+  "1e": "eerste"
+  eerste: "1e"
+  burgemeester: "burg"
+  commissaris: "comm"
+  de: "d"
+  doctor: "dr"
+  dokter: "dr"
+  doktor: "dr"
+  monseigneur: "mgr"
+  pastoor: "past"
+  prof: "pr"
+  professor: "pr"
+  saint: "st"
+  sainte: "st"
+  san: "st"
+  sankt: "st"
+  sint: "st"
+  ste: "st"
+  township: "twp"
+  van: "v"
+  vd: "van der"
+translations_en_fr:
+  ave: av
+  blvd: boul
+  east: est
+  north: nord
+  northeast: nord-est
+  northwest: nord-ouest
+  south: sud
+  southeast: sud-est
+  southwest: sud-ouest
+  st: rue
+  street: rue
+  west: ouest
+translations_fr_en:
+  av: ave
+  boul: blvd
+  est: east
+  nord: north
+  "nord-est": northeast
+  "nord-ouest": northwest
+  ouest: west
+  rue: street
+  sud: south
+  "sud-est": southeast
+  "sud-ouest": southwest
+with_diacritics: "ÀÁÂÃÄÅàáâãäåĀāĂ㥹ÇçĆćĈĉĊċČčÐðĎďĐđÈÉÊËèéêëĒēĔĕĖėĘęĚěĜĝĞğĠġĢģĤĥĦħÌÍÎÏìíîïĨĩĪīĬĭĮįİıĴĵĶķĸĹĺĻļĽľĿŀŁłÑñŃńŅņŇňʼnŊŋÒÓÔÕÖØòóôõöøŌōŎŏŐőŔŕŖŗŘřŚśŜŝŞşŠšſŢţŤťŦŧÙÚÛÜùúûüŨũŪūŬŭŮůŰűŲųŴŵÝýÿŶŷŸŹźŻżŽž"
+without_diacritics: "AAAAAAaaaaaaAaAaAaCcCcCcCcCcDdDdDdEEEEeeeeEeEeEeEeEeGgGgGgGgHhHhIIIIiiiiIiIiIiIiIiJjKkkLlLlLlLlLlNnNnNnNnnNnOOOOOOooooooOoOoOoRrRrRrSsSsSsSssTtTtTtUUUUuuuuUuUuUuUuUuUuWwYyyYyYZzZzZz"