matt/inventory
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Fixed calculations for frontend (likely still wrong but they display) + related regressions to calculate script

Browse Source
This commit is contained in:
Matt
2025-02-05 00:02:06 -05:00
parent ebffb8f912
commit 2a6a0d0a87
1 changed files with 186 additions and 120 deletions
Download Patch File Download Diff File Expand all files Collapse all files

306
inventory-server/scripts/metrics/product-metrics.js
View File

@@ -13,6 +13,7 @@ async function calculateProductMetrics(startTime, totalProducts, processedCount
const connection = await getConnection();
let success = false;
let processedOrders = 0;
const BATCH_SIZE = 5000;
try {
// Skip flags are inherited from the parent scope
@@ -44,7 +45,7 @@ async function calculateProductMetrics(startTime, totalProducts, processedCount
return {
processedProducts: processedCount,
processedOrders,
processedPurchaseOrders: 0, // This module doesn't process POs
processedPurchaseOrders: 0,
success
};
}
@@ -56,6 +57,15 @@ async function calculateProductMetrics(startTime, totalProducts, processedCount
FROM products
`);
// Get threshold settings once
const [thresholds] = await connection.query(`
SELECT critical_days, reorder_days, overstock_days, low_stock_threshold
FROM stock_thresholds
WHERE category_id IS NULL AND vendor IS NULL
LIMIT 1
`);
const defaultThresholds = thresholds[0];
// Calculate base product metrics
if (!SKIP_PRODUCT_BASE_METRICS) {
outputProgress({
@@ -82,134 +92,190 @@ async function calculateProductMetrics(startTime, totalProducts, processedCount
`);
processedOrders = orderCount[0].count;
// Calculate base metrics
// Clear temporary tables
await connection.query('TRUNCATE TABLE temp_sales_metrics');
await connection.query('TRUNCATE TABLE temp_purchase_metrics');
// Populate temp_sales_metrics with base stats and sales averages
await connection.query(`
UPDATE product_metrics pm
JOIN (
SELECT
p.pid,
p.stock_quantity,
p.cost_price,
p.cost_price * p.stock_quantity as inventory_value,
SUM(o.quantity) as total_quantity,
COUNT(DISTINCT o.order_number) as number_of_orders,
SUM(o.quantity * o.price) as total_revenue,
SUM(o.quantity * p.cost_price) as cost_of_goods_sold,
AVG(o.price) as avg_price,
STDDEV(o.price) as price_std,
MIN(o.date) as first_sale_date,
MAX(o.date) as last_sale_date,
COUNT(DISTINCT DATE(o.date)) as active_days
FROM products p
LEFT JOIN orders o ON p.pid = o.pid AND o.canceled = false
GROUP BY p.pid, p.stock_quantity, p.cost_price
) stats ON pm.pid = stats.pid
SET
pm.inventory_value = COALESCE(stats.inventory_value, 0),
pm.avg_quantity_per_order = COALESCE(stats.total_quantity / NULLIF(stats.number_of_orders, 0), 0),
pm.number_of_orders = COALESCE(stats.number_of_orders, 0),
pm.total_revenue = COALESCE(stats.total_revenue, 0),
pm.cost_of_goods_sold = COALESCE(stats.cost_of_goods_sold, 0),
pm.gross_profit = COALESCE(stats.total_revenue - stats.cost_of_goods_sold, 0),
pm.avg_margin_percent = CASE
WHEN COALESCE(stats.total_revenue, 0) > 0
THEN ((stats.total_revenue - stats.cost_of_goods_sold) / stats.total_revenue) * 100
INSERT INTO temp_sales_metrics
SELECT
p.pid,
COALESCE(SUM(o.quantity) / NULLIF(COUNT(DISTINCT DATE(o.date)), 0), 0) as daily_sales_avg,
COALESCE(SUM(o.quantity) / NULLIF(CEIL(COUNT(DISTINCT DATE(o.date)) / 7), 0), 0) as weekly_sales_avg,
COALESCE(SUM(o.quantity) / NULLIF(CEIL(COUNT(DISTINCT DATE(o.date)) / 30), 0), 0) as monthly_sales_avg,
COALESCE(SUM(o.quantity * o.price), 0) as total_revenue,
CASE
WHEN SUM(o.quantity * o.price) > 0
THEN ((SUM(o.quantity * o.price) - SUM(o.quantity * p.cost_price)) / SUM(o.quantity * o.price)) * 100
