Test nx-cugraph package instead of editable install (rapidsai#4442)

We already have the built package, so test that instead of an editable install. This helps avoid having to install the build dependencies for nx-cugraph at test time.

Authors:
  - Kyle Edwards (https://github.com/KyleFromNVIDIA)

Approvers:
  - Bradley Dice (https://github.com/bdice)
  - Ray Douglass (https://github.com/raydouglass)
  - Rick Ratzel (https://github.com/rlratzel)

URL: rapidsai#4442

ci/test_python.sh

@@ -101,10 +101,8 @@ rapids-logger "pytest nx-cugraph"
   --cov-report=term
 
 rapids-logger "pytest networkx using nx-cugraph backend"
-pushd python/nx-cugraph
-# Use editable install to make coverage work
-pip install -e . --no-deps
-./run_nx_tests.sh
+pushd python/nx-cugraph/nx_cugraph
+../run_nx_tests.sh
 # run_nx_tests.sh outputs coverage data, so check that total coverage is >0.0%
 # in case nx-cugraph failed to load but fallback mode allowed the run to pass.
 _coverage=$(coverage report|grep "^TOTAL")
@@ -114,17 +112,17 @@ echo $_coverage | awk '{ if ($NF == "0.0%") exit 1 }'
 # Run our tests again (they're fast enough) to add their coverage, then create coverage.json
 pytest \
   --pyargs nx_cugraph \
-  --config-file=./pyproject.toml \
-  --cov-config=./pyproject.toml \
+  --config-file=../pyproject.toml \
+  --cov-config=../pyproject.toml \
   --cov=nx_cugraph \
   --cov-append \
   --cov-report=
 coverage report \
   --include="*/nx_cugraph/algorithms/*" \
   --omit=__init__.py \
   --show-missing \
-  --rcfile=./pyproject.toml
-coverage json --rcfile=./pyproject.toml
+  --rcfile=../pyproject.toml
+coverage json --rcfile=../pyproject.toml
 python -m nx_cugraph.tests.ensure_algos_covered
 # Exercise (and show results of) scripts that show implemented networkx algorithms
 python -m nx_cugraph.scripts.print_tree --dispatch-name --plc --incomplete --different