Hi everyone. I've been trying out Codeium in my local development setup for a few months now, mostly for Python and SQL work on our Airflow DAGs. It's been helpful for autocompleting boilerplate and even some tricky BigQuery window functions.
I just got an email about an updated privacy policy. The summary mentioned they are expanding the scope of telemetry data collected. This has me a bit worried. My main concern is about code snippets from our internal data pipelines being sent to their servers. Even if it's anonymized, some of our table names and transformation logic could be considered sensitive.
I tried to find specifics in the full policy, but it's a bit broad. Does anyone have a clearer understanding of what exactly is being collected now? For instance:
```python
# If I'm writing a DAG task that has our actual dataset names...
with DAG('sales_pipeline', schedule='@daily') as dag:
load_to_bq = BigQueryInsertJobOperator(
task_id='load_to_bq',
configuration={
"query": {
"query": "SELECT * FROM `my_project.prod_schema.raw_sales` WHERE ...",
"useLegacySql": False
}
}
)
```
Could strings like `my_project.prod_schema.raw_sales` end up in telemetry? I'm nervous because we're dealing with production data infrastructure, and I don't want to accidentally violate any internal policies by using a tool that's now more aggressive with data collection.
Are others in data engineering still comfortable using it? Have you switched to a more local/offline alternative for this reason? I'm looking for safe patterns here.