fix: combine space-separated terms with adjacency operator in free-text search

src/pgstac/migrations/pgstac.0.9.8-unreleased.sql

@@ -227,6 +227,85 @@ END;
 $function$
 ;
 
+CREATE OR REPLACE FUNCTION pgstac.q_to_tsquery(jinput jsonb)
+ RETURNS tsquery
+ LANGUAGE plpgsql
+AS $function$
+DECLARE
+    input text;
+    processed_text text;
+    temp_text text;
+    quote_array text[];
+    placeholder text := '@QUOTE@';
+BEGIN
+    IF jsonb_typeof(jinput) = 'string' THEN
+        input := jinput->>0;
+    ELSIF jsonb_typeof(jinput) = 'array' THEN
+        input := array_to_string(
+            array(select jsonb_array_elements_text(jinput)),
+            ' OR '
+        );
+    ELSE
+        RAISE EXCEPTION 'Input must be a string or an array of strings.';
+    END IF;
+    -- Extract all quoted phrases and store in array
+    quote_array := regexp_matches(input, '"[^"]*"', 'g');
+
+    -- Replace each quoted part with a unique placeholder if there are any quoted phrases
+    IF array_length(quote_array, 1) IS NOT NULL THEN
+        processed_text := input;
+        FOR i IN array_lower(quote_array, 1) .. array_upper(quote_array, 1) LOOP
+            processed_text := replace(processed_text, quote_array[i], placeholder || i || placeholder);
+        END LOOP;
+    ELSE
+        processed_text := input;
+    END IF;
+
+    -- Replace non-quoted text using regular expressions
+
+    -- , -> |
+    processed_text := regexp_replace(processed_text, ',(?=(?:[^"]*"[^"]*")*[^"]*$)', ' | ', 'g');
+
+    -- and -> &
+    processed_text := regexp_replace(processed_text, '\s+AND\s+', ' & ', 'gi');
+
+    -- or -> |
+    processed_text := regexp_replace(processed_text, '\s+OR\s+', ' | ', 'gi');
+
+    -- + ->
+    processed_text := regexp_replace(processed_text, '^\s*\+([a-zA-Z0-9_]+)', '\1', 'g'); -- +term at start
+    processed_text := regexp_replace(processed_text, '\s*\+([a-zA-Z0-9_]+)', ' & \1', 'g'); -- +term elsewhere
+
+    -- - ->  !
+    processed_text := regexp_replace(processed_text, '^\s*\-([a-zA-Z0-9_]+)', '! \1', 'g'); -- -term at start
+    processed_text := regexp_replace(processed_text, '\s*\-([a-zA-Z0-9_]+)', ' & ! \1', 'g'); -- -term elsewhere
+
+    -- terms separated with spaces are assumed to represent adjacent terms. loop through these
+    -- occurrences and replace them with the adjacency operator (<->)
+    LOOP
+        temp_text := regexp_replace(processed_text, '([a-zA-Z0-9_]+)\s+([a-zA-Z0-9_]+)(?!\s*[&|<>])', '\1 <-> \2', 'g');
+        IF temp_text = processed_text THEN
+            EXIT; -- No more replacements were made
+        END IF;
+        processed_text := temp_text;
+    END LOOP;
+
+
+    -- Replace placeholders back with quoted phrases if there were any
+    IF array_length(quote_array, 1) IS NOT NULL THEN
+        FOR i IN array_lower(quote_array, 1) .. array_upper(quote_array, 1) LOOP
+            processed_text := replace(processed_text, placeholder || i || placeholder, '''' || substring(quote_array[i] from 2 for length(quote_array[i]) - 2) || '''');
+        END LOOP;
+    END IF;
+
+    -- Print processed_text to the console for debugging purposes
+    RAISE NOTICE 'processed_text: %', processed_text;
+
+    RETURN to_tsquery('english', processed_text);
+END;
+$function$
+;
+
 
 -- END migra calculated SQL
 DO $$

src/pgstac/migrations/pgstac.unreleased.sql

@@ -3104,6 +3104,18 @@ BEGIN
     -- - ->  !
     processed_text := regexp_replace(processed_text, '^\s*\-([a-zA-Z0-9_]+)', '! \1', 'g'); -- -term at start
     processed_text := regexp_replace(processed_text, '\s*\-([a-zA-Z0-9_]+)', ' & ! \1', 'g'); -- -term elsewhere
+
+    -- terms separated with spaces are assumed to represent adjacent terms. loop through these
+    -- occurrences and replace them with the adjacency operator (<->)
+    LOOP
+        temp_text := regexp_replace(processed_text, '([a-zA-Z0-9_]+)\s+([a-zA-Z0-9_]+)(?!\s*[&|<>])', '\1 <-> \2', 'g');
+        IF temp_text = processed_text THEN
+            EXIT; -- No more replacements were made
+        END IF;
+        processed_text := temp_text;
+    END LOOP;
+
+
     -- Replace placeholders back with quoted phrases if there were any
     IF array_length(quote_array, 1) IS NOT NULL THEN
         FOR i IN array_lower(quote_array, 1) .. array_upper(quote_array, 1) LOOP

src/pgstac/pgstac.sql

@@ -3104,6 +3104,18 @@ BEGIN
     -- - ->  !
     processed_text := regexp_replace(processed_text, '^\s*\-([a-zA-Z0-9_]+)', '! \1', 'g'); -- -term at start
     processed_text := regexp_replace(processed_text, '\s*\-([a-zA-Z0-9_]+)', ' & ! \1', 'g'); -- -term elsewhere
+
+    -- terms separated with spaces are assumed to represent adjacent terms. loop through these
+    -- occurrences and replace them with the adjacency operator (<->)
+    LOOP
+        temp_text := regexp_replace(processed_text, '([a-zA-Z0-9_]+)\s+([a-zA-Z0-9_]+)(?!\s*[&|<>])', '\1 <-> \2', 'g');
+        IF temp_text = processed_text THEN
+            EXIT; -- No more replacements were made
+        END IF;
+        processed_text := temp_text;
+    END LOOP;
+
+
     -- Replace placeholders back with quoted phrases if there were any
     IF array_length(quote_array, 1) IS NOT NULL THEN
         FOR i IN array_lower(quote_array, 1) .. array_upper(quote_array, 1) LOOP

src/pgstac/sql/004_search.sql

@@ -179,6 +179,18 @@ BEGIN
     -- - ->  !
     processed_text := regexp_replace(processed_text, '^\s*\-([a-zA-Z0-9_]+)', '! \1', 'g'); -- -term at start
     processed_text := regexp_replace(processed_text, '\s*\-([a-zA-Z0-9_]+)', ' & ! \1', 'g'); -- -term elsewhere
+
+    -- terms separated with spaces are assumed to represent adjacent terms. loop through these
+    -- occurrences and replace them with the adjacency operator (<->)
+    LOOP
+        temp_text := regexp_replace(processed_text, '([a-zA-Z0-9_]+)\s+([a-zA-Z0-9_]+)(?!\s*[&|<>])', '\1 <-> \2', 'g');
+        IF temp_text = processed_text THEN
+            EXIT; -- No more replacements were made
+        END IF;
+        processed_text := temp_text;
+    END LOOP;
+
+
     -- Replace placeholders back with quoted phrases if there were any
     IF array_length(quote_array, 1) IS NOT NULL THEN
         FOR i IN array_lower(quote_array, 1) .. array_upper(quote_array, 1) LOOP

src/pgstac/tests/basic/free_text.sql

@@ -107,3 +107,9 @@ select collection_search('{"q": "\"bear or stranger\""}');
 select collection_search('{"q": "office"}');
 
 select collection_search('{"q": ["bear", "stranger"]}');
+
+select collection_search('{"q": "large lizard"}');
+
+select collection_search('{"q": "teenagers fight monsters"}');
+
+select collection_search('{"q": "scary  monsters"}');

src/pgstac/tests/basic/free_text.sql.out

@@ -129,3 +129,12 @@ select collection_search('{"q": "office"}');
 
 select collection_search('{"q": ["bear", "stranger"]}');
  {"links": [], "collections": [{"id": "testcollection_1", "type": "Collection", "title": "Stranger Things", "extent": {"spatial": [[-180, -90, 180, 90]], "temporal": [["2016-01-01T00:00:00+00:00", "2025-12-31T23:59:59+00:00"]]}, "keywords": null, "description": "Some teenagers drop out of school to fight scary monsters", "stac_extensions": []}, {"id": "testcollection_2", "type": "Collection", "title": "The Bear", "extent": {"spatial": [[-180, -90, 180, 90]], "temporal": [["2022-01-01T00:00:00+00:00", "2025-12-31T23:59:59+00:00"]]}, "keywords": ["restaurant", "funny", "sad", "great"], "description": "Another story about why you should not start a restaurant", "stac_extensions": []}], "numberMatched": 2, "numberReturned": 2}
+
+select collection_search('{"q": "large lizard"}');
+ {"links": [], "collections": [{"id": "testcollection_3", "type": "Collection", "title": "Godzilla", "extent": {"spatial": [[-180, -90, 180, 90]], "temporal": [["1954-01-01T00:00:00+00:00", null]]}, "keywords": ["scary", "lizard", "monster"], "description": "A large lizard takes its revenge", "stac_extensions": []}], "numberMatched": 1, "numberReturned": 1}
+
+select collection_search('{"q": "teenagers fight monsters"}');
+ {"links": [], "collections": [], "numberMatched": 0, "numberReturned": 0}
+
+select collection_search('{"q": "scary  monsters"}');
+ {"links": [], "collections": [{"id": "testcollection_1", "type": "Collection", "title": "Stranger Things", "extent": {"spatial": [[-180, -90, 180, 90]], "temporal": [["2016-01-01T00:00:00+00:00", "2025-12-31T23:59:59+00:00"]]}, "keywords": null, "description": "Some teenagers drop out of school to fight scary monsters", "stac_extensions": []}], "numberMatched": 1, "numberReturned": 1}

src/pypgstac/src/pypgstac/version.py

@@ -1,3 +1,2 @@
 """Version."""
-
 __version__ = "0.9.8-dev"