bakery-admin/bakery-ia
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Improve teh securty of teh DB

Browse Source
This commit is contained in:
Urtzi Alfaro
2025-10-19 19:22:37 +02:00
parent 62971c07d7
commit 05da20357d
87 changed files with 7998 additions and 932 deletions
Download Patch File Download Diff File Expand all files Collapse all files

60
services/training/app/ml/trainer.py
View File

@@ -119,9 +119,46 @@ class EnhancedBakeryMLTrainer:
logger.info("Multiple products detected for training",
products_count=len(products))
# Event 1: Training Started (0%) - update with actual product count
# Note: Initial event was already published by API endpoint, this updates with real count
await publish_training_started(job_id, tenant_id, len(products))
# Event 1: Training Started (0%) - update with actual product count AND time estimates
# Calculate accurate time estimates now that we know the actual product count
from app.utils.time_estimation import (
calculate_initial_estimate,
calculate_estimated_completion_time,
get_historical_average_estimate
)
# Try to get historical average for more accurate estimates
try:
historical_avg = await asyncio.get_event_loop().run_in_executor(
None,
get_historical_average_estimate,
db_session,
tenant_id
)
avg_time_per_product = historical_avg if historical_avg else 60.0
logger.info("Using historical average for time estimation",
avg_time_per_product=avg_time_per_product,
has_historical_data=historical_avg is not None)
except Exception as e:
logger.warning("Could not get historical average, using default",
error=str(e))
avg_time_per_product = 60.0
estimated_duration_minutes = calculate_initial_estimate(
total_products=len(products),
avg_training_time_per_product=avg_time_per_product
)
estimated_completion_time = calculate_estimated_completion_time(estimated_duration_minutes)
# Note: Initial event was already published by API endpoint with estimated product count,
# this updates with real count and recalculated time estimates based on actual data
await publish_training_started(
job_id=job_id,
tenant_id=tenant_id,
total_products=len(products),
estimated_duration_minutes=estimated_duration_minutes,
estimated_completion_time=estimated_completion_time.isoformat()
)
# Create initial training log entry
await repos['training_log'].update_log_progress(
@@ -135,10 +172,25 @@ class EnhancedBakeryMLTrainer:
)
# Event 2: Data Analysis (20%)
# Recalculate time remaining based on elapsed time
elapsed_seconds = (datetime.now(timezone.utc) - repos['training_log']._get_start_time(job_id) if hasattr(repos['training_log'], '_get_start_time') else 0) or 0
# Estimate remaining time: we've done ~20% of work (data analysis)
# Remaining 80% includes training all products
products_to_train = len(processed_data)
estimated_remaining_seconds = int(products_to_train * avg_time_per_product)
# Recalculate estimated completion time
estimated_completion_time_data_analysis = calculate_estimated_completion_time(
estimated_remaining_seconds / 60
)
await publish_data_analysis(
job_id,
tenant_id,
f"Data analysis completed for {len(processed_data)} products"
f"Data analysis completed for {len(processed_data)} products",
estimated_time_remaining_seconds=estimated_remaining_seconds,
estimated_completion_time=estimated_completion_time_data_analysis.isoformat()
)
# Train models for each processed product with progress aggregation