Fix dbt Fusion BigQuery build

.env.example

@@ -62,12 +62,21 @@ GEMINI_API_KEY=your_gemini_api_key_here
 # =============================================================================
 # GOOGLE CLOUD (for GCS storage, Google Docs, and BigQuery)
 # =============================================================================
-# Service account credentials (JSON file path or JSON string)
-# Used for: GCS storage, Google Docs API, and BigQuery
+# Use Application Default Credentials (ADC) for Google auth.
+# Local setup:
+#   gcloud auth application-default login
+# Service account impersonation, if preferred:
+#   gcloud auth application-default login --impersonate-service-account service-account@project.iam.gserviceaccount.com
+# On GCE/Cloud Run/GitHub WIF, ADC uses the attached or federated service account.
+#
+# GOOGLE_APPLICATION_CREDENTIALS is intentionally unset for normal ADC auth.
+# Set it only if you must use a local service-account keyfile fallback.
+# GOOGLE_APPLICATION_CREDENTIALS=/path/to/service-account.json
+#
+# Used for: GCS storage, Google Docs API, and BigQuery.
 # Required scopes for Google Docs:
 #   - https://www.googleapis.com/auth/documents
 #   - https://www.googleapis.com/auth/drive
-GOOGLE_APPLICATION_CREDENTIALS=/path/to/service-account.json
 GCS_BUCKET_NAME=your_gcs_bucket_name
 
 # Google Docs Integration (optional)
@@ -89,9 +98,8 @@ GOOGLE_DRIVE_FOLDER_ID=your_google_drive_folder_id_here
 # Note: This can be the same as GOOGLE_DRIVE_FOLDER_ID if using one folder
 REALTOR_GDRIVE_FOLDER_ID=your_realtor_folder_id_here
 
-# BigQuery credentials (path to service account JSON on local dev)
-# On GCE the attached VM service account is used automatically via ADC.
-# GOOGLE_APPLICATION_CREDENTIALS=/path/to/service-account.json
+# BigQuery auth is through ADC. No service-account keyfile is required when
+# local ADC, attached service accounts, or workload identity federation are configured.
 
 # =============================================================================
 # DBT

AGENTS.md

@@ -66,7 +66,7 @@
 - Prefer small, focused commits with clear intent.
 - PRs should include a concise summary, linked issues (if applicable), and screenshots for UI changes.
 - Never merge PRs. Always leave merging to a human reviewer.
-- Add a PR comment tagging `@codex` to request review.
+- Add a PR comment tagging `@codex` to request review immediately after opening every PR. Use `gh pr comment <PR_NUMBER> --body "@codex please review this PR."` and mention the review request in the handoff.
 
 ## Data Request Scoping (Wizard Loop 1)
 

dbt_project/macros/calculate_commodity_analysis_return.sql

@@ -68,18 +68,18 @@ rolling_stats AS (
 
         MAX(current_price) OVER (
             PARTITION BY commodity_name, commodity_unit
-            ORDER BY date
-            RANGE BETWEEN INTERVAL 365 DAY PRECEDING AND CURRENT ROW
+            ORDER BY UNIX_DATE(date)
+            RANGE BETWEEN 365 PRECEDING AND CURRENT ROW
         ) AS high_1yr,
         MIN(current_price) OVER (
             PARTITION BY commodity_name, commodity_unit
-            ORDER BY date
-            RANGE BETWEEN INTERVAL 365 DAY PRECEDING AND CURRENT ROW
+            ORDER BY UNIX_DATE(date)
+            RANGE BETWEEN 365 PRECEDING AND CURRENT ROW
         ) AS low_1yr,
         STDDEV(daily_diff) OVER (
             PARTITION BY commodity_name, commodity_unit
-            ORDER BY date
-            RANGE BETWEEN INTERVAL 365 DAY PRECEDING AND CURRENT ROW
+            ORDER BY UNIX_DATE(date)
+            RANGE BETWEEN 365 PRECEDING AND CURRENT ROW
         ) AS std_diff_1yr,
         price_365d_ago AS price_start_1yr,
         CASE
@@ -90,18 +90,18 @@ rolling_stats AS (
 
         MAX(current_price) OVER (
             PARTITION BY commodity_name, commodity_unit
-            ORDER BY date
-            RANGE BETWEEN INTERVAL 270 DAY PRECEDING AND CURRENT ROW
+            ORDER BY UNIX_DATE(date)
+            RANGE BETWEEN 270 PRECEDING AND CURRENT ROW
         ) AS high_9mo,
         MIN(current_price) OVER (
             PARTITION BY commodity_name, commodity_unit
-            ORDER BY date
-            RANGE BETWEEN INTERVAL 270 DAY PRECEDING AND CURRENT ROW
+            ORDER BY UNIX_DATE(date)
+            RANGE BETWEEN 270 PRECEDING AND CURRENT ROW
         ) AS low_9mo,
         STDDEV(daily_diff) OVER (
             PARTITION BY commodity_name, commodity_unit
-            ORDER BY date
-            RANGE BETWEEN INTERVAL 270 DAY PRECEDING AND CURRENT ROW
+            ORDER BY UNIX_DATE(date)
+            RANGE BETWEEN 270 PRECEDING AND CURRENT ROW
         ) AS std_diff_9mo,
         price_270d_ago AS price_start_9mo,
         CASE
@@ -112,18 +112,18 @@ rolling_stats AS (
 
         MAX(current_price) OVER (
             PARTITION BY commodity_name, commodity_unit
-            ORDER BY date
-            RANGE BETWEEN INTERVAL 180 DAY PRECEDING AND CURRENT ROW
+            ORDER BY UNIX_DATE(date)
+            RANGE BETWEEN 180 PRECEDING AND CURRENT ROW
         ) AS high_6mo,
         MIN(current_price) OVER (
             PARTITION BY commodity_name, commodity_unit
-            ORDER BY date
-            RANGE BETWEEN INTERVAL 180 DAY PRECEDING AND CURRENT ROW
+            ORDER BY UNIX_DATE(date)
+            RANGE BETWEEN 180 PRECEDING AND CURRENT ROW
         ) AS low_6mo,
         STDDEV(daily_diff) OVER (
             PARTITION BY commodity_name, commodity_unit
-            ORDER BY date
-            RANGE BETWEEN INTERVAL 180 DAY PRECEDING AND CURRENT ROW
+            ORDER BY UNIX_DATE(date)
+            RANGE BETWEEN 180 PRECEDING AND CURRENT ROW
         ) AS std_diff_6mo,
         price_180d_ago AS price_start_6mo,
         CASE
@@ -134,18 +134,18 @@ rolling_stats AS (
 
         MAX(current_price) OVER (
             PARTITION BY commodity_name, commodity_unit
-            ORDER BY date
-            RANGE BETWEEN INTERVAL 90 DAY PRECEDING AND CURRENT ROW
+            ORDER BY UNIX_DATE(date)
+            RANGE BETWEEN 90 PRECEDING AND CURRENT ROW
         ) AS high_3mo,
         MIN(current_price) OVER (
             PARTITION BY commodity_name, commodity_unit
-            ORDER BY date
-            RANGE BETWEEN INTERVAL 90 DAY PRECEDING AND CURRENT ROW
+            ORDER BY UNIX_DATE(date)
+            RANGE BETWEEN 90 PRECEDING AND CURRENT ROW
         ) AS low_3mo,
         STDDEV(daily_diff) OVER (
             PARTITION BY commodity_name, commodity_unit
-            ORDER BY date
-            RANGE BETWEEN INTERVAL 90 DAY PRECEDING AND CURRENT ROW
+            ORDER BY UNIX_DATE(date)
+            RANGE BETWEEN 90 PRECEDING AND CURRENT ROW
         ) AS std_diff_3mo,
         price_90d_ago AS price_start_3mo,
         CASE
@@ -156,18 +156,18 @@ rolling_stats AS (
 
         MAX(current_price) OVER (
             PARTITION BY commodity_name, commodity_unit
-            ORDER BY date
-            RANGE BETWEEN INTERVAL 30 DAY PRECEDING AND CURRENT ROW
+            ORDER BY UNIX_DATE(date)
+            RANGE BETWEEN 30 PRECEDING AND CURRENT ROW
         ) AS high_1mo,
         MIN(current_price) OVER (
             PARTITION BY commodity_name, commodity_unit
-            ORDER BY date
-            RANGE BETWEEN INTERVAL 30 DAY PRECEDING AND CURRENT ROW
+            ORDER BY UNIX_DATE(date)
+            RANGE BETWEEN 30 PRECEDING AND CURRENT ROW
         ) AS low_1mo,
         STDDEV(daily_diff) OVER (
             PARTITION BY commodity_name, commodity_unit
-            ORDER BY date
-            RANGE BETWEEN INTERVAL 30 DAY PRECEDING AND CURRENT ROW
+            ORDER BY UNIX_DATE(date)
+            RANGE BETWEEN 30 PRECEDING AND CURRENT ROW
         ) AS std_diff_1mo,
         price_30d_ago AS price_start_1mo,
         CASE

dbt_project/macros/calculate_commodity_summary.sql

@@ -87,8 +87,13 @@ start_prices AS (
     FROM period_boundaries AS pb
     INNER JOIN filtered_data AS fd ON
         pb.commodity_name = fd.commodity_name
+        AND pb.commodity_unit = fd.commodity_unit
         AND pb.time_period = fd.time_period
         AND pb.period_start_date = fd.trade_date
+    QUALIFY ROW_NUMBER() OVER (
+        PARTITION BY pb.commodity_name, pb.commodity_unit, pb.time_period
+        ORDER BY fd.trade_date ASC, fd.price ASC
+    ) = 1
 ),
 
 end_prices AS (
@@ -100,8 +105,13 @@ end_prices AS (
     FROM period_boundaries AS pb
     INNER JOIN filtered_data AS fd ON
         pb.commodity_name = fd.commodity_name
+        AND pb.commodity_unit = fd.commodity_unit
         AND pb.time_period = fd.time_period
         AND pb.period_end_date = fd.trade_date
+    QUALIFY ROW_NUMBER() OVER (
+        PARTITION BY pb.commodity_name, pb.commodity_unit, pb.time_period
+        ORDER BY fd.trade_date DESC, fd.price DESC
+    ) = 1
 ),
 
 -- Main aggregation
@@ -140,10 +150,12 @@ combined_results AS (
     LEFT JOIN start_prices AS sp
         ON
             ar.commodity_name = sp.commodity_name
+            AND ar.commodity_unit = sp.commodity_unit
             AND ar.time_period = sp.time_period
     LEFT JOIN end_prices AS ep
         ON
             ar.commodity_name = ep.commodity_name
+            AND ar.commodity_unit = ep.commodity_unit
             AND ar.time_period = ep.time_period
 ),
 

dbt_project/macros/calculate_market_analysis_return.sql

@@ -88,18 +88,18 @@ rolling_stats AS (
 
         MAX(current_high) OVER (
             PARTITION BY symbol, exchange
-            ORDER BY date
-            RANGE BETWEEN INTERVAL 365 DAY PRECEDING AND CURRENT ROW
+            ORDER BY UNIX_DATE(date)
+            RANGE BETWEEN 365 PRECEDING AND CURRENT ROW
         ) AS high_1yr,
         MIN(current_low) OVER (
             PARTITION BY symbol, exchange
-            ORDER BY date
-            RANGE BETWEEN INTERVAL 365 DAY PRECEDING AND CURRENT ROW
+            ORDER BY UNIX_DATE(date)
+            RANGE BETWEEN 365 PRECEDING AND CURRENT ROW
         ) AS low_1yr,
         STDDEV(daily_diff) OVER (
             PARTITION BY symbol, exchange
-            ORDER BY date
-            RANGE BETWEEN INTERVAL 365 DAY PRECEDING AND CURRENT ROW
+            ORDER BY UNIX_DATE(date)
+            RANGE BETWEEN 365 PRECEDING AND CURRENT ROW
         ) AS std_diff_1yr,
         price_365d_ago AS price_start_1yr,
         CASE
@@ -110,18 +110,18 @@ rolling_stats AS (
 
         MAX(current_high) OVER (
             PARTITION BY symbol, exchange
-            ORDER BY date
-            RANGE BETWEEN INTERVAL 270 DAY PRECEDING AND CURRENT ROW
+            ORDER BY UNIX_DATE(date)
+            RANGE BETWEEN 270 PRECEDING AND CURRENT ROW
         ) AS high_9mo,
         MIN(current_low) OVER (
             PARTITION BY symbol, exchange
-            ORDER BY date
-            RANGE BETWEEN INTERVAL 270 DAY PRECEDING AND CURRENT ROW
+            ORDER BY UNIX_DATE(date)
+            RANGE BETWEEN 270 PRECEDING AND CURRENT ROW
         ) AS low_9mo,
         STDDEV(daily_diff) OVER (
             PARTITION BY symbol, exchange
-            ORDER BY date
-            RANGE BETWEEN INTERVAL 270 DAY PRECEDING AND CURRENT ROW
+            ORDER BY UNIX_DATE(date)
+            RANGE BETWEEN 270 PRECEDING AND CURRENT ROW
         ) AS std_diff_9mo,
         price_270d_ago AS price_start_9mo,
         CASE
@@ -132,18 +132,18 @@ rolling_stats AS (
 
         MAX(current_high) OVER (
             PARTITION BY symbol, exchange
-            ORDER BY date
-            RANGE BETWEEN INTERVAL 180 DAY PRECEDING AND CURRENT ROW
+            ORDER BY UNIX_DATE(date)
+            RANGE BETWEEN 180 PRECEDING AND CURRENT ROW
         ) AS high_6mo,
         MIN(current_low) OVER (
             PARTITION BY symbol, exchange
-            ORDER BY date
-            RANGE BETWEEN INTERVAL 180 DAY PRECEDING AND CURRENT ROW
+            ORDER BY UNIX_DATE(date)
+            RANGE BETWEEN 180 PRECEDING AND CURRENT ROW
         ) AS low_6mo,
         STDDEV(daily_diff) OVER (
             PARTITION BY symbol, exchange
-            ORDER BY date
-            RANGE BETWEEN INTERVAL 180 DAY PRECEDING AND CURRENT ROW
+            ORDER BY UNIX_DATE(date)
+            RANGE BETWEEN 180 PRECEDING AND CURRENT ROW
         ) AS std_diff_6mo,
         price_180d_ago AS price_start_6mo,
         CASE
@@ -154,18 +154,18 @@ rolling_stats AS (
 
         MAX(current_high) OVER (
             PARTITION BY symbol, exchange
-            ORDER BY date
-            RANGE BETWEEN INTERVAL 90 DAY PRECEDING AND CURRENT ROW
+            ORDER BY UNIX_DATE(date)
+            RANGE BETWEEN 90 PRECEDING AND CURRENT ROW
         ) AS high_3mo,
         MIN(current_low) OVER (
             PARTITION BY symbol, exchange
-            ORDER BY date
-            RANGE BETWEEN INTERVAL 90 DAY PRECEDING AND CURRENT ROW
+            ORDER BY UNIX_DATE(date)
+            RANGE BETWEEN 90 PRECEDING AND CURRENT ROW
         ) AS low_3mo,
         STDDEV(daily_diff) OVER (
             PARTITION BY symbol, exchange
-            ORDER BY date
-            RANGE BETWEEN INTERVAL 90 DAY PRECEDING AND CURRENT ROW
+            ORDER BY UNIX_DATE(date)
+            RANGE BETWEEN 90 PRECEDING AND CURRENT ROW
         ) AS std_diff_3mo,
         price_90d_ago AS price_start_3mo,
         CASE
@@ -176,18 +176,18 @@ rolling_stats AS (
 
         MAX(current_high) OVER (
             PARTITION BY symbol, exchange
-            ORDER BY date
-            RANGE BETWEEN INTERVAL 30 DAY PRECEDING AND CURRENT ROW
+            ORDER BY UNIX_DATE(date)
+            RANGE BETWEEN 30 PRECEDING AND CURRENT ROW
         ) AS high_1mo,
         MIN(current_low) OVER (
             PARTITION BY symbol, exchange
-            ORDER BY date
-            RANGE BETWEEN INTERVAL 30 DAY PRECEDING AND CURRENT ROW
+            ORDER BY UNIX_DATE(date)
+            RANGE BETWEEN 30 PRECEDING AND CURRENT ROW
         ) AS low_1mo,
         STDDEV(daily_diff) OVER (
             PARTITION BY symbol, exchange
-            ORDER BY date
-            RANGE BETWEEN INTERVAL 30 DAY PRECEDING AND CURRENT ROW
+            ORDER BY UNIX_DATE(date)
+            RANGE BETWEEN 30 PRECEDING AND CURRENT ROW
         ) AS std_diff_1mo,
         price_30d_ago AS price_start_1mo,
         CASE