run-pytest

name	run-pytest
description	Run the pytest suite for the Israeli supermarket parsers repo, locally or via Docker. Use when running tests, checking test output, or verifying a specific chain passes. Covers fast path, Docker path, and exhaustive (unlimited samples) mode.

Run pytest

Fast path (use first)

mkdir -p temp
python -m pytest il_supermarket_parsers/parsers/tests/test_all.py -vv -n auto 2>&1 | tee temp/pytest-run.txt

• -n auto — pytest-xdist picks worker count from os.cpu_count().
• Output lands in temp/ (same mount CI uses).

Run a single chain

python -m pytest il_supermarket_parsers/parsers/tests/test_all.py -vv -k "<ChainName>" 2>&1 | tee temp/pytest-run.txt

---

Docker path (use when fast path passes but CI still fails)

Mirrors .github/workflows/test-suite.yml exactly:

docker build -t erlichsefi/israeli-supermarket-parsers:test --target test .
mkdir -p temp
docker run --rm -v ./temp:/usr/src/app/temp erlichsefi/israeli-supermarket-parsers:test 2>&1 | tee temp/pytest-run.txt

Inside the image, CMD uses -n 2; only the local fast path above uses -n auto.

---

Exhaustive mode (new source or uncertain parser)

By default each test downloads 10 sample files (NUM_SAMPLES=10). To fetch every file from a source, temporarily patch il_supermarket_parsers/parsers/tests/test_case.py lines 106-109:

Before:

try:
    self.limit = int(os.environ.get("NUM_SAMPLES", 10))
except ValueError:
    self.limit = 10

After:

self.limit = None

Then run only the affected chain:

python -m pytest il_supermarket_parsers/parsers/tests/test_all.py -vv -k "<ChainName>" 2>&1 | tee temp/pytest-run-full.txt

Revert before committing — limit = None causes CI to download unbounded data and time out.

---

Scan for failures

grep -E "^FAILED|ERROR" temp/pytest-run.txt

Each failure line: FAILED test_all.py::ChainTestCase::test_name.