IFRCGo · Rup-Narayan-Rajbanshi · Jun 5, 2025 · Jun 4, 2025 · Jun 5, 2025 · Jun 5, 2025
diff --git a/pystac_monty/sources/common.py b/pystac_monty/sources/common.py
@@ -1,12 +1,13 @@
 import abc
 import json
+import tempfile
 import typing
-from dataclasses import dataclass
+from dataclasses import dataclass, field
 from enum import Enum
-from typing import List, Literal, Tuple, Union
+from typing import List, Literal, Optional, Tuple, Union
 
 import requests
-from pydantic import BaseModel, Field
+from pydantic import BaseModel, ConfigDict, Field
 from pystac import Collection, Item
 
 from pystac_monty.geocoding import MontyGeoCoder
@@ -42,14 +43,18 @@ def success_rows(self) -> int:
         return self.total_rows - self.failed_rows
 
 
+class BaseModelWithExtra(BaseModel):
+    model_config = ConfigDict(extra="ignore", arbitrary_types_allowed=True)
+
+
 class DataType(Enum):
     FILE = "file"
     MEMORY = "memory"
 
 
-class File(BaseModel):
+class File(BaseModelWithExtra):
     data_type: Literal[DataType.FILE]
-    path: str
+    path: str | tempfile._TemporaryFileWrapper
 
 
 class Memory(BaseModel):
@@ -84,6 +89,13 @@ class USGSDataSourceType(BaseModel):
     loss_data: Union[File, Memory, None] = None
 
 
+class DesInventarData(BaseModel):
+    tmp_zip_file: File
+    country_code: str
+    iso3: str
+    source_url: str | None = None
+
+
 @dataclass
 class MontyDataSource:
     source_url: str
@@ -123,14 +135,15 @@ def get_data(self) -> typing.Any:
 
 @dataclass
 class MontyDataSourceV3:
-    root: Union[GenericDataSource, GdacsDataSourceType, USGSDataSourceType]
+    root: Union[GenericDataSource, GdacsDataSourceType, USGSDataSourceType, DesInventarData]
+    source_url: Optional[str] = field(init=False)
 
     def __post_init__(self):
         self.source_url = self.root.source_url
         if isinstance(self.root, GenericDataSource):
             self.input_data = self.root.input_data
 
-    def get_source_url(self) -> str:
+    def get_source_url(self) -> Optional[str]:
         """Get the Source URL"""
         return self.source_url
 

diff --git a/pystac_monty/sources/desinventar.py b/pystac_monty/sources/desinventar.py
@@ -20,7 +20,7 @@
     MontyImpactType,
 )
 from pystac_monty.hazard_profiles import MontyHazardProfiles
-from pystac_monty.sources.common import MontyDataTransformer
+from pystac_monty.sources.common import DesInventarData, MontyDataSourceV3, MontyDataTransformer
 from pystac_monty.validators.desinventar import (
     STAC_EVENT_ID_PREFIX,
     STAC_HAZARD_ID_PREFIX,
@@ -164,17 +164,16 @@ def parse_row_data(
 
 
 # FIXME: cleanup named temporary file
-class DesinventarDataSource:
+class DesinventarDataSource(MontyDataSourceV3):
     tmp_zip_file: tempfile._TemporaryFileWrapper
-    source_url: str | None
     country_code: str
     iso3: str
 
-    def __init__(self, tmp_zip_file: tempfile._TemporaryFileWrapper, country_code: str, iso3: str, source_url: str | None = None):
-        self.tmp_zip_file = tmp_zip_file
-        self.country_code = country_code
-        self.iso3 = iso3
-        self.source_url = source_url
+    def __init__(self, data: DesInventarData):
+        super().__init__(root=data)
+        self.tmp_zip_file = data.tmp_zip_file.path
+        self.country_code = data.country_code
+        self.iso3 = data.iso3
 
     @classmethod
     def from_zip_file(cls, zip_file: ZipFile, country_code: str, iso3: str):

diff --git a/pystac_monty/sources/emdat.py b/pystac_monty/sources/emdat.py
@@ -37,7 +37,6 @@ class EMDATDataSource(MontyDataSourceV3):
 
     df: pd.DataFrame
     file_path: str
-    source_url: str
     data_source: Union[File, Memory]
 
     def __init__(self, data: GenericDataSource):

diff --git a/pystac_monty/sources/usgs.py b/pystac_monty/sources/usgs.py
@@ -4,6 +4,7 @@
 import logging
 import os
 import typing
+from dataclasses import dataclass
 from datetime import datetime
 from enum import Enum
 
@@ -47,6 +48,7 @@ class USGSLossData(BaseModel):
     data: GenericDataSource
 
 
+@dataclass
 class USGSDataSource(MontyDataSourceV3):
     type: USGSSourceType
     source_url: str

diff --git a/tests/extensions/cassettes/test_desinventar/DesinventarTest.test_transformer_0_grd.yaml b/tests/extensions/cassettes/test_desinventar/DesinventarTest.test_transformer_0_grd.yaml
@@ -137,7 +137,7 @@ interactions:
       Access-Control-Allow-Origin:
       - '*'
       Age:
-      - '0'
+      - '398'
       Cache-Control:
       - max-age=600
       Connection:
@@ -147,11 +147,11 @@ interactions:
       Content-Type:
       - application/json; charset=utf-8
       Date:
-      - Fri, 30 May 2025 06:34:39 GMT
+      - Thu, 05 Jun 2025 06:30:26 GMT
       ETag:
-      - '"68362516-231e"'
+      - '"68362515-231e"'
       Last-Modified:
-      - Tue, 27 May 2025 20:48:22 GMT
+      - Tue, 27 May 2025 20:48:21 GMT
       Server:
       - GitHub.com
       Vary:
@@ -161,17 +161,17 @@ interactions:
       X-Cache:
       - HIT
       X-Cache-Hits:
-      - '0'
+      - '3'
       X-Fastly-Request-ID:
-      - 9f9f09ad2221afd85d14852d6401eb4f2c12feb6
+      - e4d6e8043d19816a287408e60606f714666e4b51
       X-GitHub-Request-Id:
-      - 5F40:10F1EC:134EF0:172ED1:68394806
+      - 4E22:64F80:87F27:A7133:684137F4
       X-Served-By:
-      - cache-maa10225-MAA
+      - cache-maa10226-MAA
       X-Timer:
-      - S1748586879.013310,VS0,VE233
+      - S1749105027.752903,VS0,VE1
       expires:
-      - Fri, 30 May 2025 06:04:15 GMT
+      - Thu, 05 Jun 2025 06:33:49 GMT
       x-proxy-cache:
       - MISS
     status:

diff --git a/tests/extensions/cassettes/test_emdat/EMDATTest.test_transformer_0.yaml b/tests/extensions/cassettes/test_emdat/EMDATTest.test_transformer_0.yaml
@@ -7,7 +7,7 @@ interactions:
       Host:
       - ifrcgo.org
       User-Agent:
-      - Python-urllib/3.13
+      - Python-urllib/3.12
     method: GET
     uri: https://ifrcgo.org/monty-stac-extension/v1.0.0/schema.json
   response:
@@ -137,7 +137,7 @@ interactions:
       Access-Control-Allow-Origin:
       - '*'
       Age:
-      - '0'
+      - '402'
       Cache-Control:
       - max-age=600
       Connection:
@@ -147,11 +147,11 @@ interactions:
       Content-Type:
       - application/json; charset=utf-8
       Date:
-      - Fri, 30 May 2025 08:09:35 GMT
+      - Thu, 05 Jun 2025 06:30:31 GMT
       ETag:
-      - '"68362516-231e"'
+      - '"68362515-231e"'
       Last-Modified:
-      - Tue, 27 May 2025 20:48:22 GMT
+      - Tue, 27 May 2025 20:48:21 GMT
       Server:
       - GitHub.com
       Vary:
@@ -163,15 +163,15 @@ interactions:
       X-Cache-Hits:
       - '0'
       X-Fastly-Request-ID:
-      - 080038baf25e880038eff8ebe24c59d0029b2905
+      - 0979feb4560e90b0ff7cda2945f2dae7293b3f36
       X-GitHub-Request-Id:
-      - 5F40:10F1EC:134EF0:172ED1:68394806
+      - 4E22:64F80:87F27:A7133:684137F4
       X-Served-By:
-      - cache-maa10227-MAA
+      - cache-maa10222-MAA
       X-Timer:
-      - S1748592576.517644,VS0,VE229
+      - S1749105032.564379,VS0,VE1
       expires:
-      - Fri, 30 May 2025 06:04:15 GMT
+      - Thu, 05 Jun 2025 06:33:49 GMT
       x-proxy-cache:
       - MISS
     status:

diff --git a/tests/extensions/cassettes/test_emdat/EMDATTest.test_transformer_with_file_data_0.yaml b/tests/extensions/cassettes/test_emdat/EMDATTest.test_transformer_with_file_data_0.yaml
@@ -7,7 +7,7 @@ interactions:
       Host:
       - ifrcgo.org
       User-Agent:
-      - Python-urllib/3.13
+      - Python-urllib/3.12
     method: GET
     uri: https://ifrcgo.org/monty-stac-extension/v1.0.0/schema.json
   response:
@@ -137,7 +137,7 @@ interactions:
       Access-Control-Allow-Origin:
       - '*'
       Age:
-      - '0'
+      - '403'
       Cache-Control:
       - max-age=600
       Connection:
@@ -147,11 +147,11 @@ interactions:
       Content-Type:
       - application/json; charset=utf-8
       Date:
-      - Fri, 30 May 2025 08:09:35 GMT
+      - Thu, 05 Jun 2025 06:30:31 GMT
       ETag:
-      - '"68362516-231e"'
+      - '"68362515-231e"'
       Last-Modified:
-      - Tue, 27 May 2025 20:48:22 GMT
+      - Tue, 27 May 2025 20:48:21 GMT
       Server:
       - GitHub.com
       Vary:
@@ -163,15 +163,15 @@ interactions:
       X-Cache-Hits:
       - '1'
       X-Fastly-Request-ID:
-      - d9b2de839b64e00a1b827aaf9f5f3be4c965eeca
+      - 60b661c66a2e62c55c4f3fd8c1a4186f37601210
       X-GitHub-Request-Id:
-      - 5F40:10F1EC:134EF0:172ED1:68394806
+      - 4E22:64F80:87F27:A7133:684137F4
       X-Served-By:
-      - cache-maa10236-MAA
+      - cache-maa10246-MAA
       X-Timer:
-      - S1748592576.976088,VS0,VE2
+      - S1749105032.850369,VS0,VE2
       expires:
-      - Fri, 30 May 2025 06:04:15 GMT
+      - Thu, 05 Jun 2025 06:33:49 GMT
       x-proxy-cache:
       - MISS
     status:

diff --git a/tests/extensions/cassettes/test_emdat/EMDATTest.test_transformer_with_json_data_0.yaml b/tests/extensions/cassettes/test_emdat/EMDATTest.test_transformer_with_json_data_0.yaml
@@ -7,7 +7,7 @@ interactions:
       Host:
       - ifrcgo.org
       User-Agent:
-      - Python-urllib/3.13
+      - Python-urllib/3.12
     method: GET
     uri: https://ifrcgo.org/monty-stac-extension/v1.0.0/schema.json
   response:
@@ -137,7 +137,7 @@ interactions:
       Access-Control-Allow-Origin:
       - '*'
       Age:
-      - '0'
+      - '403'
       Cache-Control:
       - max-age=600
       Connection:
@@ -147,11 +147,11 @@ interactions:
       Content-Type:
       - application/json; charset=utf-8
       Date:
-      - Fri, 30 May 2025 08:09:36 GMT
+      - Thu, 05 Jun 2025 06:30:32 GMT
       ETag:
-      - '"68362516-231e"'
+      - '"68362515-231e"'
       Last-Modified:
-      - Tue, 27 May 2025 20:48:22 GMT
+      - Tue, 27 May 2025 20:48:21 GMT
       Server:
       - GitHub.com
       Vary:
@@ -161,17 +161,17 @@ interactions:
       X-Cache:
       - HIT
       X-Cache-Hits:
-      - '1'
+      - '2'
       X-Fastly-Request-ID:
-      - f324a7226f606c69b43ad983df6bbb6235085e5e
+      - 468580308f9a6ef5698d0ee1e07b944b6e1134d2
       X-GitHub-Request-Id:
-      - 5F40:10F1EC:134EF0:172ED1:68394806
+      - 4E22:64F80:87F27:A7133:684137F4
       X-Served-By:
-      - cache-maa10229-MAA
+      - cache-maa10234-MAA
       X-Timer:
-      - S1748592576.182067,VS0,VE1
+      - S1749105032.112576,VS0,VE1
       expires:
-      - Fri, 30 May 2025 06:04:15 GMT
+      - Thu, 05 Jun 2025 06:33:49 GMT
       x-proxy-cache:
       - MISS
     status:

diff --git a/tests/extensions/cassettes/test_gdacs/GDACSTest.test_transformer_with_file_data_0.yaml b/tests/extensions/cassettes/test_gdacs/GDACSTest.test_transformer_with_file_data_0.yaml
@@ -137,7 +137,7 @@ interactions:
       Access-Control-Allow-Origin:
       - '*'
       Age:
-      - '6'
+      - '403'
       Cache-Control:
       - max-age=600
       Connection:
@@ -147,11 +147,11 @@ interactions:
       Content-Type:
       - application/json; charset=utf-8
       Date:
-      - Fri, 30 May 2025 06:34:45 GMT
+      - Thu, 05 Jun 2025 06:30:32 GMT
       ETag:
-      - '"68362516-231e"'
+      - '"68362515-231e"'
       Last-Modified:
-      - Tue, 27 May 2025 20:48:22 GMT
+      - Tue, 27 May 2025 20:48:21 GMT
       Server:
       - GitHub.com
       Vary:
@@ -161,17 +161,17 @@ interactions:
       X-Cache:
       - HIT
       X-Cache-Hits:
-      - '1'
+      - '4'
       X-Fastly-Request-ID:
-      - 7d28007237f6bc60f24a1a8b01656708bf608f24
+      - cd6f5bc3db7a0b1d45012b300aad504352f34fd1
       X-GitHub-Request-Id:
-      - 5F40:10F1EC:134EF0:172ED1:68394806
+      - 4E22:64F80:87F27:A7133:684137F4
       X-Served-By:
-      - cache-maa10237-MAA
+      - cache-maa10226-MAA
       X-Timer:
-      - S1748586886.542049,VS0,VE2
+      - S1749105032.388272,VS0,VE1
       expires:
-      - Fri, 30 May 2025 06:04:15 GMT
+      - Thu, 05 Jun 2025 06:33:49 GMT
       x-proxy-cache:
       - MISS
     status: