Automate Data Capture at Scale with Document AI: Challenge Lab Reviews

Titichaya V. · Reviewed 14 days ago

the scripts is using cloud run function gen1 but the instruction said we need to deploy gen2 so the code will not work

Prachya K. · Reviewed 14 days ago

siddhi k. · Reviewed 14 days ago

To the Google Cloud Skills Boost / Lab Content Team, I am writing to report that Task 4 and Task 5 of this lab are fundamentally broken and desperately need an immediate update. The provided instructions contain multiple critical errors that force students to spend hours debugging the lab's poorly maintained code rather than learning the actual GCP concepts. Here is a list of the critical bugs in the current lab instructions: Outdated Deployment Flags: The instructions use --trigger-resource=gs://... which is invalid for Cloud Run Functions Gen 2. It must be updated to --trigger-bucket. Missing Service Account (404 Error): The deploy command includes --service-account=${PROJECT_ID}@appspot.gserviceaccount.com, but this default App Engine SA is NOT provisioned in the lab environment, causing instant deployment failure. Wrong Bucket Name (403 Error): The deploy command references a bucket ending in -input-in, but the actual bucket we are instructed to create earlier ends in -input-invoices. This typo causes confusing Eventarc permission errors. Code Crash on Startup (Healthcheck Failed): The main.py code tries to parse a TIMEOUT environment variable (int(os.environ.get('TIMEOUT'))), but the provided .env.yaml file does not include it. This causes a NoneType error that crashes the container during the initial startup. Hardcoded Project ID (gRPC Permission Denied): The main.py file has the string YourGCPProjectID hardcoded instead of dynamically pulling the student's project ID. This causes a gRPC Permission Denied error when the function tries to write data to BigQuery. I had to manually run a sed command to fix your provided source code. Flawed File Copy Command: In Task 5, the command gsutil -m cp -r gs://cloud-training/gsp367/* ... copies Python scripts and YAML files into the input bucket alongside the PDFs. This immediately crashes the Document AI processor, which expects only document files. As a Data Engineer trying to upskill, it is extremely frustrating to waste time troubleshooting outdated lab environments and untested code. Please fix these issues so future learners don't have to face the same terrible experience. Regards, Benz

Thanachit S. · Reviewed 14 days ago

Ark A. · Reviewed 15 days ago

Martin K. · Reviewed 16 days ago

Peter F. · Reviewed 17 days ago

RAUSHAN Y. · Reviewed 17 days ago

Sam L. · Reviewed 19 days ago

tom v. · Reviewed 20 days ago

kitisak J. · Reviewed 20 days ago

Artist R. · Reviewed 21 days ago

Abbishek A. · Reviewed 22 days ago

Pranjal K. · Reviewed 22 days ago

KITAR C. · Reviewed 22 days ago

Sam L. · Reviewed 22 days ago

Sam L. · Reviewed 22 days ago

Pratiksha T. · Reviewed 23 days ago

Ankit G. · Reviewed 24 days ago

Chandrashekhar S. · Reviewed 24 days ago

Divya .. · Reviewed 26 days ago

Gregorio I. · Reviewed 27 days ago

Valentin G. · Reviewed 28 days ago

Anil K. · Reviewed 28 days ago

Emir L. · Reviewed 29 days ago