fix(2-B): normalize WFIGS field formats

WFIGS returns ISO 3166-2 state codes (US-MT) and 2-letter incident
type codes (WF, RX). Normalize at parse boundary:

- normalize_state: strips US- prefix (US-MT -> MT)
- normalize_incident_type: maps codes to names (WF -> wildfire)

Fixes:
- category was fire.incident.wf, now fire.incident.wildfire
- region was US-US-MT-GLACIER, now US-MT-GLACIER

Both raw and normalized values stored in event.data.

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
This commit is contained in:
Matt Johnson 2026-05-19 03:04:27 +00:00
commit dfad7ef45d
4 changed files with 202 additions and 61 deletions

View file

@ -17,6 +17,35 @@ WFIGS_PERIMETERS_URL = (
# Fall-off sweep window: 14 days (matches WFIGS's longest fall-off: large fires)
FALLOFF_WINDOW_DAYS = 14
# Incident type code mappings (WFIGS uses 2-letter codes)
INCIDENT_TYPE_MAP = {
"WF": "wildfire",
"RX": "prescribed_fire",
"CX": "complex",
"FA": "false_alarm",
}
def normalize_state(state: str | None) -> str | None:
"""Strip 'US-' prefix from POOState (ISO 3166-2 -> 2-letter)."""
if not state:
return None
if state.startswith("US-") and len(state) == 5:
return state[3:]
if len(state) == 2:
return state
return state # unknown shape, pass through
def normalize_incident_type(code: str | None) -> str:
"""Map IncidentTypeCategory code to a readable name."""
if not code:
return "unknown"
upper = code.upper()
if upper in INCIDENT_TYPE_MAP:
return INCIDENT_TYPE_MAP[upper]
return code.lower()
def severity_from_acres(acres: float | None) -> int:
"""Map DailyAcres to severity level 0-4."""
@ -42,6 +71,7 @@ def build_regions(state: str | None, county: str | None) -> tuple[list[str], str
"""
Build geo.regions list and primary_region from POOState and POOCounty.
Expects normalized 2-letter state codes (e.g., "MT" not "US-MT").
Returns (regions, primary_region).
"""
if not state:
@ -62,6 +92,7 @@ def subject_suffix(state: str | None, county: str | None) -> str:
"""
Build subject suffix from state and county.
Expects normalized 2-letter state codes.
Returns lowercase state.county (county with spacesunderscores).
Falls back to "unknown" if state is not available.
"""

View file

@ -25,6 +25,8 @@ from central.adapters.wfigs_common import (
extract_centroid,
get_observed_guids,
init_observed_table,
normalize_incident_type,
normalize_state,
parse_wfigs_timestamp,
point_in_bbox,
severity_from_acres,
@ -185,6 +187,7 @@ class WFIGSIncidentsAdapter(SourceAdapter):
return f"central.fire.incident.removed.{state}"
# Regular incidents: central.fire.incident.<state>.<county>
# POOState is already normalized (2-letter code)
state = event.data.get("POOState")
county = event.data.get("POOCounty")
suffix = subject_suffix(state, county)
@ -273,17 +276,22 @@ class WFIGSIncidentsAdapter(SourceAdapter):
):
continue
# Track this GUID as observed (for fall-off detection)
state = props.get("POOState")
# Normalize at parse boundary
state_raw = props.get("POOState")
state = normalize_state(state_raw)
county = props.get("POOCounty")
incident_type_raw = props.get("IncidentTypeCategory")
incident_type = normalize_incident_type(incident_type_raw)
# Track this GUID as observed (for fall-off detection)
# Store normalized state for consistency
current_guids[irwin_id] = (state, county)
# Parse fields
incident_type = props.get("IncidentTypeCategory", "unknown").lower()
discovery_time = parse_wfigs_timestamp(props.get("FireDiscoveryDateTime"))
daily_acres = props.get("DailyAcres")
# Build regions
# Build regions (expects normalized 2-letter state code)
regions, primary_region = build_regions(state, county)
# Build geo
@ -297,7 +305,7 @@ class WFIGSIncidentsAdapter(SourceAdapter):
else:
geo = Geo(regions=regions, primary_region=primary_region)
# Build event
# Build event with normalized values in data
event = Event(
id=irwin_id,
adapter=self.name,
@ -308,12 +316,14 @@ class WFIGSIncidentsAdapter(SourceAdapter):
data={
"IrwinID": irwin_id,
"IncidentName": props.get("IncidentName"),
"IncidentTypeCategory": props.get("IncidentTypeCategory"),
"IncidentTypeCategory": incident_type,
"IncidentTypeCategory_raw": incident_type_raw,
"DailyAcres": daily_acres,
"PercentContained": props.get("PercentContained"),
"FireDiscoveryDateTime": props.get("FireDiscoveryDateTime"),
"ModifiedOnDateTime": props.get("ModifiedOnDateTime"),
"POOState": state,
"POOState_raw": state_raw,
"POOCounty": county,
"raw": props,
},

View file

@ -25,6 +25,8 @@ from central.adapters.wfigs_common import (
extract_centroid,
get_observed_guids,
init_observed_table,
normalize_incident_type,
normalize_state,
parse_wfigs_timestamp,
polygon_intersects_bbox,
severity_from_acres,
@ -185,6 +187,7 @@ class WFIGSPerimetersAdapter(SourceAdapter):
return f"central.fire.perimeter.removed.{state}"
# Regular perimeters: central.fire.perimeter.<state>.<county>
# POOState is already normalized (2-letter code)
state = event.data.get("POOState")
county = event.data.get("POOCounty")
suffix = subject_suffix(state, county)
@ -271,18 +274,23 @@ class WFIGSPerimetersAdapter(SourceAdapter):
):
continue
# Track this GUID as observed (for fall-off detection)
state = props.get("attr_POOState")
# Normalize at parse boundary
state_raw = props.get("attr_POOState")
state = normalize_state(state_raw)
county = props.get("attr_POOCounty")
incident_type_raw = props.get("attr_IncidentTypeCategory")
incident_type = normalize_incident_type(incident_type_raw)
# Track this GUID as observed (for fall-off detection)
# Store normalized state for consistency
current_guids[irwin_id] = (state, county)
# Parse fields using prefixed names
incident_type = props.get("attr_IncidentTypeCategory", "unknown").lower()
discovery_time = parse_wfigs_timestamp(props.get("attr_FireDiscoveryDateTime"))
# Use poly_GISAcres or attr_IncidentSize for acreage
daily_acres = props.get("attr_IncidentSize") or props.get("poly_GISAcres")
# Build regions
# Build regions (expects normalized 2-letter state code)
regions, primary_region = build_regions(state, county)
# Extract centroid for geo
@ -320,13 +328,15 @@ class WFIGSPerimetersAdapter(SourceAdapter):
data={
"IrwinID": irwin_id,
"IncidentName": props.get("attr_IncidentName") or props.get("poly_IncidentName"),
"IncidentTypeCategory": props.get("attr_IncidentTypeCategory"),
"IncidentTypeCategory": incident_type,
"IncidentTypeCategory_raw": incident_type_raw,
"DailyAcres": props.get("attr_IncidentSize"),
"GISAcres": props.get("poly_GISAcres"),
"PercentContained": props.get("attr_PercentContained"),
"FireDiscoveryDateTime": props.get("attr_FireDiscoveryDateTime"),
"ModifiedOnDateTime": props.get("attr_ModifiedOnDateTime_dt"),
"POOState": state,
"POOState_raw": state_raw,
"POOCounty": county,
"geometry": geometry, # Full GeoJSON polygon
"raw": props,