microsoft/presidio

Presidio Analyser tests have high memory usage

hhobson opened this issue · 0 comments

hhobson commented

Describe the bug
Unable to run full Presido Analyser test suit on Github 2-Core (8GB RAM) Codespace due to high memory usage.

Max. memory usage was >11GB before #1424, now ~8GB (see memory usage graph below). Memray used to monitor memory usage.

Memray memory usage graph

To Reproduce
Steps to reproduce the behavior:

  1. On commit hash 2fe6ad70c75b56384bd495a1d61c71f17ba06b1d
  2. Run cd presidio-analyzer
  3. Run poetry run pip install memray
  4. Run poetry run memray run -m pytest
  5. Run memray flamegraph command given in memray run command output
  6. View flamegraph html file

Expected behavior
Test suits memory usage should be kept as low as possible

Screenshots
If applicable, add screenshots to help explain your problem.

Additional context
Add any other context about the problem here.