ELSE 0
END,
pm.first_sale_date = stats.first_sale_date,
pm.last_sale_date = stats.last_sale_date,
pm.days_of_inventory = CASE
WHEN COALESCE(stats.total_quantity / NULLIF(stats.active_days, 0), 0) > 0
THEN FLOOR(stats.stock_quantity / (stats.total_quantity / stats.active_days))
ELSE NULL
END,
pm.weeks_of_inventory = CASE
WHEN COALESCE(stats.total_quantity / NULLIF(stats.active_days, 0), 0) > 0
THEN FLOOR(stats.stock_quantity / (stats.total_quantity / stats.active_days) / 7)
ELSE NULL
END,
pm.gmroi = CASE
WHEN COALESCE(stats.inventory_value, 0) > 0
THEN (stats.total_revenue - stats.cost_of_goods_sold) / stats.inventory_value
ELSE 0
END,
pm.last_calculated_at = NOW()
END as avg_margin_percent,
MIN(o.date) as first_sale_date,
MAX(o.date) as last_sale_date
FROM products p
LEFT JOIN orders o ON p.pid = o.pid
AND o.canceled = false
AND o.date >= DATE_SUB(CURDATE(), INTERVAL 90 DAY)
GROUP BY p.pid
`);
// Calculate forecast accuracy and bias
// Populate temp_purchase_metrics
await connection.query(`
WITH forecast_accuracy AS (
SELECT
sf.pid,
AVG(CASE
WHEN o.quantity > 0
THEN ABS(sf.forecast_units - o.quantity) / o.quantity * 100
ELSE 100
END) as avg_forecast_error,
AVG(CASE
WHEN o.quantity > 0
THEN (sf.forecast_units - o.quantity) / o.quantity * 100
INSERT INTO temp_purchase_metrics
SELECT
p.pid,
AVG(DATEDIFF(po.received_date, po.date)) as avg_lead_time_days,
MAX(po.date) as last_purchase_date,
MIN(po.received_date) as first_received_date,
MAX(po.received_date) as last_received_date
FROM products p
LEFT JOIN purchase_orders po ON p.pid = po.pid
AND po.received_date IS NOT NULL
AND po.date >= DATE_SUB(CURDATE(), INTERVAL 365 DAY)
GROUP BY p.pid
`);
// Process updates in batches
let lastPid = 0;
while (true) {
if (isCancelled) break;
const [batch] = await connection.query(
'SELECT pid FROM products WHERE pid > ? ORDER BY pid LIMIT ?',
[lastPid, BATCH_SIZE]
);
if (batch.length === 0) break;
await connection.query(`
UPDATE product_metrics pm
JOIN products p ON pm.pid = p.pid
LEFT JOIN temp_sales_metrics sm ON pm.pid = sm.pid
LEFT JOIN temp_purchase_metrics lm ON pm.pid = lm.pid
SET
pm.inventory_value = p.stock_quantity * p.cost_price,
pm.daily_sales_avg = COALESCE(sm.daily_sales_avg, 0),
pm.weekly_sales_avg = COALESCE(sm.weekly_sales_avg, 0),
pm.monthly_sales_avg = COALESCE(sm.monthly_sales_avg, 0),
pm.total_revenue = COALESCE(sm.total_revenue, 0),
pm.avg_margin_percent = COALESCE(sm.avg_margin_percent, 0),
pm.first_sale_date = sm.first_sale_date,
pm.last_sale_date = sm.last_sale_date,
pm.avg_lead_time_days = COALESCE(lm.avg_lead_time_days, 30),
pm.days_of_inventory = CASE
WHEN COALESCE(sm.daily_sales_avg, 0) > 0
THEN FLOOR(p.stock_quantity / sm.daily_sales_avg)
ELSE NULL
END,
pm.weeks_of_inventory = CASE
WHEN COALESCE(sm.weekly_sales_avg, 0) > 0
THEN FLOOR(p.stock_quantity / sm.weekly_sales_avg)
ELSE NULL
END,
pm.stock_status = CASE
WHEN p.stock_quantity <= 0 THEN 'Out of Stock'
WHEN COALESCE(sm.daily_sales_avg, 0) = 0 AND p.stock_quantity <= ? THEN 'Low Stock'
WHEN COALESCE(sm.daily_sales_avg, 0) = 0 THEN 'In Stock'
WHEN p.stock_quantity / NULLIF(sm.daily_sales_avg, 0) <= ? THEN 'Critical'
WHEN p.stock_quantity / NULLIF(sm.daily_sales_avg, 0) <= ? THEN 'Reorder'
WHEN p.stock_quantity / NULLIF(sm.daily_sales_avg, 0) > ? THEN 'Overstocked'
ELSE 'Healthy'
END,
pm.reorder_qty = CASE
WHEN COALESCE(sm.daily_sales_avg, 0) > 0 THEN
GREATEST(
CEIL(sm.daily_sales_avg * COALESCE(lm.avg_lead_time_days, 30) * 1.96),
?
)
ELSE ?
END,
pm.overstocked_amt = CASE
WHEN p.stock_quantity / NULLIF(sm.daily_sales_avg, 0) > ?
THEN GREATEST(0, p.stock_quantity - CEIL(sm.daily_sales_avg * ?))
