Merge remote-tracking branch 'origin/main' into ibtracs

Author: emmanuelmathot

pystac_monty/geocoding.py

@@ -4,6 +4,7 @@
 from typing import Any, Dict, List, Optional, Set, Union
 
 import fiona  # type: ignore
+import requests
 from shapely.geometry import mapping, shape  # type: ignore
 from shapely.ops import unary_union  # type: ignore
 
@@ -191,7 +192,7 @@ class GAULGeocoder(MontyGeoCoder):
     Loads features dynamically as needed.
     """
 
-    def __init__(self, gpkg_path: str, simplify_tolerance: float = 0.01) -> None:
+    def __init__(self, gpkg_path: Optional[str], service_base_url: Optional[str], simplify_tolerance: float = 0.01) -> None:
         """
         Initialize GAULGeocoder
 
@@ -206,8 +207,16 @@ def __init__(self, gpkg_path: str, simplify_tolerance: float = 0.01) -> None:
         self._simplify_tolerance = simplify_tolerance
         self._cache: Dict[str, Union[Dict[str, Any], int, None]] = {}  # Cache for frequently accessed geometries
         self._file_handle = None
-        self._initialize_path()
-        self._open_file()
+        
+        if not gpkg_path and not service_base_url:
+            raise ValueError("At least the gpkg_path or service_base_url should be set.")
+
+        if self.gpkg_path:
+            self._initialize_path()
+            self._open_file()
+        else:
+            self.service_base_url = service_base_url
+            self.request_timeout = 30
 
     def __enter__(self) -> "GAULGeocoder":
         """Context manager entry point"""
@@ -380,6 +389,12 @@ def _get_name_to_adm0_mapping(self, name: str) -> Optional[int]:
                 return adm0_code
         return None
 
+    def _service_request_handler(self, service_url: str, params: dict):
+        response = requests.get(service_url, params=params, timeout=self.request_timeout)
+        if response.status_code == 200:
+            return response.json()
+        return None
+
     def get_geometry_from_admin_units(self, admin_units: str) -> Optional[Dict[str, Any]]:
         """
         Get geometry from admin units JSON string
@@ -453,6 +468,12 @@ def get_geometry_by_country_name(self, country_name: str) -> Optional[Dict[str,
         Returns:
             Dictionary containing geometry and bbox if found
         """
+
+        if not self.gpkg_path:
+            params = {"country_name": country_name}
+            service_url = f"{self.service_base_url}/by_country_name"
+            return self._service_request_handler(service_url=service_url, params=params)
+
         if not country_name or not self._path:
             return None
 

pystac_monty/sources/emdat.py

@@ -2,6 +2,7 @@
 from datetime import datetime
 from typing import List, Optional, Union
 
+import numpy as np
 import pandas as pd
 import pytz
 from pystac import Item, Link
@@ -18,6 +19,7 @@
 from pystac_monty.geocoding import MontyGeoCoder
 from pystac_monty.hazard_profiles import MontyHazardProfiles
 from pystac_monty.sources.common import MontyDataSource, MontyDataTransformer
+from pystac_monty.utils import rename_columns
 
 STAC_EVENT_ID_PREFIX = "emdat-event-"
 STAC_HAZARD_ID_PREFIX = "emdat-hazard-"
@@ -37,8 +39,13 @@ def __init__(self, source_url: str, data: Union[str, pd.DataFrame]):
             self.df = pd.read_excel(data)
         elif isinstance(data, pd.DataFrame):
             self.df = data
+        elif isinstance(data, dict):
+            # If data is a dict, assume it's Json content
+            data = data["data"]["public_emdat"]["data"]
+            df = pd.DataFrame(data)
+            self.df = rename_columns(df)
         else:
-            raise ValueError("Data must be either Excel content (str) or pandas DataFrame")
+            raise ValueError("Data must be either Excel content (str) or pandas DataFrame or Json")
 
     def get_data(self) -> pd.DataFrame:
         return self.df
@@ -109,7 +116,7 @@ def _create_event_item_from_row(self, row: pd.Series) -> Optional[Item]:
         bbox = None
 
         # 1. Try admin units first if geocoder is available
-        if self.geocoder and not pd.isna(row.get("Admin Units")):
+        if self.geocoder and np.any(pd.notna(row.get("Admin Units"))):
             geom_data = self.geocoder.get_geometry_from_admin_units(row.get("Admin Units"))
             if geom_data:
                 geometry = geom_data["geometry"]

pystac_monty/sources/pdc.py

@@ -56,7 +56,13 @@ def __init__(self, pdc_data_src: PDCDataSource):
                 self.exposure_detail = json.loads(f.read())
 
         self.uuid = self.config_data.get("uuid", None)
-        self.episode_number = int(float(self.config_data.get("exposure_timestamp", 0)))
+        # Note: We might need to handle this differently if the exposure_timestamp
+        # is other than numeric (e.g. alphabetic)
+        # For now, we have assigned 0 to the episode_number
+        try:
+            self.episode_number = int(float(self.config_data.get("exposure_timestamp", 0)))
+        except ValueError:
+            self.episode_number = 0
         if "geojson_file_path" in self.config_data and os.path.exists(self.config_data["geojson_file_path"]):
             with open(self.config_data["geojson_file_path"], "r", encoding="utf-8") as f:
                 self.geojson_data = json.loads(f.read())

pystac_monty/utils.py

@@ -0,0 +1,31 @@
+def rename_columns(df):
+    """Rename columns of a pandas DataFrame"""
+    return df.rename(
+        columns={
+            "disno": "DisNo.",
+            "admin_units": "Admin Units",
+            "latitude": "Latitude",
+            "longitude": "Longitude",
+            "country": "Country",
+            "classif_key": "Classification Key",
+            "iso": "ISO",
+            "total_deaths": "Total Deaths",
+            "no_injured": "No Injured",
+            "no_affected": "No Affected",
+            "no_homeless": "No Homeless",
+            "total_affected": "Total Affected",
+            "total_dam": "Total Damages ('000 US$)",
+            "start_year": "Start Year",
+            "start_month": "Start Month",
+            "start_day": "Start Day",
+            "end_year": "End Year",
+            "end_month": "End Month",
+            "end_day": "End Day",
+            "magnitude": "Magnitude",
+            "magnitude_scale": "Magnitude Sacle",
+            "name": "Event Name",
+            "type": "Disaster Type",
+            "subtype": "Disaster Subtype",
+            "location": "Location",
+        }
+    )