Skip to content
Tests

Update dependencies #59

Sign in to view logs

Update dependencies

Update dependencies #59

Workflow file for this run

.github/workflows/tests.yml at 5513af9
name: Tests
on:
push:
branches:
- main
pull_request:
jobs:
tests:
name: Tests
runs-on: ubuntu-latest
steps:
- name: Install poppler-utils and docx2txt
run: |
sudo apt-get update
sudo apt-get install -y poppler-utils docx2txt
- name: Checkout code
uses: actions/checkout@v4
- name: Install Poetry
uses: snok/install-poetry@v1
- name: Setup Python
uses: actions/setup-python@v5
with:
python-version-file: 'pyproject.toml'
cache: 'poetry'
- name: Install Poetry dependencies
run: poetry install
- name: Write SERVICE_ACCOUNT_CREDENTIALS to service_account_credentials.json
uses: jsdaniell/[email protected]
with:
name: "service_account_credentials.json"
json: ${{ secrets.SERVICE_ACCOUNT_CREDENTIALS }}
- name: Test Tahweel on 1 page PDF file
run: |
poetry run tahweel ".github/data/test-case-1-page.pdf" --service-account-credentials service_account_credentials.json
test -f ".github/data/test-case-1-page.txt" || exit 1
test -f ".github/data/test-case-1-page.docx" || exit 1
grep -q "^بسم الله الرحمن الرحيم$" ".github/data/test-case-1-page.txt" || exit 1
rm -f .github/data/*.txt .github/data/*.docx
- name: Test Tahweel on 3 pages PDF file
run: |
poetry run tahweel ".github/data/test-case-3-pages.pdf" --service-account-credentials service_account_credentials.json
test -f ".github/data/test-case-3-pages.txt" || exit 1
test -f ".github/data/test-case-3-pages.docx" || exit 1
grep -o "PAGE_SEPARATOR" ".github/data/test-case-3-pages.txt" | wc -l | grep -q "^2$" || exit 1
head -n 1 ".github/data/test-case-3-pages.txt" | grep -q "^بسم الله الرحمن الرحيم$" || exit 1
tail -n 1 ".github/data/test-case-3-pages.txt" | grep -q "^والصلاة والسلام على أشرف الأنبياء والمرسلين$" || exit 1
rm -f .github/data/*.txt .github/data/*.docx
- name: Test Tahweel on multiple PDF files
run: |
poetry run tahweel ".github/data/test-case-1-page.pdf" ".github/data/test-case-3-pages.pdf" --service-account-credentials service_account_credentials.json
test -f ".github/data/test-case-1-page.txt" || exit 1
test -f ".github/data/test-case-1-page.docx" || exit 1
test -f ".github/data/test-case-3-pages.txt" || exit 1
test -f ".github/data/test-case-3-pages.docx" || exit 1
rm -f .github/data/*.txt .github/data/*.docx
- name: Test Tahweel on a directory with --dir-output-type tree_to_tree
run: |
poetry run tahweel ".github/data/" --service-account-credentials service_account_credentials.json --dir-output-type tree_to_tree
test -f ".github/data - Tahweel TXT/test-case-1-page.txt" || exit 1
test -f ".github/data - Tahweel DOCX/test-case-1-page.docx" || exit 1
test -f ".github/data - Tahweel TXT/test-case-3-pages.txt" || exit 1
test -f ".github/data - Tahweel DOCX/test-case-3-pages.docx" || exit 1
test -f ".github/data - Tahweel TXT/test-case-multilines.txt" || exit 1
test -f ".github/data - Tahweel DOCX/test-case-multilines.docx" || exit 1
test -f ".github/data - Tahweel TXT/test-case.txt" || exit 1
test -f ".github/data - Tahweel DOCX/test-case.docx" || exit 1
rm -rf ".github/data - Tahweel TXT" ".github/data - Tahweel DOCX"
- name: Test Tahweel on a directory with --dir-output-type side_by_side
run: |
poetry run tahweel ".github/data/" --service-account-credentials service_account_credentials.json --dir-output-type side_by_side
test -f ".github/data/test-case-1-page.txt" || exit 1
test -f ".github/data/test-case-1-page.docx" || exit 1
test -f ".github/data/test-case-3-pages.txt" || exit 1
test -f ".github/data/test-case-3-pages.docx" || exit 1
test -f ".github/data/test-case-multilines.txt" || exit 1
test -f ".github/data/test-case-multilines.docx" || exit 1
test -f ".github/data/test-case.txt" || exit 1
test -f ".github/data/test-case.docx" || exit 1
rm -f .github/data/*.txt .github/data/*.docx
- name: Test Tahweel on 3 pages PDF file with --txt-page-separator ANYTHING
run: |
poetry run tahweel ".github/data/test-case-3-pages.pdf" --service-account-credentials service_account_credentials.json --txt-page-separator ANYTHING
test -f ".github/data/test-case-3-pages.txt" || exit 1
test -f ".github/data/test-case-3-pages.docx" || exit 1
grep -o "ANYTHING" ".github/data/test-case-3-pages.txt" | wc -l | grep -q "^2$" || exit 1
rm -f .github/data/*.txt .github/data/*.docx
- name: Test Tahweel on multilines PDF file with --docx-remove-newlines
run: |
poetry run tahweel ".github/data/test-case-multilines.pdf" --service-account-credentials service_account_credentials.json --docx-remove-newlines
test -f ".github/data/test-case-multilines.txt" || exit 1
test -f ".github/data/test-case-multilines.docx" || exit 1
docx2txt ".github/data/test-case-multilines.docx" - | grep -q "^بسم الله الرحمن الرحيم والصلاة والسلام على أشرف الأنبياء والمرسلين$" || exit 1
rm -f .github/data/*.txt .github/data/*.docx
- name: Test Tahweel on 1 page PDF file with --output-formats txt
run: |
poetry run tahweel ".github/data/test-case-1-page.pdf" --service-account-credentials service_account_credentials.json --output-formats txt
test -f ".github/data/test-case-1-page.txt" || exit 1
test ! -f ".github/data/test-case-1-page.docx" || exit 1
rm -f .github/data/*.txt .github/data/*.docx
- name: Test Tahweel on 1 page PDF file with --output-dir ".github/custom-output-dir"
run: |
poetry run tahweel ".github/data/test-case-1-page.pdf" --service-account-credentials service_account_credentials.json --output-dir ".github/custom-output-dir"
test -f ".github/custom-output-dir/test-case-1-page.txt" || exit 1
test -f ".github/custom-output-dir/test-case-1-page.docx" || exit 1
rm -rf .github/custom-output-dir
- name: Test Tahweel on a directory with --dir-output-type tree_to_tree --output-dir ".github/custom-output-dir"
run: |
poetry run tahweel ".github/data/" --service-account-credentials service_account_credentials.json --dir-output-type tree_to_tree --output-dir ".github/custom-output-dir"
test -f ".github/custom-output-dir/Tahweel TXT/test-case-1-page.txt" || exit 1
test -f ".github/custom-output-dir/Tahweel DOCX/test-case-1-page.docx" || exit 1
test -f ".github/custom-output-dir/Tahweel TXT/test-case-3-pages.txt" || exit 1
test -f ".github/custom-output-dir/Tahweel DOCX/test-case-3-pages.docx" || exit 1
test -f ".github/custom-output-dir/Tahweel TXT/test-case-multilines.txt" || exit 1
test -f ".github/custom-output-dir/Tahweel DOCX/test-case-multilines.docx" || exit 1
test -f ".github/custom-output-dir/Tahweel TXT/test-case.txt" || exit 1
test -f ".github/custom-output-dir/Tahweel DOCX/test-case.docx" || exit 1
rm -rf .github/custom-output-dir
- name: Test Tahweel on a directory with --dir-output-type side_by_side --output-dir ".github/custom-output-dir"
run: |
poetry run tahweel ".github/data/" --service-account-credentials service_account_credentials.json --dir-output-type side_by_side --output-dir ".github/custom-output-dir"
test -f ".github/custom-output-dir/test-case-1-page.txt" || exit 1
test -f ".github/custom-output-dir/test-case-1-page.docx" || exit 1
test -f ".github/custom-output-dir/test-case-3-pages.txt" || exit 1
test -f ".github/custom-output-dir/test-case-3-pages.docx" || exit 1
test -f ".github/custom-output-dir/test-case-multilines.txt" || exit 1
test -f ".github/custom-output-dir/test-case-multilines.docx" || exit 1
test -f ".github/custom-output-dir/test-case.txt" || exit 1
test -f ".github/custom-output-dir/test-case.docx" || exit 1
rm -rf .github/custom-output-dir
- name: Test Tahweel on an image file
run: |
poetry run tahweel ".github/data/test-case.jpg" --service-account-credentials service_account_credentials.json
test -f ".github/data/test-case.txt" || exit 1
test -f ".github/data/test-case.docx" || exit 1
rm -f .github/data/*.txt .github/data/*.docx