ELSE 0
END) as avg_forecast_bias,
MAX(sf.forecast_date) as last_forecast_date
FROM sales_forecasts sf
JOIN orders o ON sf.pid = o.pid
AND DATE(o.date) = sf.forecast_date
WHERE o.canceled = false
AND sf.forecast_date >= DATE_SUB(CURRENT_DATE, INTERVAL 90 DAY)
GROUP BY sf.pid
)
UPDATE product_metrics pm
JOIN forecast_accuracy fa ON pm.pid = fa.pid
SET
pm.forecast_accuracy = GREATEST(0, 100 - LEAST(fa.avg_forecast_error, 100)),
pm.forecast_bias = GREATEST(-100, LEAST(fa.avg_forecast_bias, 100)),
pm.last_forecast_date = fa.last_forecast_date,
pm.last_calculated_at = NOW()
`);
END,
pm.last_calculated_at = NOW()
WHERE p.pid IN (?)
`, [
defaultThresholds.low_stock_threshold,
defaultThresholds.critical_days,
defaultThresholds.reorder_days,
defaultThresholds.overstock_days,
defaultThresholds.low_stock_threshold,
defaultThresholds.low_stock_threshold,
defaultThresholds.overstock_days,
defaultThresholds.overstock_days,
batch.map(row => row.pid)
]);
processedCount = Math.floor(totalProducts * 0.4);
outputProgress({
status: 'running',
operation: 'Base product metrics calculated',
current: processedCount || 0,
total: totalProducts || 0,
elapsed: formatElapsedTime(startTime),
remaining: estimateRemaining(startTime, processedCount || 0, totalProducts || 0),
rate: calculateRate(startTime, processedCount || 0),
percentage: (((processedCount || 0) / (totalProducts || 1)) * 100).toFixed(1),
timing: {
start_time: new Date(startTime).toISOString(),
end_time: new Date().toISOString(),
elapsed_seconds: Math.round((Date.now() - startTime) / 1000)
}
});
} else {
processedCount = Math.floor(totalProducts * 0.4);
outputProgress({
status: 'running',
operation: 'Skipping base product metrics calculation',
current: processedCount || 0,
total: totalProducts || 0,
elapsed: formatElapsedTime(startTime),
remaining: estimateRemaining(startTime, processedCount || 0, totalProducts || 0),
rate: calculateRate(startTime, processedCount || 0),
percentage: (((processedCount || 0) / (totalProducts || 1)) * 100).toFixed(1),
timing: {
start_time: new Date(startTime).toISOString(),
end_time: new Date().toISOString(),
elapsed_seconds: Math.round((Date.now() - startTime) / 1000)
}
});
lastPid = batch[batch.length - 1].pid;
processedCount += batch.length;
outputProgress({
status: 'running',
operation: 'Processing base metrics batch',
current: processedCount,
total: totalProducts,
elapsed: formatElapsedTime(startTime),
remaining: estimateRemaining(startTime, processedCount, totalProducts),
rate: calculateRate(startTime, processedCount),
percentage: ((processedCount / totalProducts) * 100).toFixed(1),
timing: {
start_time: new Date(startTime).toISOString(),
end_time: new Date().toISOString(),
elapsed_seconds: Math.round((Date.now() - startTime) / 1000)
}
});
}
// Calculate forecast accuracy and bias in batches
lastPid = 0;
while (true) {
if (isCancelled) break;
const [batch] = await connection.query(
'SELECT pid FROM products WHERE pid > ? ORDER BY pid LIMIT ?',
[lastPid, BATCH_SIZE]
);
if (batch.length === 0) break;
await connection.query(`
UPDATE product_metrics pm
JOIN (
SELECT
sf.pid,
AVG(CASE
WHEN o.quantity > 0
THEN ABS(sf.forecast_units - o.quantity) / o.quantity * 100
ELSE 100
END) as avg_forecast_error,
AVG(CASE
WHEN o.quantity > 0
THEN (sf.forecast_units - o.quantity) / o.quantity * 100
ELSE 0
END) as avg_forecast_bias,
MAX(sf.forecast_date) as last_forecast_date
FROM sales_forecasts sf
JOIN orders o ON sf.pid = o.pid
AND DATE(o.date) = sf.forecast_date
WHERE o.canceled = false
AND sf.forecast_date >= DATE_SUB(CURRENT_DATE, INTERVAL 90 DAY)
AND sf.pid IN (?)
GROUP BY sf.pid
) fa ON pm.pid = fa.pid
SET
pm.forecast_accuracy = GREATEST(0, 100 - LEAST(fa.avg_forecast_error, 100)),
pm.forecast_bias = GREATEST(-100, LEAST(fa.avg_forecast_bias, 100)),
pm.last_forecast_date = fa.last_forecast_date,
pm.last_calculated_at = NOW()
WHERE pm.pid IN (?)
`, [batch.map(row => row.pid), batch.map(row => row.pid)]);
lastPid = batch[batch.length - 1].pid;
}
}
if (isCancelled) return {
processedProducts: processedCount || 0,
processedOrders: processedOrders || 0,
processedPurchaseOrders: 0, // This module doesn't process POs
success
};
// Calculate product time aggregates
if (!SKIP_PRODUCT_TIME_AGGREGATES) {
outputProgress({