xmet/lib/xmet/raob.py

419 lines
12 KiB
Python

import io
import re
import enum
import datetime
from xmet.util import each_chunk
from xmet.afos import RE_ID, RE_ISSUANCE, RE_PRODUCT
from xmet.sounding import Sounding, SoundingSample
CHUNK_SEP = "\x01"
CHUNK_STRIP_CHARS = "\x01\x03\x0a\x20"
def meters_second(knots: float) -> float:
return knots / 1.944
class RAOBReaderException(Exception):
...
class RAOBObs():
DATA_SOURCE = 'UCAR'
def __init__(self, kind: str):
self.kind: str = kind
self.tokens: list[str] = list()
def read(self, token: str):
self.tokens.append(token)
def parse_timestamp(self, token: str):
if token[0:2] == '//':
return None
day = int(token[0:2])
hour = int(token[2:4])
now = datetime.datetime.now(datetime.UTC)
return datetime.datetime(
year = now.year,
month = now.month,
day = day if day < 51 else day - 50
)
def parse_temp_dewpoint(self, token: str):
if token[2] == '/':
return {
'temp': None,
'dewpoint': None
}
if token[0:2] == '//':
temp = None
else:
tenths = int(token[2])
sign = 1 if tenths % 2 == 0 else -1
temp = sign * 0.1 * float(token[0:3])
if token[3:5] == '//':
dewpoint = None
else:
dda = int(token[3:5])
dd = dda * 0.1 if dda <= 50 else dda - 50
dewpoint = temp - dd
return {
'temp': temp,
'dewpoint': dewpoint
}
def parse_wind(self, token: str):
base_speed = 0
base_dir = 0
if token == '=':
return
if token[2] != '/':
value = int(token[2])
if value >= 5:
base_speed = 100 * (value % 5)
base_dir = 5
else:
base_speed = 100 * value
if token[0:2] == '//':
wind_dir = None
else:
wind_dir = meters_second(float(token[0:3]) + base_dir)
if token[3:5] == '//':
wind_speed = None
else:
wind_speed = meters_second(float(token[3:5]) + base_speed)
return {
'dir': wind_dir,
'speed': wind_speed
}
TTAA_PRESSURES = {
'00': 1000, '92': 925, '85': 850,
'70': 700, '50': 500, '40': 400,
'30': 300, '25': 250, '20': 200,
'15': 150, '10': 100,
}
def code_to_pressure(self, code: str):
if code in self.TTAA_PRESSURES:
return self.TTAA_PRESSURES[code]
return None
def calc_1000mb_height(self, value: float) -> float:
if value >= 500:
return 0 - (value - 500)
return value
def calc_850mb_height(self, value: float) -> float:
return 1000.0 + value
def calc_700mb_height(self, value: float) -> float:
if value >= 500:
return 2000.0 + value
else:
return 3000.0 + value
def calc_500mb_height(self, value: float) -> float:
return 10.0 * value
def calc_250mb_height(self, value: float) -> float:
if value >= 500:
return value * 10
else:
return 10.0 * (1000.0 + value)
def calc_100mb_height(self, value: float) -> float:
return 10.0 * (1000.0 + value)
def parse_height_pressure(self, token: str):
code = token[0:2]
num = token[2:5]
#
# Ignore tokens where height is not known.
#
if num == '///':
return None
#
# Ignore the tropopause or height of max wind velocity.
#
if code == '77' or code == '88':
return None
pressure = self.code_to_pressure(code)
if pressure is None:
return None
elif pressure == 1000:
height = self.calc_1000mb_height(float(num))
elif pressure <= 850 and pressure > 700:
height = self.calc_850mb_height(float(num))
elif pressure <= 700 and pressure > 500:
height = self.calc_700mb_height(float(num))
elif pressure <= 500 and pressure > 250:
height = self.calc_500mb_height(float(num))
elif pressure <= 250 and pressure > 100:
height = self.calc_250mb_height(float(num))
elif pressure <= 100:
height = self.calc_100mb_height(float(num))
else:
height = float(num)
return {
'pressure': pressure,
'height': height
}
PRESSURE_SIG = {
'11': True, '22': True, '33': True, '44': True, '55': True,
'66': True, '77': True, '88': True, '99': True
}
def parse_significant_pressure(self, token: str):
code, pressure = token[0:2], token[2:5]
if code in self.PRESSURE_SIG:
return {
'height': None,
'pressure': float(pressure)
}
def parse_surface_pressure(self, token: str):
code, pressure = token[0:2], token[2:5]
if code == '99':
return {
'height': None,
'pressure': float(pressure)
}
def parse_sample_tokens(self, tokens: list[str]) -> SoundingSample:
sample = SoundingSample()
if tokens[0][0:2] == '99':
sample.elapsed = 0
hp = self.parse_surface_pressure(tokens[0])
else:
hp = self.parse_height_pressure(tokens[0])
if hp is None:
return None
td = self.parse_temp_dewpoint(tokens[1])
wind = self.parse_wind(tokens[2])
sample.height = hp['height']
sample.pressure = hp['pressure']
sample.temp = td['temp'] if td is not None else None
sample.dewpoint = td['dewpoint'] if td is not None else None
sample.wind_dir = wind['dir'] if wind is not None else None
sample.wind_speed = wind['speed'] if wind is not None else None
sample.pressure_qa = ' '
sample.height_qa = ' '
sample.temp_qa = ' '
return sample
def parse_ttaa(self) -> Sounding:
