add missing migration

Author: DolceTriade

backend/migrations/0013_extract_context_match_spans.sql

@@ -0,0 +1,91 @@
+DROP FUNCTION IF EXISTS extract_context_with_highlight(TEXT, TEXT, INT, BOOLEAN);
+
+CREATE OR REPLACE FUNCTION extract_context_with_highlight(
+    p_text TEXT,
+    p_substring TEXT,
+    p_context_lines INT,
+    p_case_sensitive BOOLEAN DEFAULT FALSE
+)
+RETURNS TABLE (
+    match_line_number INT,
+    snippet_start_line_number INT,
+    context_snippet TEXT,
+    match_spans JSONB
+)
+LANGUAGE plpgsql
+AS $$
+DECLARE
+    lines TEXT[];
+    max_line_num INT;
+    current_line_num INT;
+    line_content TEXT;
+    end_line_num INT;
+    previous_context TEXT;
+    flags TEXT;
+    search_start INT;
+    match_start_char INT;
+    match_end_char INT;
+    line_offset_bytes INT;
+BEGIN
+    lines := regexp_split_to_array(COALESCE(p_text, ''), E'\n');
+    max_line_num := COALESCE(array_length(lines, 1), 0);
+    IF max_line_num = 0 THEN
+        RETURN;
+    END IF;
+
+    flags := CASE WHEN p_case_sensitive THEN '' ELSE 'i' END;
+
+    FOR current_line_num IN 1..max_line_num LOOP
+        line_content := lines[current_line_num];
+        IF (
+            (p_case_sensitive AND line_content ~ p_substring)
+            OR ((NOT p_case_sensitive) AND line_content ~* p_substring)
+        ) THEN
+            match_line_number := current_line_num;
+            snippet_start_line_number := GREATEST(1, current_line_num - p_context_lines);
+            end_line_num := LEAST(max_line_num, current_line_num + p_context_lines);
+            context_snippet := array_to_string(lines[snippet_start_line_number:end_line_num], E'\n');
+
+            previous_context := CASE
+                WHEN current_line_num > snippet_start_line_number
+                    THEN array_to_string(lines[snippet_start_line_number:current_line_num - 1], E'\n')
+                ELSE ''
+            END;
+            line_offset_bytes := octet_length(previous_context);
+            IF current_line_num > snippet_start_line_number THEN
+                line_offset_bytes := line_offset_bytes + 1;
+            END IF;
+
+            match_spans := '[]'::jsonb;
+            search_start := 1;
+
+            LOOP
+                match_start_char := regexp_instr(line_content, p_substring, search_start, 1, 0, flags);
+                EXIT WHEN match_start_char = 0;
+
+                match_end_char := regexp_instr(line_content, p_substring, search_start, 1, 1, flags);
+                IF match_end_char = 0 THEN
+                    match_end_char := char_length(line_content) + 1;
+                END IF;
+
+                match_spans := match_spans || jsonb_build_array(
+                    jsonb_build_object(
+                        'start', line_offset_bytes + octet_length(left(line_content, match_start_char - 1)),
+                        'end', line_offset_bytes + octet_length(left(line_content, match_end_char - 1))
+                    )
+                );
+
+                IF match_end_char <= match_start_char THEN
+                    search_start := match_start_char + 1;
+                ELSE
+                    search_start := match_end_char;
+                END IF;
+            END LOOP;
+
+            RETURN NEXT;
+        END IF;
+    END LOOP;
+
+    RETURN;
+END;
+$$;