Skip to content

Commit

Permalink
Change in behaviour, WSI no longer estimated by default, arg now requ… (
Browse files Browse the repository at this point in the history 
#26)

* Change in behaviour, WSI no longer estimated by default, arg now required in call to transform / as_geojson.
Updated tests
Removal of remote ref in schema.

* Revert of JSON schema.

* Update test_bufr2geojson.py

---------

Co-authored-by: Tom Kralidis <[email protected]>
  • Loading branch information
@david-i-berry @tomkralidis
david-i-berry and tomkralidis authored Feb 6, 2024
1 parent 90b560b commit 5d0634d
Show file tree
Hide file tree
Showing 3 changed files with 69 additions and 43 deletions.
105 changes: 65 additions & 40 deletions bufr2geojson/__init__.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -417,14 +417,14 @@ def get_time(self) -> str:

return time_

def get_wsi(self) -> str:
def get_wsi(self, guess_wsi: bool = False) -> str:
"""
Function returns WIGOS station ID as string
:returns: WIGOS station ID.
"""

return self.get_identification()["wsi"]
return self.get_identification(guess_wsi)["wsi"]

def get_tsi(self) -> str:
"""
Expand All @@ -435,13 +435,16 @@ def get_tsi(self) -> str:

return self.get_identification()["tsi"]

def get_identification(self) -> dict:
def get_identification(self, guess_wsi: bool = False) -> dict:
"""
Function extracts identification information from qualifiers.
:returns: dictionary containing any class 01 qualifiers and WSI as dict. # noqa
"""

# default WSI value
wsi = None

# see https://library.wmo.int/doc_num.php?explnum_id=11021
# page 19 for allocation of WSI if not set
# check to see what identification we have
Expand All @@ -467,27 +470,35 @@ def get_identification(self) -> dict:
if all(x in self.qualifiers["01"] for x in _types): # noqa
block = self.get_qualifier("01", "block_number")
station = self.get_qualifier("01", "station_number")
wsi_series = 0
wsi_issuer = 20000
wsi_number = 0
wsi_local = strip2(f"{block:02d}{station:03d}")
tsi = strip2(f"{block:02d}{station:03d}")
if guess_wsi:
wsi_series = 0
wsi_issuer = 20000
wsi_number = 0
wsi_local = tsi
wsi = f"{wsi_series}-{wsi_issuer}-{wsi_number}-{wsi_local}"

return {
"wsi": f"{wsi_series}-{wsi_issuer}-{wsi_number}-{wsi_local}",
"tsi": wsi_local,
"wsi": wsi,
"tsi": tsi,
"type": "{}_and_{}".format(*_types)
}

# ship or mobile land station identifier (001011)
_type = "ship_or_mobile_land_station_identifier"
if _type in self.qualifiers["01"]:
callsign = self.get_qualifier("01", _type)
wsi_series = 0
wsi_issuer = 20004
wsi_number = 0
wsi_local = strip2(callsign)
tsi = strip2(callsign)
if guess_wsi:
wsi_series = 0
wsi_issuer = 20004
wsi_number = 0
wsi_local = tsi
wsi = f"{wsi_series}-{wsi_issuer}-{wsi_number}-{wsi_local}"

return {
"wsi": f"{wsi_series}-{wsi_issuer}-{wsi_number}-{wsi_local}",
"tsi": wsi_local,
"wsi": wsi,
"tsi": tsi,
"type": _type
}

Expand All @@ -499,13 +510,17 @@ def get_identification(self) -> dict:
wmo_region = self.get_qualifier("region_number")
wmo_subregion = self.get_qualifier("wmo_region_sub_area")
wmo_number = self.get_qualifier("buoy_or_platform_identifier")
wsi_series = 0
wsi_issuer = 20002
wsi_number = 0
wsi_local = strip2(f"{wmo_region:01d}{wmo_subregion:01d}{wmo_number:05d}") # noqa
tsi = strip2(f"{wmo_region:01d}{wmo_subregion:01d}{wmo_number:05d}") # noqa
if guess_wsi:
wsi_series = 0
wsi_issuer = 20002
wsi_number = 0
wsi_local = tsi # noqa
wsi = f"{wsi_series}-{wsi_issuer}-{wsi_number}-{wsi_local}"

return {
"wsi": f"{wsi_series}-{wsi_issuer}-{wsi_number}-{wsi_local}",
"tsi": wsi_local,
"wsi": wsi,
"tsi": tsi,
"type": "5_digit_marine_observing_platform_identifier"
}

Expand All @@ -514,13 +529,17 @@ def get_identification(self) -> dict:
_type = "stationary_buoy_platform_identifier_e_g_c_man_buoys"
if _type in self.qualifiers["01"]:
id_ = self.get_qualifier("01", _type)
wsi_series = 0
wsi_issuer = 20002
wsi_number = 0
wsi_local = strip2(id_)
tsi = strip2(id_)
if guess_wsi:
wsi_series = 0
wsi_issuer = 20002
wsi_number = 0
wsi_local = tsi
wsi = f"{wsi_series}-{wsi_issuer}-{wsi_number}-{wsi_local}"

return {
"wsi": f"{wsi_series}-{wsi_issuer}-{wsi_number}-{wsi_local}",
"tsi": wsi_local,
"wsi": wsi,
"tsi": tsi,
"type": _type
}

