Skip to content

Latest commit

 

History

History
379 lines (305 loc) · 18.6 KB

README.md

File metadata and controls

379 lines (305 loc) · 18.6 KB
Raw

PYTEST TESTING TOOL (PTT)

What is a Pytest

Simply put, a Pytest is a pass or fail Python test. For instance, with the WCS step, we have Python scripts (which we are calling auxiliary code within the frame of the testing tool) that compare the pipeline product with the ESA corresponding intermediary file, and calculates a difference. The Pytest is to asses if that difference is less than or equal to an X threshold value. Hence, a failed test means that the condition was not met. If an error should occur with the pipeline, the test will be flagged as an error.

Quick Start Guide

NOTE: This guide assumes that Conda has been installed. If you have not yet done so, please follow the instructions at: https://astroconda.readthedocs.io/en/latest/ Please use python 3.5

  1. Create the conda environment for testing and get the configuration files.

a. Conda environment for this testing campaign:

  • Current testing version is 7.1 for release candidate 0.7.9.0. However, this version has a broken version of the assign WCS step. Hence, we are currently testing with the development version of the pipeline (please see the NOTE below).
  • In case you need to test some specific step of release candidate 0.7.9.0, please follow these instructions to create the build7.1 environment. In a terminal type:
conda create -n jwst_b7.1 --file url_depending_on_your_system python=3.5

for the current release candidate, the ulr options are:

NOTE: If you need to use the development version of the pipeline then do the following:

conda create -n jwst_dev -c http://ssb.stsci.edu/conda-dev jwst python=3.5

Then, to update the development environment, activate the environment and then type:

conda update --override-channels -c http://ssb.stsci.edu/conda-dev -c defaults --all

b. Configuration files corresponding to this build. Create a directory (e.g. b7.1cfg_files) somewhere in your testing working space, and cd into it. Now type the following command within the conda environment you just created (see step 2).

collect_pipeline_cfgs .
  1. Activate the conda environment for testing the pipeline, e.g. type:
source activate jwst_b7.1

From here on, every step of this guide should happen within the conda testig environment.

NOTE:

  • If you forget what did you name your new environment type:
conda env list

this will list all environments you have created.

  • If you want to remove an old testing environment type:
conda env remove -n name_of_your_old_environment
  1. Install the pytest html plugin. Within the conda testing environment, type:
pip install pytest-html

NOTE: Every time you create a new conda environment you need to install html plugin.

  1. Clone the repository so you have the PTT. To do this click at the top right of this page, in the dropdown button that says clone or download, then copy the ulr that appears there. Now, within the conda testing environment, go to or create the directory where you want the PTT to "live" in. However, make sure that the configuration files directory is at least at the top level of the directory tree where the PTT will live, e.g. the b7.1cfg_files directory and the nirspec_testing_tool directory can be at the same level, but the b7.1cfg_files directory cannot be inside the nirspec_testing_tool directory because otherwise the .cfg files will be picked up by Git and it will try to put them in your space.
git clone the_ulr_you_copied

After this is done, you should see a full copy of the PTT in your directory.

NOTE:

  • If you have already cloned the repository, in the terminal go to where you placed the nirspec_pipe_testing_tool directory. Then, use the following command to update the code: git pull
  • If, however, you had written script(s) in the tool's directory tree, git will not let you continue until you move the script(s) to another directory.
  1. Prepare the data to run through the pipeline. To do this:

a. Copy the test data you will use from the NIRSpec vault directory. Go to the directory where you want to have the testing data, and from there type:

cp /grp/jwst/wit4/nirspec_vault/prelaunch_data/testing_sets/b7.1_pipeline_testing/
                                             test_data_suite/the_data_you_want_to_copy .

b. In the directory where you copied the test data, you will need to run a script PER fits file you want to test. Do not worry, this will only be done once. This script will create a new subdirectory with the necessary input file to run the SSB script that converts raw data into uncal type files. You can choose to either keep this subdirectory, or tell the script to remove it after the operation is done. In the terminal type:

python /path_to_the_testing_tool/nirspec_pipe_testing_tool/utils/
                                        prepare_data2run.py fits_file.fits MODE -u

where the MODE is expected to be one of: FS, MOS, IFU. This command will update the uncal keyword header without creating a new file, and will also keep the subdirectory. To remove it, simply add -rm at the end. To save the keyword changes in a new fits file (instead of updating), remove the -u. The new uncal fits file is now ready for pipeline ingest.

c. Optional. If you want to see the header of any file, you can use the another script in the utils directory of the PTT. If you just want to see on-screen the header, go where your fits file "lives" and type:

python /path_to_the_testing_tool/nirspec_pipe_testing_tool/utils/
                                                        read_hdr.py fits_file.fits

This command will show the main header. To save the header to a text file add a -s at the end. If you want to see/save a different extension add at the end -e=1 for extension 1, and so on.

d. Now, the data is ready to be ran through cal_detector1. Please go ahead with step 6 of this guide to do that.

  1. Set the PTT configuration file. In a text file editor, you are going to modify the configuration file named cwspec2_config.cfg, which lives at /nirspec_pipetesting_tool/calwebb_spec2_pytests/. This is the file that controls all the input that the tool needs. Please open it and make sure that:
  • All the paths point to the right places. The files can be located anywhere, but both, the pipeline and the tool, will run faster if the files are local on your computer.
  • The input file for the PTT is the final output file from calwebb_detector1.
  • The adequate mode for the data to be tested is set correctly, choices are: FS, IFU, or MOS.
  • The steps that you want to be ran or not are set to True or False.
  • In the bottom part of the file, all the additional arguments for the PTT are correct, e.g. threshold values, figure switches, and additional fits files.
  1. Run the calwebb_detector1 pipeline. The final output of this is the level 2 data required to run the PTT. In a terminal, go into the directory where the testing tool lives (i.e. at the level of the calwebb_spec2_pytests directory), and make sure that the testing conda environment is on. Now type:
python ../utils/run_cal_detector1.py /path_where_the_uncal_file_lives/uncal_file.fits -sbs

This script will execute the calwebb detector 1 pipeline step by step. If you want to run it in a single run, using the configuration file that you have for it in the utils directory, simply remove the -sbs from previous command. If everything went well, you will see a text file called cal_detector1_outputs_and_times.txt, which contains the steps ran, the name of the output fits file, and the time each step took to run. However, if you chose to run the calwebb detector1 in a single run, you will only see the total running time. This text file, along with the intermediary products will be located in the path you set for the working_directory variable in the configuration file of the PTT.

You now are able to run the MESA calwebb detector 1 testing tool. Steps to obtain the MESA testing tool:

a. At the same level as the top directory of the PTT (i.e. the nirspec_pipe_testing_tool directory), create a new directory called MESA_cal_detector1.

b. Inside MESA_cal_detector1, you will clone their Git repository. Please follow the directions for this at: http://calibration-pipeline-testing-tool.readthedocs.io/en/latest/

c. Now, in the utils directory of nirspec_pipe_testing_tool, you will find a sample json file that you can modify in order to use as input for the MESA calwebb detector 1 testing tool. Notice that it differs from the example given in the MESA documentation. In our example, the steps are in the order of the pipeline, to make it easier to determine what file is input to which step. Please copy our sample json file from the utils directory into the MESA_cal_detector1 directory you created, and or modify the json file with the name and path of your intermediary fits products from running the calwebb_detector1 pipeline.

d. Run the MESA testing tool with the following command:

test_pipeline --config ./cal_detector1_input.json

Please record your progress. Make a PDF export of the html report, and place it in the corresponding column of Table 2 of our testing campaign Confluence page: https://confluence.stsci.edu/display/JWST/NIRSpec+Pipeline+Testing+Build+7.1+part+1

  1. Ready to run PTT. Go back to the directory where PTT lives and into the calwebb_spec2_pytests directory, copy final output file from calwebb detector1 into the working directory you indicated in the cwspec2_config.cfg file, and make sure that the input file for the PTT matches the file you just copied into the working directory. Now, to ensure that everything is in order, and to see what pytests will be executed and in which order type:
pytest --collect-only
  1. Do the first PTT run. As an output of the testing tool you will see an html file, report.html, and an intermediary product file name map will appear in the calwebb_spec2_pytests directory. The fits files of intermediary products will be saved in the path you indicated at the cwspec2_config.cfg file with the variable working_directory. In the terminal type:
pytest -s --config_file=cwspec2_config.cfg --html=report.html

The -s will capture all the print statements in the code on screen.

