Merge pull request #34 from cedadev/conditional_method

Conditional method

Author: rhysrevans3

docs/conf.py

@@ -13,7 +13,7 @@
 project = "Extraction Methods"
 copyright = "2025, Rhys Evans"
 author = "Rhys Evans"
-release = "1.1.2"
+release = "1.1.3"
 
 # -- General configuration ---------------------------------------------------
 # https://www.sphinx-doc.org/en/master/usage/configuration.html#general-configuration

extraction_methods/plugins/assets/backends/elasticsearch.py

@@ -86,41 +86,15 @@ def run(self, body: dict[str, Any]) -> Iterator[dict[str, Any]]:
 
         es = Elasticsearch_client(**self.input.client_kwargs)
 
-        es_body = {
-            "query": {
-                "bool": {
-                    "must": [
-                        {
-                            "regexp": {
-                                f"{self.input.search_field}.keyword": {
-                                    "value": self.input.regex,
-                                }
-                            }
-                        },
-                        {"exists": {"field": "md5"}},
-                    ],
-                    "must_not": [{"exists": {"field": "removed"}}],
-                },
-            },
-            "_source": [self.input.href_term, self.input.search_field]
-            + [extra_field.key for extra_field in self.input.extra_fields],
-        }
-
         # Run search
         result = es.search(
             index=self.input.index,
-            body=es_body,
+            body=self.input.body,
             timeout=f"{self.input.request_timeout}s",
         )
 
         for hit in result["hits"]["hits"]:
             source = hit["_source"]
-            asset = {
-                "href": source[self.input.href_term],
-            }
-
-            for field in self.input.extra_fields:
-                if value := source.get(field.key):
-                    asset[field.output_key] = value
+            source["href"] = source.pop(self.input.href_term)
 
-            yield asset
+            yield source

pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "extraction-methods"
-version = "1.1.2"
+version = "1.1.3"
 description = "Methods to enable the extraction of metadata"
 authors = [
     {"name" = "rhysrevans3", "email" = "<[email protected]>"},