C360 Fashion Retail demo setup with Docker Compose, Postgres schema, ystem instructions update and CSV data seeding

PromptQl Config file update

Author: srinigsan

.data/c360-fashion-retail/compose.yaml

@@ -20,6 +20,6 @@ services:
       timeout: 5s
       retries: 5
 
-      
+
 volumes:
   postgres_data:

.data/c360-fashion-retail/postgres/1-create-retail-pg.sql

@@ -1,5 +1,3 @@
-CREATE DATABASE retail_fashion;
-
 \c retail_fashion;
 
 -- Dumped from database version 17.6 (Debian 17.6-1.pgdg13+1)
@@ -8,7 +6,6 @@ CREATE DATABASE retail_fashion;
 SET statement_timeout = 0;
 SET lock_timeout = 0;
 SET idle_in_transaction_session_timeout = 0;
-SET transaction_timeout = 0;
 SET client_encoding = 'UTF8';
 SET standard_conforming_strings = on;
 SELECT pg_catalog.set_config('search_path', '', false);
@@ -513,26 +510,6 @@ CREATE TABLE public.customers (
 
 ALTER TABLE public.customers OWNER TO postgres;
 
---
--- Name: data_compatibility_report; Type: VIEW; Schema: public; Owner: postgres
---
-
-CREATE VIEW public.data_compatibility_report AS
- SELECT assessment_type,
-    table_name,
-    column_name,
-    total_rows,
-    non_null_rows,
-    convertible_rows,
-    conversion_success_rate,
-    sample_unconvertible,
-    max_length,
-    recommended_action
-   FROM public.assess_data_compatibility() assess_data_compatibility(assessment_type, table_name, column_name, total_rows, non_null_rows, convertible_rows, conversion_success_rate, sample_unconvertible, max_length, recommended_action);
-
-
-ALTER VIEW public.data_compatibility_report OWNER TO postgres;
-
 --
 -- Name: email_engagement; Type: TABLE; Schema: public; Owner: postgres
 --
@@ -1548,5 +1525,4 @@ ALTER TABLE ONLY public.style_similarity_matches
 -- PostgreSQL database dump complete
 --
 
-\unrestrict g7bUKWIv6sLlBRnt7yPasCEbL1SUmP43hOZ92X8UWUFG248Sh2it96UBTnHVpgk
 

.data/c360-fashion-retail/postgres/2-import-data.sql

@@ -43,79 +43,46 @@
 -- Load style similarity matches (depends on products)
 \COPY style_similarity_matches FROM '/docker-entrypoint-initdb.d/style_similarity_matches.csv' CSV HEADER;
 
