2025-01-30 15:53:26 +02:00
|
|
|
package grib
|
|
|
|
|
|
|
|
|
|
import cats.effect._
|
|
|
|
|
import cats.effect.unsafe.implicits.global
|
|
|
|
|
import fs2.io.file.{Files, Path}
|
2025-01-30 22:53:00 +02:00
|
|
|
import io.circe.generic.auto._
|
|
|
|
|
import io.circe.syntax._
|
2025-01-30 15:53:26 +02:00
|
|
|
|
|
|
|
|
import java.nio.ByteBuffer
|
2025-01-30 22:44:25 +02:00
|
|
|
import java.time.temporal.ChronoUnit
|
|
|
|
|
import java.time.{ZoneOffset, ZonedDateTime}
|
2025-01-30 15:53:26 +02:00
|
|
|
|
|
|
|
|
object GribParser {
|
2025-01-30 22:44:25 +02:00
|
|
|
def main(args: Array[String]): Unit = {
|
|
|
|
|
val fileName = "data/HARMONIE_DINI_SF_2025-01-24T030000Z_2025-01-26T010000Z.grib"
|
|
|
|
|
val path = Path(fileName)
|
|
|
|
|
|
|
|
|
|
val program = for {
|
2025-01-30 22:53:00 +02:00
|
|
|
gribList <- parseFile(path)
|
|
|
|
|
json = gribList.asJson
|
|
|
|
|
_ <- IO.println(json.spaces2)
|
2025-01-30 22:44:25 +02:00
|
|
|
} yield ()
|
|
|
|
|
|
|
|
|
|
program.unsafeRunSync()
|
|
|
|
|
}
|
2025-01-30 15:53:26 +02:00
|
|
|
def parseFile(path: Path): IO[List[Grib]] = {
|
|
|
|
|
def loop(ptr: Long, acc: List[Grib]): IO[List[Grib]] = {
|
|
|
|
|
for {
|
|
|
|
|
fileSize <- Files[IO].size(path)
|
|
|
|
|
result <- if (ptr >= fileSize) {
|
|
|
|
|
IO.pure(acc.reverse)
|
|
|
|
|
} else {
|
|
|
|
|
for {
|
|
|
|
|
grib <- parseGrib(path, ptr)
|
|
|
|
|
nextResult <- loop(ptr + grib.length, grib :: acc)
|
|
|
|
|
} yield nextResult
|
|
|
|
|
}
|
|
|
|
|
} yield result
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
loop(0L, List.empty)
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
def parseGrib(path: Path, ptr: Long): IO[Grib] = {
|
|
|
|
|
for {
|
2025-01-30 22:44:25 +02:00
|
|
|
res0 <- parse0(path, ptr)
|
|
|
|
|
(version, discipline, gribLength, len0) = res0
|
|
|
|
|
ptr1 = ptr + len0
|
|
|
|
|
res1 <- parse1(path, ptr1)
|
|
|
|
|
(referenceTime, len1) = res1
|
|
|
|
|
ptr3 = ptr1 + len1
|
|
|
|
|
res3 <- parse3(path, ptr3)
|
|
|
|
|
(grid, len3) = res3
|
|
|
|
|
ptr4 = ptr3 + len3
|
|
|
|
|
res4 <- parse4(path, ptr4, discipline, referenceTime)
|
|
|
|
|
(meteo, time, len4) = res4
|
|
|
|
|
// _ <- IO.println(time)
|
|
|
|
|
ptr5 = ptr4 + len4
|
|
|
|
|
res5 <- parse5(path, ptr5)
|
|
|
|
|
(conversion, bitsPerDataPoint, len5) = res5
|
|
|
|
|
ptr6 = ptr5 + len5
|
|
|
|
|
len6 <- parse6(path, ptr6)
|
|
|
|
|
sections = List(
|
|
|
|
|
GribSection(0, ptr, len0),
|
|
|
|
|
GribSection(1, ptr1, len1),
|
|
|
|
|
GribSection(3, ptr3, len3),
|
|
|
|
|
GribSection(4, ptr4, len4),
|
|
|
|
|
GribSection(5, ptr5, len5),
|
|
|
|
|
GribSection(6, ptr6, len6),
|
|
|
|
|
)
|
|
|
|
|
grib = Grib(version, gribLength, grid, meteo, time, conversion, bitsPerDataPoint, sections)
|
|
|
|
|
} yield grib
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
def parse0(path: Path, ptr: Long): IO[(Int, Int, Long, Int)] = {
|
|
|
|
|
val length = 16
|
|
|
|
|
for {
|
|
|
|
|
bytes <- readBytes(path, ptr, length)
|
|
|
|
|
identifier = new String(bytes.take(4), "ASCII") // GRIB
|
2025-01-30 15:53:26 +02:00
|
|
|
discipline = bytes(6)
|
|
|
|
|
version = bytes(7)
|
2025-01-30 22:44:25 +02:00
|
|
|
gribLength = ByteBuffer.wrap(bytes.slice(8, 16)).getLong
|
|
|
|
|
} yield (version, discipline, gribLength, length)
|
2025-01-30 15:53:26 +02:00
|
|
|
}
|
|
|
|
|
|
2025-01-30 22:44:25 +02:00
|
|
|
def parse1(path: Path, ptr: Long): IO[(ZonedDateTime, Int)] = {
|
|
|
|
|
for {
|
|
|
|
|
bytes <- readBytes(path, ptr, 64)
|
|
|
|
|
length = ByteBuffer.wrap(bytes.slice(0, 4)).getInt
|
|
|
|
|
year = ByteBuffer.wrap(bytes.slice(12, 14)).getShort
|
|
|
|
|
month = bytes(14)
|
|
|
|
|
day = bytes(15)
|
|
|
|
|
hour = bytes(16)
|
|
|
|
|
minute = bytes(17)
|
|
|
|
|
second = bytes(18)
|
|
|
|
|
referenceTime = ZonedDateTime.of(
|
|
|
|
|
year, month, day,
|
|
|
|
|
hour, minute, second, 0, // last 0 is nanos
|
|
|
|
|
ZoneOffset.UTC // This is what 'Z' represents - UTC/Zero offset
|
|
|
|
|
)
|
|
|
|
|
// result = (f"$year%04d-$month%02d-$day%02dT$hour%02d:$minute%02d:$second%02dZ", length)
|
|
|
|
|
} yield (referenceTime, length)
|
|
|
|
|
}
|
2025-01-30 15:53:26 +02:00
|
|
|
|
2025-01-30 22:44:25 +02:00
|
|
|
def parse3(path: Path, ptr: Long): IO[(GribGrid, Int)] = {
|
|
|
|
|
for {
|
|
|
|
|
bytes <- readBytes(path, ptr, 64)
|
|
|
|
|
length = ByteBuffer.wrap(bytes.slice(0, 4)).getInt
|
|
|
|
|
template = ByteBuffer.wrap(bytes.slice(12, 14)).getShort
|
|
|
|
|
cols = ByteBuffer.wrap(bytes.slice(30, 34)).getInt
|
|
|
|
|
rows = ByteBuffer.wrap(bytes.slice(34, 38)).getInt
|
|
|
|
|
} yield (GribGrid(template, cols, rows), length)
|
2025-01-30 15:53:26 +02:00
|
|
|
}
|
|
|
|
|
|
2025-01-30 22:44:25 +02:00
|
|
|
def parse4(path: Path, ptr: Long, discipline: Int, referenceTime: ZonedDateTime): IO[(MeteoParam, GribTime, Int)] = {
|
|
|
|
|
for {
|
|
|
|
|
bytes <- readBytes(path, ptr, 64)
|
|
|
|
|
length = ByteBuffer.wrap(bytes.slice(0, 4)).getInt
|
|
|
|
|
category = bytes(9) & 0xFF // to uint
|
|
|
|
|
product = bytes(10) & 0xFF
|
|
|
|
|
subtype = if (bytes(8) == 1) "now" else "period"
|
|
|
|
|
levelType = bytes(22) & 0xFF
|
|
|
|
|
levelValue = bytes(27) & 0xFF
|
|
|
|
|
meteoParam = MeteoParam(discipline, category, product, subtype, levelType, levelValue)
|
|
|
|
|
forecastTime = if (bytes(8) == 1) {
|
|
|
|
|
val leadTime = ByteBuffer.wrap(bytes.slice(20, 22)).getShort
|
|
|
|
|
referenceTime.toInstant.plus(leadTime, ChronoUnit.HOURS)
|
|
|
|
|
} else {
|
|
|
|
|
val year = ByteBuffer.wrap(bytes.slice(37, 39)).getShort
|
|
|
|
|
val month = bytes(39)
|
|
|
|
|
val day = bytes(40)
|
|
|
|
|
val hour = bytes(41)
|
|
|
|
|
val minute = bytes(42)
|
|
|
|
|
val second = bytes(43)
|
|
|
|
|
ZonedDateTime.of(
|
|
|
|
|
year, month, day,
|
|
|
|
|
hour, minute, second, 0, // last 0 is nanos
|
|
|
|
|
ZoneOffset.UTC // This is what 'Z' represents - UTC/Zero offset
|
|
|
|
|
).toInstant
|
|
|
|
|
}
|
|
|
|
|
time = GribTime(referenceTime.toString, forecastTime.toString)
|
|
|
|
|
} yield (meteoParam, time, length)
|
|
|
|
|
}
|
2025-01-30 15:53:26 +02:00
|
|
|
|
2025-01-30 22:44:25 +02:00
|
|
|
def parse5(path: Path, ptr: Long): IO[(MeteoConversion, Int, Int)] = {
|
|
|
|
|
for {
|
|
|
|
|
bytes <- readBytes(path, ptr, 64)
|
|
|
|
|
length = ByteBuffer.wrap(bytes.slice(0, 4)).getInt
|
|
|
|
|
bitsPerDataPoint = bytes(19).toInt
|
|
|
|
|
reference = ByteBuffer.wrap(bytes.slice(11, 15)).getInt
|
|
|
|
|
binaryScale = -(bytes(16) & 0xFF)
|
|
|
|
|
decimalScale = bytes(18) & 0xFF
|
|
|
|
|
conversion = MeteoConversion(reference, binaryScale, decimalScale)
|
|
|
|
|
} yield (conversion, bitsPerDataPoint, length)
|
|
|
|
|
}
|
2025-01-30 15:53:26 +02:00
|
|
|
|
2025-01-30 22:44:25 +02:00
|
|
|
def parse6(path: Path, ptr: Long): IO[Int] = {
|
|
|
|
|
for {
|
|
|
|
|
bytes <- readBytes(path, ptr, 64)
|
|
|
|
|
length = ByteBuffer.wrap(bytes.slice(0, 4)).getInt
|
|
|
|
|
} yield length
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
def readBytes(path: Path, ptr: Long, len: Int): IO[Array[Byte]] = {
|
|
|
|
|
Files[IO].readRange(path, 1024, ptr, ptr + len)
|
|
|
|
|
.compile
|
|
|
|
|
.to(Array)
|
2025-01-30 15:53:26 +02:00
|
|
|
}
|
|
|
|
|
}
|