hestia-earth-aggregation 0.21.17__tar.gz → 0.21.18__tar.gz

{hestia_earth_aggregation-0.21.17/hestia_earth_aggregation.egg-info → hestia_earth_aggregation-0.21.18}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: hestia_earth_aggregation
-Version: 0.21.17
+Version: 0.21.18
 Summary: HESTIA's aggregation engine.
 Home-page: https://gitlab.com/hestia-earth/hestia-aggregation-engine
 Author: HESTIA Team

{hestia_earth_aggregation-0.21.17 → hestia_earth_aggregation-0.21.18}/hestia_earth/aggregation/utils/management.py

@@ -1,6 +1,7 @@
 from hestia_earth.schema import ManagementJSONLD
 from hestia_earth.utils.model import linked_node
-from hestia_earth.utils.tools import safe_parse_date
+from hestia_earth.utils.tools import safe_parse_date, flatten
+from hestia_earth.utils.date import gapfill_datestr
 
 from . import match_dates, _set_dict_single
 from .queries import TIME_PERIOD, _get_time_ranges, _current_date
@@ -45,6 +46,34 @@ def aggregated_dates(blank_node: dict):
     )
 
 
+def split_by_dates(blank_nodes: list, start_year: int):
+    # if the blank_node overlaps with the start period of the aggregation, we need to split it in 2 parts
+    def split_data(blank_node: dict):
+        start_date = safe_parse_date(blank_node.get("startDate"), default=None)
+        end_date = safe_parse_date(blank_node.get("endDate"), default=None)
+        is_overlapping = all([start_date is not None, end_date is not None]) and all(
+            [start_date.year < start_year, start_year < end_date.year]
+        )
+        return (
+            [
+                blank_node
+                | {
+                    "startDate": blank_node.get("startDate"),
+                    "endDate": gapfill_datestr(start_year - 1, "end"),
+                },
+                blank_node
+                | {
+                    "startDate": gapfill_datestr(start_year, "start"),
+                    "endDate": blank_node.get("endDate"),
+                },
+            ]
+            if is_overlapping
+            else blank_node
+        )
+
+    return flatten(map(split_data, blank_nodes))
+
+
 def filter_management(blank_nodes: list, start_year: int = None, end_year: int = None):
     def update_dates(blank_node: dict):
         start_date = safe_parse_date(blank_node.get("startDate"), default=None)
@@ -72,7 +101,9 @@ def filter_management(blank_nodes: list, start_year: int = None, end_year: int =
     """
     return [
         update_dates(v)
-        for v in blank_nodes
+        for v in (
+            split_by_dates(blank_nodes, start_year) if start_year else blank_nodes
+        )
         if not start_year
         or not end_year
         or match_dates(v, start_year - TIME_PERIOD, end_year)

{hestia_earth_aggregation-0.21.17 → hestia_earth_aggregation-0.21.18}/hestia_earth/aggregation/utils/queries.py

@@ -15,7 +15,7 @@ from hestia_earth.aggregation.log import logger
 from . import HestiaError, _save_json, pick
 from .term import DEFAULT_COUNTRY_NAME, _fetch_countries, is_global as is_global_country
 
-SEARCH_LIMIT = 10000
+SEARCH_LIMIT = int(os.getenv("SEARCH_LIMIT", "10000"))
 _MATCH_AGGREGATED_QUERY = {"match": {"aggregated": "true"}}
 _MATCH_AGGREGATED_VALIDATED_QUERY = {"match": {"aggregatedDataValidated": "true"}}
 # enable matching any product

hestia_earth_aggregation-0.21.18/hestia_earth/aggregation/version.py

@@ -0,0 +1 @@
+VERSION = "0.21.18"

{hestia_earth_aggregation-0.21.17 → hestia_earth_aggregation-0.21.18/hestia_earth_aggregation.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: hestia_earth_aggregation
-Version: 0.21.17
+Version: 0.21.18
 Summary: HESTIA's aggregation engine.
 Home-page: https://gitlab.com/hestia-earth/hestia-aggregation-engine
 Author: HESTIA Team

hestia_earth_aggregation-0.21.17/hestia_earth/aggregation/version.py

@@ -1 +0,0 @@
-VERSION = "0.21.17"