--- Load additional tables if CSV files exist (these might be empty)
--- Use conditional loading to avoid errors if files don't exist
-
+-- Load optional/auxiliary CSVs if they exist (psql \COPY with ON_ERROR_STOP avoids failure)
+\COPY campaigns FROM '/docker-entrypoint-initdb.d/campaigns.csv' CSV HEADER;
+\COPY campaign_responses FROM '/docker-entrypoint-initdb.d/campaign_responses.csv' CSV HEADER;
+\COPY customer_service_interactions FROM '/docker-entrypoint-initdb.d/customer_service_interactions.csv' CSV HEADER;
+\COPY email_engagement FROM '/docker-entrypoint-initdb.d/email_engagement.csv' CSV HEADER;
+\COPY loyalty_activities FROM '/docker-entrypoint-initdb.d/loyalty_activities.csv' CSV HEADER;
+\COPY loyalty_profiles FROM '/docker-entrypoint-initdb.d/loyalty_profiles.csv' CSV HEADER;
+\COPY session_summary FROM '/docker-entrypoint-initdb.d/session_summary.csv' CSV HEADER;
+
+-- Ensure ON CONFLICT inserts work (add unique constraint on session_summary.date if not exists)
 DO $$
 BEGIN
-    -- Try to load campaigns
-    BEGIN
-        COPY campaigns FROM '/docker-entrypoint-initdb.d/campaigns.csv' CSV HEADER;
-    EXCEPTION WHEN OTHERS THEN
-        RAISE NOTICE 'campaigns.csv not found or empty, skipping...';
-    END;
-
-    -- Try to load campaign responses
-    BEGIN
-        COPY campaign_responses FROM '/docker-entrypoint-initdb.d/campaign_responses.csv' CSV HEADER;
-    EXCEPTION WHEN OTHERS THEN
-        RAISE NOTICE 'campaign_responses.csv not found or empty, skipping...';
-    END;
-
-    -- Try to load customer service interactions
-    BEGIN
-        COPY customer_service_interactions FROM '/docker-entrypoint-initdb.d/customer_service_interactions.csv' CSV HEADER;
-    EXCEPTION WHEN OTHERS THEN
-        RAISE NOTICE 'customer_service_interactions.csv not found or empty, skipping...';
-    END;
-
-    -- Try to load email engagement
-    BEGIN
-        COPY email_engagement FROM '/docker-entrypoint-initdb.d/email_engagement.csv' CSV HEADER;
-    EXCEPTION WHEN OTHERS THEN
-        RAISE NOTICE 'email_engagement.csv not found or empty, skipping...';
-    END;
-
-    -- Try to load loyalty activities
-    BEGIN
-        COPY loyalty_activities FROM '/docker-entrypoint-initdb.d/loyalty_activities.csv' CSV HEADER;
-    EXCEPTION WHEN OTHERS THEN
-        RAISE NOTICE 'loyalty_activities.csv not found or empty, skipping...';
-    END;
-
-    -- Try to load loyalty profiles
-    BEGIN
-        COPY loyalty_profiles FROM '/docker-entrypoint-initdb.d/loyalty_profiles.csv' CSV HEADER;
-    EXCEPTION WHEN OTHERS THEN
-        RAISE NOTICE 'loyalty_profiles.csv not found or empty, skipping...';
-    END;
-
-    -- Try to load session summary
-    BEGIN
-        COPY session_summary FROM '/docker-entrypoint-initdb.d/session_summary.csv' CSV HEADER;
-    EXCEPTION WHEN OTHERS THEN
-        RAISE NOTICE 'session_summary.csv not found or empty, skipping...';
-    END;
-
+    IF NOT EXISTS (
+        SELECT 1 
+        FROM pg_constraint 
+        WHERE conname = 'session_summary_date_key'
+    ) THEN
+        EXECUTE 'ALTER TABLE session_summary ADD CONSTRAINT session_summary_date_key UNIQUE (summary_date)';
+    END IF;
 END $$;
 
 -- Update sequences to match the highest IDs from imported data
--- Only update sequences for tables that have auto-incrementing IDs
 DO $$
 DECLARE
     seq_name TEXT;
-    table_name TEXT;
+    tbl_name TEXT;
+    col_name TEXT;
     max_id INTEGER;
 BEGIN
-    -- Update sequences for tables with serial columns
-    FOR seq_name, table_name IN
-        SELECT sequence_name, table_name
+    -- Find each sequence and its owning table/column
+    FOR seq_name, tbl_name, col_name IN
+        SELECT s.sequence_name, c.table_name, c.column_name
         FROM information_schema.sequences s
-        JOIN information_schema.columns c ON c.column_default LIKE '%' || s.sequence_name || '%'
+        JOIN information_schema.columns c 
+          ON c.column_default LIKE ('%' || s.sequence_name || '%')
         WHERE s.sequence_schema = 'public'
     LOOP
-        EXECUTE format('SELECT COALESCE(MAX(%I), 0) FROM %I',
-                      replace(seq_name, '_seq', ''), table_name) INTO max_id;
+        EXECUTE format('SELECT COALESCE(MAX(%I), 0) FROM %I', col_name, tbl_name)
+        INTO max_id;
+
         IF max_id > 0 THEN
             EXECUTE format('SELECT setval(%L, %s, true)', seq_name, max_id);
             RAISE NOTICE 'Updated sequence % to %', seq_name, max_id;
@@ -147,17 +114,17 @@ BEGIN
 
     FOR rec IN
         SELECT
-            t.tablename,
+            t.table_name,
             COALESCE(s.n_tup_ins, 0) as rows_loaded
         FROM information_schema.tables t
-        LEFT JOIN pg_stat_user_tables s ON s.tablename = t.table_name AND s.schemaname = 'public'
+        LEFT JOIN pg_stat_user_tables s 
+          ON s.relname = t.table_name AND s.schemaname = 'public'
         WHERE t.table_schema = 'public'
-        AND t.table_type = 'BASE TABLE'
-        -- Exclude staging and utility tables
-        AND t.table_name NOT LIKE 'staging_%'
+          AND t.table_type = 'BASE TABLE'
+          AND t.table_name NOT LIKE 'staging_%'
         ORDER BY t.table_name
     LOOP
-        RAISE NOTICE 'Table: % - Rows: %', rec.tablename, rec.rows_loaded;
+        RAISE NOTICE 'Table: % - Rows: %', rec.table_name, rec.rows_loaded;
         total_tables := total_tables + 1;
         total_rows := total_rows + rec.rows_loaded;
     END LOOP;
@@ -176,17 +143,13 @@ SELECT
     s.last_analyze,
     pg_size_pretty(pg_total_relation_size('public.'||t.table_name)) as table_size
 FROM information_schema.tables t
-LEFT JOIN pg_stat_user_tables s ON s.tablename = t.table_name AND s.schemaname = 'public'
+LEFT JOIN pg_stat_user_tables s ON s.relname = t.table_name AND s.schemaname = 'public'
 WHERE t.table_schema = 'public'
-AND t.table_type = 'BASE TABLE'
-AND t.table_name NOT LIKE 'staging_%'
+  AND t.table_type = 'BASE TABLE'
+  AND t.table_name NOT LIKE 'staging_%'
 ORDER BY s.n_tup_ins DESC NULLS LAST, t.table_name;
 
--- Grant permissions (if needed)
--- GRANT SELECT, INSERT, UPDATE, DELETE ON ALL TABLES IN SCHEMA public TO your_app_user;
--- GRANT USAGE, SELECT ON ALL SEQUENCES IN SCHEMA public TO your_app_user;
-
 ANALYZE; -- Update table statistics for query optimization
 
 \echo 'C360 Retail Fashion Database loading completed successfully!'
-\echo 'Use: SELECT * FROM table_row_counts; to see loaded data summary'
+\echo 'Use: SELECT * FROM table_row_counts; to see loaded data summary'

demos/C360-fashion-retail/globals/metadata/promptql-config.hml

@@ -57,44 +57,7 @@ definition:
       - Output format confirmed (side-by-side vs long)
     </pre_execution_validation>
 
-    <mandatory_checkpoint_enforcement>
-      **HARD STOP PROTOCOL - NO EXCEPTIONS**
-      
-      Before executing ANY query or analysis, you MUST complete the following checkpoint sequence:
-
-      **CHECKPOINT 1: QUESTION CLASSIFICATION**
-      - Type: "QUESTION CLASSIFICATION:"
-      - Identify entity-level
-      - If unclear, STOP and ask user for clarification.
-      - **BLOCKING RULE**: Cannot proceed without explicit classification.
-
-      **CHECKPOINT 2: COMPOSITE KEY VALIDATION**  
-      - Type: "COMPOSITE KEY REQUIRED:"
-      - Identify composite key fields relevant to classification 
-      - Type: "Composite key identified = YES"
-      - **BLOCKING RULE**: Cannot write SQL without identifying required composite key.
-
-      **CHECKPOINT 3: BASE CTE CONFIRMATION**
-      - Type: "BASE CTE STRUCTURE:"
-      - Write: "WITH [table_name]_base AS (SELECT DISTINCT [composite_key_fields] FROM...)"
-      - Type: "Base CTE planned = YES"
-      - **BLOCKING RULE**: Cannot aggregate without base CTE structure.
-
-      **CHECKPOINT 4: AGGREGATION VALIDATION**
-      - Type: "AGGREGATION APPROACH:"
-      - Confirm: "Aggregating FROM base CTE, NOT from raw table"
-      - Type: "Aggregation method validated = YES"
-      - **BLOCKING RULE**: Cannot execute query that aggregates directly from raw table.
-
-      **CHECKPOINT 5: EXECUTION CLEARANCE**
-      - Type: "EXECUTION CLEARANCE CHECKLIST:"
-      - "✓ Classification completed"
-      - "✓ Composite key identified" 
-      - "✓ Base CTE planned"
-      - "✓ Aggregation validated"
-      - Type: "CLEARED FOR EXECUTION = YES"
-      - **BLOCKING RULE**: Cannot execute without all four checkmarks.
-    </mandatory_checkpoint_enforcement>
+  
 
     <execution_protocol>
       Always follow this mandatory 7-step execution method: