From 05a035579657e2c56efd020c83c43053257de8d7 Mon Sep 17 00:00:00 2001
From: perf3ct <jonfuller2012@gmail.com>
Date: Tue, 8 Jul 2025 16:52:23 +0000
Subject: [PATCH] fix(tests): fix the crazy metrics collection issue

---
 ...708000001_fix_ocr_queue_stats_function.sql | 42 +++++++++++++++++++
 src/ocr/queue.rs                              | 32 +++++++++++++-
 2 files changed, 72 insertions(+), 2 deletions(-)
 create mode 100644 migrations/20250708000001_fix_ocr_queue_stats_function.sql

diff --git a/migrations/20250708000001_fix_ocr_queue_stats_function.sql b/migrations/20250708000001_fix_ocr_queue_stats_function.sql
new file mode 100644
index 0000000..0370004
--- /dev/null
+++ b/migrations/20250708000001_fix_ocr_queue_stats_function.sql
@@ -0,0 +1,42 @@
+-- Fix the get_ocr_queue_stats function to ensure it matches the expected structure
+-- This migration ensures the function correctly gets completed_today from documents table
+-- and handles the case where migration 20250620100019 may have failed silently
+
+CREATE OR REPLACE FUNCTION get_ocr_queue_stats()
+RETURNS TABLE (
+    pending_count BIGINT,
+    processing_count BIGINT,
+    failed_count BIGINT,
+    completed_today BIGINT,
+    avg_wait_time_minutes DOUBLE PRECISION,
+    oldest_pending_minutes DOUBLE PRECISION
+) AS $$
+BEGIN
+    RETURN QUERY
+    WITH queue_stats AS (
+        SELECT 
+            COUNT(*) FILTER (WHERE status = 'pending') as pending_count,
+            COUNT(*) FILTER (WHERE status = 'processing') as processing_count,
+            COUNT(*) FILTER (WHERE status = 'failed' AND attempts >= max_attempts) as failed_count,
+            CAST(AVG(EXTRACT(EPOCH FROM (COALESCE(started_at, NOW()) - created_at))/60) FILTER (WHERE status IN ('processing', 'completed')) AS DOUBLE PRECISION) as avg_wait_time_minutes,
+            CAST(MAX(EXTRACT(EPOCH FROM (NOW() - created_at))/60) FILTER (WHERE status = 'pending') AS DOUBLE PRECISION) as oldest_pending_minutes
+        FROM ocr_queue
+    ),
+    document_stats AS (
+        -- Count documents that completed OCR today (looking at documents table where actual completion is tracked)
+        SELECT COUNT(*) as completed_today
+        FROM documents
+        WHERE ocr_status = 'completed'
+        AND updated_at >= CURRENT_DATE
+        AND updated_at < CURRENT_DATE + INTERVAL '1 day'
+    )
+    SELECT 
+        queue_stats.pending_count,
+        queue_stats.processing_count,
+        queue_stats.failed_count,
+        document_stats.completed_today,
+        queue_stats.avg_wait_time_minutes,
+        queue_stats.oldest_pending_minutes
+    FROM queue_stats, document_stats;
+END;
+$$ LANGUAGE plpgsql;
\ No newline at end of file
diff --git a/src/ocr/queue.rs b/src/ocr/queue.rs
index 10e5e8a..1386fe7 100644
--- a/src/ocr/queue.rs
+++ b/src/ocr/queue.rs
@@ -795,6 +795,31 @@ impl OcrQueueService {
     pub async fn get_stats(&self) -> Result<QueueStats> {
         tracing::debug!("OCR Queue: Starting get_stats() call");
         
+        // First, let's check if the function exists and what it returns
+        let function_exists = sqlx::query_scalar::<_, bool>(
+            r#"
+            SELECT EXISTS (
+                SELECT 1 FROM pg_proc p
+                JOIN pg_namespace n ON p.pronamespace = n.oid
+                WHERE n.nspname = 'public' AND p.proname = 'get_ocr_queue_stats'
+            )
+            "#
+        )
+        .fetch_one(&self.pool)
+        .await
+        .map_err(|e| {
+            tracing::error!("OCR Queue: Failed to check if function exists: {}", e);
+            e
+        })?;
+        
+        if !function_exists {
+            tracing::error!("OCR Queue: Function get_ocr_queue_stats() does not exist");
+            return Err(anyhow::anyhow!("Function get_ocr_queue_stats() does not exist"));
+        }
+        
+        tracing::debug!("OCR Queue: Function get_ocr_queue_stats() exists, attempting to call it");
+        
+        // Call the function
         let stats = sqlx::query(
             r#"
             SELECT * FROM get_ocr_queue_stats()
@@ -809,10 +834,13 @@ impl OcrQueueService {
 
         tracing::debug!("OCR Queue: Successfully fetched stats row");
         
-        // Debug: Print all column names and their types
+        // Debug: Print all column names, their types, and their values
         let columns = stats.columns();
         for (i, column) in columns.iter().enumerate() {
-            tracing::debug!("OCR Queue: Column {}: name='{}', type='{:?}'", i, column.name(), column.type_info());
+            let column_name = column.name();
+            let column_type = column.type_info();
+            
+            tracing::debug!("OCR Queue: Column {}: name='{}', type='{:?}'", i, column_name, column_type);
         }
 
         // Try to extract each field with detailed error handling