Tests and pylint

Running

$ sh tests/check.sh

runs the test.sh script and runs pylint with MTH’s preferred configuration (in dev/pylintrc)

maintainer-tests

A few of the tests in the peyotl test suite (including some tiggered by the maintainer-test.sh but not by python setup.py test). require more setup than we expect a typical user to want to do. In particular, you need some phylesystem-style repos on which you can perform git actions - including pushes to remotes.

$ bash tests/maintainer-tests.sh

invokes the richer test suite. However, these tests require some set up.

The tests of logging require that you do not have a ~/.peyotl/config file. See the configuration page for details on how to use environmental variables instead of the default config file location.

tests requiring local clones of mini testing repos

Running

$ bash bootstrap.sh

should clone the testing repos that you need to run tests on the phyleystem, amendments, and collections tests.

If you do not run that command, you will seem some tests skipped because of lack of mini_system or mini_phyl mini_amendments and mini_collections. Skipping these tests is normal is normal. Mark and Emily Jane have some unittests of git interactions that require privileged access to a particular testing repository.

If you are interested in running these “maintainer” tests you can run ‘bootstrap.sh’ file at the top of the peyotl repo. Or you can do the setup manually by:

1. fork https://github.com/mtholder/mini_phyl and
   https://github.com/mtholder/mini_system on github (or some other
   server on which you have write permissions).

2. From the top of the phylesystem repo directory:

mkdir peyotl/test/data/mini_par
cd peyotl/test/data/mini_par

3. Used `git clone ...` to create clones of mini_phyl and mini_system
    in this directory.

At that point you should not see tests in the peyotl being skipped with messages that refer to Peyotl not being configured for maintainer tests.

If you have not checked out these testing repos with the git protocol then the push-to-mirror operations will not be tested.

Testing wrappers of web-services

Put RUN_WEB_SERVICE_TESTS in your environment if you want the tests that call open tree of life web services to run. Without this variable, those tests are skipped.

Some tricks

These are probably mainly of interest to people who develop peyotl

curl calls for web-services

When debugging or writing issues, it is nice to have a curl version of a web service call. If PEYOTL_CURL_LOG_FILE is in the env when peyotl is executing, the api wrappers will write a curl version of their activity to the filepath indicated by that variable. These files get big, so you probably don’t want to have this on by default. MTH adds the following 3 functions to bash when developing peyotl; then log-peyotl-curl turns on logging (and flushes the previous log!), cat-peyotl-curl shows the logged commands, and stop-log-peyotl-curl turns off logging.

function log-peyotl-curl {
    export PEYOTL_CURL_LOG_FILE=/tmp/peyotl-curl-log.txt
    if test -f "$PEYOTL_CURL_LOG_FILE"
    then
        rm "$PEYOTL_CURL_LOG_FILE"
    fi
}
function cat-peyotl-curl {
    if ! test -f "$PEYOTL_CURL_LOG_FILE"
    then
        cat /tmp/peyotl-curl-log.txt
    else
        cat "$PEYOTL_CURL_LOG_FILE"
    fi
}

function stop-log-peyotl-curl {
    unset PEYOTL_CURL_LOG_FILE
}



$ sh tests/maintainer-test.sh

Roundtrip tests

The following tests are triggered by running bash tests/integration-tests.sh

A test of the available format conversions (without NeXML validation) can be run with:

$ sh peyotl/test/check_nexson_nexml_clique.sh peyotl/test/data/nexson/otu.json scripts/nexson/nexson_nexml.py

If you alias your nexml validation tool to the name “validate-nexml” then you can run the check_nexml_roundrip.sh and check_nexson_roundrip.sh

Other dependencies for these test scripts are xmllint and saxon-xslt. Note that these are not dependencies for normal usage of

Caveat: check_nexml_roundrip.sh will fail if the attribute order differs from the order used by nexson_nexml.py

validate-nexml command.

MTH’s validate-nexml is shell script:

#!/bin/sh
java -jar "${NEXML_PARENT}/xml-validator/target/xml-validator-1.0-SNAPSHOT-jar-with-dependencies.jar" -s "${NEXML_PARENT}/nexml/xsd/nexml.xsd" $@

where xml-validator is a compiled clone of https://github.com/wiztools/xml-validator (you have to run mvn package to build) and nexml is a clone of https://github.com/nexml/nexml

You can tweak this by deciding on your NEXML_PARENT dir and running:

$ cd "${NEXML_PARENT}"
$ svn checkout http://xml-validator.googlecode.com/svn/trunk/ xml-validator-read-only
$ git clone https://github.com/nexml/nexml.git
$ cd xml-validator-read-only
$ mvn package