#
# Return None if there is no height data up to 100mb.
#
if self.tokens[0][4] != '1':
return None
#
# Return None if there is no station identifier.
#
if self.tokens[1][0:3] == 'NIL':
return None
sample = self.parse_sample_tokens(self.tokens[2:5])
if sample is None:
return
timestamp = self.parse_timestamp(self.tokens[0])
sounding = Sounding()
sounding.samples = [sample]
sounding.station = self.tokens[1]
sounding.data_source_pressure = self.DATA_SOURCE
sounding.data_source_other = self.DATA_SOURCE
sounding.timestamp_observed = timestamp
sounding.timestamp_released = timestamp - datetime.timedelta(minutes=45)
for i in range(5, len(self.tokens), 3):
if len(self.tokens) < i+3 or self.tokens[i][-1] == '=':
break
#
# Stop parsing tokens at the tropopause.
#
if self.tokens[i][0:2] == '88':
break
sample = self.parse_sample_tokens(self.tokens[i:i+3])
if sample is None:
continue
sounding.samples.append(sample)
return sounding
class RAOBChunk():
def __init__(self,
wfo: str,
product: str,
tokens: list[str]):
self.wfo = wfo
self.product = product
self.tokens = tokens
def is_obs_start(self, token: str) -> bool:
return token == 'TTAA' or token == 'TTBB' \
or token == 'TTCC' or token == 'TTDD' \
or token == 'PPAA' or token == 'PPBB' \
or token == 'PPCC' or token == 'PPDD'
def each_obs(self):
obs = None
for token in self.tokens:
if self.is_obs_start(token):
if obs is not None:
yield obs
obs = RAOBObs(token)
elif obs is not None:
obs.read(token)
if obs is not None:
yield obs
def each_sounding(self):
for obs in self.each_obs():
if obs.kind == 'TTAA':
sounding = obs.parse_ttaa()
if sounding is None or len(sounding.samples) == 0:
continue
yield sounding
class RAOBReader():
"""
A reader for the global `Current.rawins` file provided by UCAR:
https://weather.rap.ucar.edu/data/upper/Current.rawins
The format is best documented here:
https://www.atmos.albany.edu/facstaff/ralazear/ATM211/Home_files/RAOB_Code_packet.pdf
"""
def __init__(self, fh: io.TextIOBase):
self.fh = fh
self.soundings = dict()
def parse_chunk(self, text: str) -> RAOBChunk:
meta = {
'wfo': None, # NWS forecast office
'product': None # NWS product code
}
line_index = 0
#
# Split each line in the text chunk. Not all chunks will have the
# same amount of metadata, so parse accordingly.
#
lines = list(map(lambda s: s.strip(), text.split("\n")))
#
# The `Current.rawins` feed from UCAR includes basic AFOS header
# information in the first two lines. Validate this. Note the first
# line is a sort of sequence number which has no public significance.
#
match = RE_ID.match(lines[0])
if match is None:
raise RAOBReaderException(f"First chunk line not 3-digit identifier ({lines[0]})")
else:
line_index += 1
#
# The `Current.rawins` feed from UCAR should also include a product
# issuance code indicating the WFO and validity time. This can also
# be validated.
#
match = RE_ISSUANCE.match(lines[1])
if match is None:
raise RAOBReaderException('Second chunk line not product issuance')
else:
meta['wfo'] = match['wfo']
line_index += 1
#
# Finally, sometimes, the `Current.rawins` feed has an AFOS header
# which indicates the product code followed by the three-character
# WFO code. Capture the product code purely for posterity.
#
match = RE_PRODUCT.match(lines[2])
if match is not None:
meta['product'] = match['product']
line_index += 1
#
# Split each whitespace-delimited column of each line into one big
# list of lines for the remainder of the current text chunk.
#
tokens = list()
for line in lines[line_index:]:
tokens.extend(re.split(r'\s+', line))
return RAOBChunk(meta['wfo'],
meta['product'],
tokens)
def each_chunk(self):
for text in each_chunk(self.fh, CHUNK_SEP, CHUNK_STRIP_CHARS):
yield self.parse_chunk(text)
def each_obs(self):
for chunk in self.each_chunk():
yield from chunk.each_obs()
def each_sounding(self):
for chunk in self.each_chunk():
yield from chunk.each_sounding()
@staticmethod
def each_sounding_from_fh(fh: io.TextIOBase):
reader = RAOBReader(fh)
yield from reader.each_sounding()
@staticmethod
def each_sounding_from_file(path: str):
with open(path, 'r') as fh:
yield from RAOBReader.each_sounding_from_fh(fh)