Expand All @@ -529,13 +548,17 @@ def get_identification(self) -> dict:
_type = "marine_observing_platform_identifier"
if _type in self.qualifiers["01"]:
id_ = self.get_qualifier("01", _type)
wsi_series = 0
wsi_issuer = 20002
wsi_number = 0
wsi_local = strip2(id_)
tsi = strip2(id_)
if guess_wsi:
wsi_series = 0
wsi_issuer = 20002
wsi_number = 0
wsi_local = tsi
wsi = f"{wsi_series}-{wsi_issuer}-{wsi_number}-{wsi_local}"

return {
"wsi": f"{wsi_series}-{wsi_issuer}-{wsi_number}-{wsi_local}",
"tsi": wsi_local,
"wsi": wsi,
"tsi": tsi,
"type": "7_digit_marine_observing_platform_identifier"
}

Expand Down Expand Up @@ -576,7 +599,7 @@ def get_code_value(self, fxxyyy: str, code: int) -> str:
return decoded

def as_geojson(self, bufr_handle: int, id: str,
serialize: bool = False) -> dict:
serialize: bool = False, guess_wsi: bool = False) -> dict:
"""
Function to return GeoJSON representation of BUFR message
Expand Down Expand Up @@ -729,7 +752,7 @@ def as_geojson(self, bufr_handle: int, id: str,
attributes, append)
continue
if value is not None:
self.get_identification()
# self.get_identification()
metadata = self.get_qualifiers()
metadata_hash = hashlib.md5(json.dumps(metadata).encode("utf-8")).hexdigest() # noqa
md = {
Expand All @@ -738,7 +761,7 @@ def as_geojson(self, bufr_handle: int, id: str,
}
for idx in range(len(metadata)):
md["metadata"].append(metadata[idx])
wsi = self.get_wsi()
wsi = self.get_wsi(guess_wsi)
feature_id = f"WIGOS_{wsi}_{characteristic_date}T{characteristic_time}" # noqa
feature_id = f"{feature_id}{id}-{index}"
phenomenon_time = self.get_time()
Expand Down Expand Up @@ -786,12 +809,14 @@ def as_geojson(self, bufr_handle: int, id: str,
return data


def transform(data: bytes, serialize: bool = False) -> Iterator[dict]:
def transform(data: bytes, serialize: bool = False,
guess_wsi: bool = False) -> Iterator[dict]:
"""
Main transformation
:param data: byte string of BUFR data
:param serialize: whether to return as JSON string (default is False)
:param guess_wsi: whether to 'guess' WSI based on TSI and allocaiotn rules
:returns: `generator` of GeoJSON features
"""
Expand All @@ -803,8 +828,7 @@ def transform(data: bytes, serialize: bool = False) -> Iterator[dict]:
with open(tmp.name, 'wb') as f:
f.write(data)

# check data type, only in situ supported
# not yet implemented
# check data type, only in situ supported (not yet implemented)
# split subsets into individual messages and process
imsg = 0
messages_remaining = True
Expand Down Expand Up @@ -846,7 +870,8 @@ def transform(data: bytes, serialize: bool = False) -> Iterator[dict]:
tag = f"-{idx}"
try:
data = parser.as_geojson(single_subset, id=tag,
serialize=serialize)
serialize=serialize,
guess_wsi=guess_wsi) # noqa

except Exception as e:
LOGGER.error("Error parsing BUFR to GeoJSON, no data written") # noqa
Expand Down
2 changes: 1 addition & 1 deletion bufr2geojson/resources/schemas/wmo-om-profile-geojson.yaml
View file
Original file line number Diff line number Diff line change
Expand Up @@ -99,4 +99,4 @@ examples:
type: application/geo+json
hreflang: en
title: 'Trierer Strasse 70, 53115 Bonn'
length: 0
length: 0
5 changes: 3 additions & 2 deletions tests/test_bufr2geojson.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -128,7 +128,7 @@ def geojson_output():


def test_multi(multimsg_bufr):
results = transform(multimsg_bufr)
results = transform(multimsg_bufr, guess_wsi=True)
# count number of geojsons
icount = 0
for res in results:
Expand All @@ -140,7 +140,8 @@ def test_multi(multimsg_bufr):
def test_transform(geojson_schema, geojson_output):
test_bufr_file = 'A_ISIA21EIDB202100_C_EDZW_20220320210902_11839953.bin'
with open(test_bufr_file, 'rb') as fh:
messages1, messages2 = itertools.tee(transform(fh.read()))
messages1, messages2 = itertools.tee(transform(fh.read(),
guess_wsi=True))

# validate against JSON Schema
for message in messages1:
Expand Down

0 comments on commit 5d0634d

Please sign in to comment.