Module: Tina4Ultipa::Codec

Defined in:
lib/tina4_ultipa/codec.rb

Defined Under Namespace

Classes: Reader

Constant Summary collapse

NUMERIC =

PropertyType (as the enum symbol google-protobuf hands back) -> pack/unpack directive for the fixed-width little-endian numeric types.

{
  PROPERTY_TYPE_INT32:  "l<",
  PROPERTY_TYPE_UINT32: "L<",
  PROPERTY_TYPE_INT64:  "q<",
  PROPERTY_TYPE_UINT64: "Q<",
  PROPERTY_TYPE_FLOAT:  "e",
  PROPERTY_TYPE_DOUBLE: "E",
}.freeze
STRINGY =
%i[PROPERTY_TYPE_STRING PROPERTY_TYPE_TEXT].freeze
NULLISH =
%i[PROPERTY_TYPE_NULL PROPERTY_TYPE_UNSET].freeze
LISTY =
%i[PROPERTY_TYPE_LIST PROPERTY_TYPE_SET].freeze
MAPPY =
%i[PROPERTY_TYPE_MAP PROPERTY_TYPE_RECORD].freeze
LOCAL_DT =
%i[PROPERTY_TYPE_DATETIME PROPERTY_TYPE_LOCAL_DATETIME].freeze
INT_TO_SYM =

A google-protobuf enum field usually reads back as a Symbol, but falls back to the raw Integer for values it does not know. Normalise to the symbol.

{
  0 => :PROPERTY_TYPE_UNSET, 1 => :PROPERTY_TYPE_INT32, 2 => :PROPERTY_TYPE_UINT32,
  3 => :PROPERTY_TYPE_INT64, 4 => :PROPERTY_TYPE_UINT64, 5 => :PROPERTY_TYPE_FLOAT,
  6 => :PROPERTY_TYPE_DOUBLE, 7 => :PROPERTY_TYPE_STRING,
  8 => :PROPERTY_TYPE_DATETIME, 9 => :PROPERTY_TYPE_TIMESTAMP, 10 => :PROPERTY_TYPE_TEXT,
  11 => :PROPERTY_TYPE_BLOB, 12 => :PROPERTY_TYPE_POINT, 13 => :PROPERTY_TYPE_DECIMAL,
  14 => :PROPERTY_TYPE_LIST, 15 => :PROPERTY_TYPE_SET, 16 => :PROPERTY_TYPE_MAP,
  17 => :PROPERTY_TYPE_NULL, 18 => :PROPERTY_TYPE_BOOL,
  19 => :PROPERTY_TYPE_LOCAL_DATETIME, 20 => :PROPERTY_TYPE_ZONED_DATETIME,
  21 => :PROPERTY_TYPE_DATE, 22 => :PROPERTY_TYPE_ZONED_TIME,
  23 => :PROPERTY_TYPE_LOCAL_TIME, 24 => :PROPERTY_TYPE_YEAR_TO_MONTH,
  25 => :PROPERTY_TYPE_DAY_TO_SECOND, 26 => :PROPERTY_TYPE_RECORD,
  27 => :PROPERTY_TYPE_POINT3D, 28 => :PROPERTY_TYPE_VECTOR,
  29 => :PROPERTY_TYPE_TABLE, 30 => :PROPERTY_TYPE_PATH, 31 => :PROPERTY_TYPE_ERROR,
  32 => :PROPERTY_TYPE_NODE, 33 => :PROPERTY_TYPE_EDGE
}.freeze

Class Method Summary collapse

Class Method Details

.date(data) ⇒ Object

── temporal / spatial helpers ──────────────────────────────────────



59
60
61
62
63
64
# File 'lib/tina4_ultipa/codec.rb', line 59

def date(data)
  return nil if data.nil? || data.bytesize < 4

  y = data.byteslice(0, 2).unpack1("s<") # signed year (BCE round-trips)
  format("%04d-%02d-%02d", y, data.getbyte(2), data.getbyte(3))
end

.datetime_frac(base, nanos) ⇒ Object



94
95
96
97
98
# File 'lib/tina4_ultipa/codec.rb', line 94

def datetime_frac(base, nanos)
  return base if nanos.nil? || nanos.zero?

  base + format(".%09d", nanos).sub(/0+\z/, "").sub(/\.\z/, "")
end

.day_to_second(data) ⇒ Object



184
185
186
187
188
# File 'lib/tina4_ultipa/codec.rb', line 184

def day_to_second(data)
  return { "seconds" => 0, "nanoseconds" => 0 } if data.nil? || data.bytesize < 12

  { "seconds" => data.unpack1("q<"), "nanoseconds" => data.unpack1("@8 L<") }
end

.decode(tv) ⇒ Object

gqldb.TypedValue -> a Ruby value (nil for null; Hash for node/edge/...).



362
363
364
365
366
# File 'lib/tina4_ultipa/codec.rb', line 362

def decode(tv)
  return nil if tv.is_null

  scalar(tv.type, tv.data)
end

.edge(data) ⇒ Object



289
290
291
292
293
294
295
296
297
298
# File 'lib/tina4_ultipa/codec.rb', line 289

def edge(data)
  r = Reader.new(data)
  eid = r.str
  etype = r.str
  src = r.str
  dst = r.str
  props = r.props
  { "_kind" => "edge", "id" => eid, "type" => etype, "from" => src, "to" => dst,
    "properties" => props, "uuid" => trailer(r) }
end

.encode(value) ⇒ Object

A Ruby value -> gqldb.TypedValue, for query parameters.



369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
# File 'lib/tina4_ultipa/codec.rb', line 369

def encode(value)
  case value
  when nil
    Gqldb::TypedValue.new(type: :PROPERTY_TYPE_NULL, is_null: true)
  when true, false
    Gqldb::TypedValue.new(type: :PROPERTY_TYPE_BOOL, data: (value ? "\x01" : "\x00").b)
  when Integer
    Gqldb::TypedValue.new(type: :PROPERTY_TYPE_INT64, data: [value].pack("q<"))
  when Float
    Gqldb::TypedValue.new(type: :PROPERTY_TYPE_DOUBLE, data: [value].pack("E"))
  when String
    # A binary (ASCII-8BIT) String is the Ruby analogue of Python `bytes`:
    # send it as a BLOB so raw bytes round-trip intact. UTF-8/other -> STRING.
    if value.encoding == Encoding::ASCII_8BIT
      Gqldb::TypedValue.new(type: :PROPERTY_TYPE_BLOB, data: value)
    else
      Gqldb::TypedValue.new(type: :PROPERTY_TYPE_STRING, data: value.b)
    end
  else
    # Anything else -> UTF-8 bytes.
    Gqldb::TypedValue.new(type: :PROPERTY_TYPE_STRING, data: value.to_s.b)
  end
end

.error(data) ⇒ Object



190
191
192
193
194
195
196
# File 'lib/tina4_ultipa/codec.rb', line 190

def error(data)
  s = (data || "".b).dup.force_encoding("UTF-8")
  obj = JSON.parse(s)
  { "_kind" => "error", "code" => (obj["code"] || 0), "message" => (obj["message"] || "") }
rescue StandardError
  { "_kind" => "error", "code" => 0, "message" => s }
end

.list(data) ⇒ Object

── composites ──────────────────────────────────────────────────────



263
264
265
266
# File 'lib/tina4_ultipa/codec.rb', line 263

def list(data)
  r = Reader.new(data)
  Array.new(r.u16) { r.tve }
end

.local_datetime(data) ⇒ Object



100
101
102
103
104
105
106
107
108
109
# File 'lib/tina4_ultipa/codec.rb', line 100

def local_datetime(data)
  if !data.nil? && data.bytesize >= 11
    y = data.unpack1("s<")
    nanos = data.unpack1("@7 L<")
    base = format("%04d-%02d-%02d %02d:%02d:%02d", y, data.getbyte(2), data.getbyte(3),
                  data.getbyte(4), data.getbyte(5), data.getbyte(6))
    return datetime_frac(base, nanos)
  end
  unix(data) # legacy 8-byte unix fallback
end

.local_time(data) ⇒ Object



73
74
75
76
77
78
# File 'lib/tina4_ultipa/codec.rb', line 73

def local_time(data)
  return nil if data.nil? || data.bytesize < 8

  nanos = data.unpack1("@4 L<")
  time_str(data.getbyte(0), data.getbyte(1), data.getbyte(2), nanos)
end

.node(data) ⇒ Object



280
281
282
283
284
285
286
287
# File 'lib/tina4_ultipa/codec.rb', line 280

def node(data)
  r = Reader.new(data)
  nid = r.str
  labels = Array.new(r.u16) { r.str }
  props = r.props
  { "_kind" => "node", "id" => nid, "labels" => labels, "properties" => props,
    "uuid" => trailer(r) }
end

.offset(minutes) ⇒ Object



88
89
90
91
92
# File 'lib/tina4_ultipa/codec.rb', line 88

def offset(minutes)
  sign = minutes >= 0 ? "+" : "-"
  m = minutes.abs
  format("%s%02d:%02d", sign, m / 60, m % 60)
end

.path(data) ⇒ Object



300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
# File 'lib/tina4_ultipa/codec.rb', line 300

def path(data)
  r = Reader.new(data)
  nodes = Array.new(r.u16) do
    n = r.u32
    v = node(r.d.byteslice(r.p, n))
    r.p += n
    v
  end
  edges = Array.new(r.u16) do
    n = r.u32
    v = edge(r.d.byteslice(r.p, n))
    r.p += n
    v
  end
  { "_kind" => "path", "nodes" => nodes, "edges" => edges }
end

.point(data) ⇒ Object



150
151
152
153
154
155
156
157
# File 'lib/tina4_ultipa/codec.rb', line 150

def point(data)
  return nil if data.nil? || data.bytesize < 16

  x = data.unpack1("E")
  y = data.unpack1("@8 E")
  srid = data.bytesize >= 20 ? data.unpack1("@16 L<") : 7203
  { "_kind" => "point", "x" => x, "y" => y, "srid" => srid }
end

.point3d(data) ⇒ Object



159
160
161
162
163
164
165
166
167
# File 'lib/tina4_ultipa/codec.rb', line 159

def point3d(data)
  return nil if data.nil? || data.bytesize < 24

  x = data.unpack1("E")
  y = data.unpack1("@8 E")
  z = data.unpack1("@16 E")
  srid = data.bytesize >= 28 ? data.unpack1("@24 L<") : 9757
  { "_kind" => "point3d", "x" => x, "y" => y, "z" => z, "srid" => srid }
end

.scalar(ptype, data) ⇒ Object

Decode a raw property/value payload of a given PropertyType. Used both for the top-level TypedValue and for each value inside a composite.



321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
# File 'lib/tina4_ultipa/codec.rb', line 321

def scalar(ptype, data)
  t = ptype
  t = INT_TO_SYM[t] || t if t.is_a?(Integer)
  return nil if NULLISH.include?(t)

  fmt = NUMERIC[t]
  unless fmt.nil?
    return nil if data.nil? || data.empty?

    return data.unpack1(fmt)
  end
  return data.dup.force_encoding("UTF-8") if STRINGY.include?(t)
  return !(data.nil? || data.empty? || data.getbyte(0).zero?) if t == :PROPERTY_TYPE_BOOL
  # BLOB: raw bytes as a binary string.
  return (data || "".b).dup.force_encoding("BINARY") if t == :PROPERTY_TYPE_BLOB
  # DECIMAL is a precision-preserving UTF-8 string: return it verbatim, never to_f.
  return data.dup.force_encoding("UTF-8") if t == :PROPERTY_TYPE_DECIMAL
  return date(data) if t == :PROPERTY_TYPE_DATE
  return local_datetime(data) if LOCAL_DT.include?(t)
  return zoned_datetime(data) if t == :PROPERTY_TYPE_ZONED_DATETIME
  return timestamp(data) if t == :PROPERTY_TYPE_TIMESTAMP
  return local_time(data) if t == :PROPERTY_TYPE_LOCAL_TIME
  return zoned_time(data) if t == :PROPERTY_TYPE_ZONED_TIME
  return year_to_month(data) if t == :PROPERTY_TYPE_YEAR_TO_MONTH
  return day_to_second(data) if t == :PROPERTY_TYPE_DAY_TO_SECOND
  return point(data) if t == :PROPERTY_TYPE_POINT
  return point3d(data) if t == :PROPERTY_TYPE_POINT3D
  return vector(data) if t == :PROPERTY_TYPE_VECTOR
  return list(data) if LISTY.include?(t)
  return Reader.new(data).props if MAPPY.include?(t)
  return table(data) if t == :PROPERTY_TYPE_TABLE
  return node(data) if t == :PROPERTY_TYPE_NODE
  return edge(data) if t == :PROPERTY_TYPE_EDGE
  return path(data) if t == :PROPERTY_TYPE_PATH
  return error(data) if t == :PROPERTY_TYPE_ERROR

  # Unmapped/future types: hand back the raw bytes rather than lose or guess.
  data
end

.table(data) ⇒ Object



268
269
270
271
272
273
# File 'lib/tina4_ultipa/codec.rb', line 268

def table(data)
  r = Reader.new(data)
  cols = Array.new(r.u16) { r.str }
  rows = Array.new(r.u16) { Array.new(r.u16) { r.tve } }
  { "_kind" => "table", "columns" => cols, "rows" => rows }
end

.time_str(hour, minute, sec, nanos) ⇒ Object



66
67
68
69
70
71
# File 'lib/tina4_ultipa/codec.rb', line 66

def time_str(hour, minute, sec, nanos)
  base = format("%02d:%02d:%02d", hour, minute, sec)
  return base if nanos.nil? || nanos.zero?

  base + format(".%09d", nanos).sub(/0+\z/, "").sub(/\.\z/, "")
end

.timestamp(data) ⇒ Object



123
124
125
126
127
128
129
130
131
132
133
134
135
# File 'lib/tina4_ultipa/codec.rb', line 123

def timestamp(data)
  if !data.nil? && data.bytesize >= 12
    secs = data.unpack1("q<")
    nanos = data.unpack1("@8 L<")
    begin
      base = Time.at(secs).utc.strftime("%Y-%m-%d %H:%M:%S")
      return datetime_frac(base, nanos)
    rescue StandardError, RangeError
      return secs
    end
  end
  unix(data)
end

.trailer(reader) ⇒ Object

Optional 8-byte little-endian uint64 internal-id trailer (gqldb 6.1.147+).



276
277
278
# File 'lib/tina4_ultipa/codec.rb', line 276

def trailer(reader)
  reader.p + 8 <= reader.d.bytesize ? reader.u64.to_s : ""
end

.unix(data) ⇒ Object

Legacy 8-byte unix time (seconds/ms/us) -> 'YYYY-MM-DD HH:MM:SS' (UTC).



138
139
140
141
142
143
144
145
146
147
148
# File 'lib/tina4_ultipa/codec.rb', line 138

def unix(data)
  return nil if data.nil? || data.bytesize < 8

  v = data.byteslice(0, 8).unpack1("Q<")
  secs = v > 1e14 ? v / 1e6 : (v > 1e11 ? v / 1e3 : v)
  begin
    Time.at(secs).utc.strftime("%Y-%m-%d %H:%M:%S")
  rescue StandardError, RangeError
    v
  end
end

.vector(data) ⇒ Object



169
170
171
172
173
174
175
176
# File 'lib/tina4_ultipa/codec.rb', line 169

def vector(data)
  return [] if data.nil? || data.bytesize < 4

  dim = data.unpack1("L<")
  return [] if data.bytesize < 4 + dim * 4

  Array.new(dim) { |i| data.unpack1("@#{4 + i * 4} e") }
end

.year_to_month(data) ⇒ Object



178
179
180
181
182
# File 'lib/tina4_ultipa/codec.rb', line 178

def year_to_month(data)
  return { "months" => 0 } if data.nil? || data.bytesize < 4

  { "months" => data.unpack1("l<") }
end

.zoned_datetime(data) ⇒ Object



111
112
113
114
115
116
117
118
119
120
121
# File 'lib/tina4_ultipa/codec.rb', line 111

def zoned_datetime(data)
  if !data.nil? && data.bytesize >= 13
    y = data.unpack1("s<")
    nanos = data.unpack1("@7 L<")
    off = data.unpack1("@11 s<")
    base = format("%04d-%02d-%02d %02d:%02d:%02d", y, data.getbyte(2), data.getbyte(3),
                  data.getbyte(4), data.getbyte(5), data.getbyte(6))
    return datetime_frac(base, nanos) + offset(off)
  end
  unix(data)
end

.zoned_time(data) ⇒ Object



80
81
82
83
84
85
86
# File 'lib/tina4_ultipa/codec.rb', line 80

def zoned_time(data)
  return nil if data.nil? || data.bytesize < 10

  nanos = data.unpack1("@4 L<")
  off = data.unpack1("@8 s<")
  time_str(data.getbyte(0), data.getbyte(1), data.getbyte(2), nanos) + offset(off)
end