Implement policies based on chunk creation time

- The idea with these changes is to support INTERVAL boundaries for
  INTEGER type columns in policy APIs. The boundary calculation will
  be done using chunk creation time metadata.
- Updated show_chunks, drop_chunks and retention policy APIs to get the
  affected chunks using chunk creation time metadata based on the
  INTERVAL boundary specified for the INTEGER columns.
- Honor "integer_now" function if it's specified so as to keep backwards
  compatibility with the existing behavior

Co-authored-by: Dipesh Pandit <[email protected]>

sql/policy_internal.sql

@@ -152,6 +152,7 @@ DECLARE
   dimtypeinput        REGPROC;
   compress_after      TEXT;
   lag_value           TEXT;
+  lag_bigint_value    BIGINT;
   htid                INTEGER;
   htoid               REGCLASS;
   chunk_rec           RECORD;
@@ -193,9 +194,19 @@ BEGIN
 
   lag_value := jsonb_object_field_text(config, 'compress_after');
 
+  -- We now allow INTERVAL lag values for INT-like dimensions. So
+  -- try to typecast to BIGINT (since it should be able to handle
+  -- INT/SMALLINT/BIGINT). If the typecasting fails then we know
+  -- it's an interval
+  BEGIN
+    lag_bigint_value := lag_value::BIGINT;
+  EXCEPTION WHEN OTHERS THEN
+    dimtype := 'INTERVAL' ::regtype;
+  END;
+
   -- execute the properly type casts for the lag value
   CASE dimtype
-    WHEN 'TIMESTAMP'::regtype, 'TIMESTAMPTZ'::regtype, 'DATE'::regtype THEN
+    WHEN 'TIMESTAMP'::regtype, 'TIMESTAMPTZ'::regtype, 'DATE'::regtype, 'INTERVAL' ::regtype THEN
       CALL _timescaledb_functions.policy_compression_execute(
         job_id, htid, lag_value::INTERVAL,
         maxchunks, verbose_log, recompress_enabled