  1. Report your findings. If all went well, you should have the html report in the calwebb_spec2_pytests directory, along with the file name map of intermediary products that appeared in the working directory. Each section of the report specifies what tests passed or failed, and, if there were any errors, the report will tell you (with the name of the script where the error occurred) if it is a pipeline error or a PTT error. Keep updating your testing progress in our testing campaign Confluence page: https://confluence.stsci.edu/display/JWST/NIRSpec+Pipeline+Testing+Build+7.1+part+1

Please follow these actions for reporting your progress, depending on the outcome:

  • Do not place the html file in the Confluence page because it will get corrupted. Instead, please create a PDF export from the report and use that file for the Confluence page and for any sharing purposes.
  • If there are no testing tool errors and some of the pytests failed, you can check
    off the steps and please put that report in Table 2 of the Confluence page.
  • If there are testing tool errors please place your report in the Confluence page, add a small comment in the "Testing Tool" column of Table 2 in the Confluence page, and send an email to Maria Pena-Guerrero ([email protected]), so that the PTT error can be addressed as soon as possible.
  • If there are errors with the pipeline, please create a Basecamp message (https://stsci-ins.basecamphq.com/projects/11477312-jwst-pipeline/posts), following the NIRSpec guidelines for doing so, which you can find at: https://confluence.stsci.edu/display/JWST/NIRSpec+Guidelines+for+Pipeline+Issue+Reporting and copy the link in the corresponding column of Table 2 of the Confluence page.
  • Keep updating the html report as you continue with the testing, and including the final report.
  1. Put final results in the NIRSpec vault. At the end of the testing campaign, please do the following within your working directory:

a. Create a new directory called yourname_MODE_Name_of_the_raw_data_file_results, e.g. maria_IFU_NRSSMOS-MOD-G1M-17-5344175105_1_491_SE_2015-12-10T18h00m06_results. (If you already copied a previous version of your results and you want to keep it then add a _v2 -or the version number that corresponds- after the word results, e.g. maria_IFU_NRSSMOS-MOD-G1M-17-5344175105_1_491_SE_2015-12-10T18h00m06_results_v2)

b. Inside that new direcoty place all the intermediary fits products as well as the html report

c. Also inside that directory, create a text file named YourNameMODEresults.txt, e.g. mariaIFU_NRSSMOS-MOD-G1M-17-5344175105_1_491_SE_2015-12-10T18h00m06_results.txt. In this text file you will only type the full path where you obtained the testing data, e.g.

/grp/jwst/wit4/nirspec_vault/prelaunch_data/testing_sets/b7.1_pipeline_testing/
                                                              test_data_suite/IFU_CV3

d. Finally, place the results directory you created in the staging directory of the NIRSpec vault, and send Gray an email ([email protected]). The path of the staging directory is: /grp/jwst/wit4/nirspec_vault/staging

TO KEEP IN MIND

  • A text file containing an intermediary product name map will be created in the pytests directory.
  • If any of the central store directory calls do not respond (e.g. when looking at the flats), the pytest will be skipped even if the step is set to True in the config file. The failing message will say that the step was set to False (this is a known bug). To force the tests run, you will have to download the files the tool is calling, and change the corresponding paths in the configuration file.
  • The output in the terminal can be a bit overwhelming if there was a failed test or an error, since it shows both, the pipeline messages and the PTT messages. In the html report is much clearer to understand what happened.
  • As part of the testing campaign, it is important that you run the pipeline from the command line as well, and that you make sure that the outcome intermediary files are consistent with the ones ran with scripts, i.e. the PTT. This sanity check is minor but important to verify. In the utils directory there are two text files named terminal_commands_calwebb_detector1_steps.txt and terminal_commands_calwebb_spec2_steps.txt. These files contain all the commands you can use from the terminal to run the calwebb_detector1 and calwebb spec2 steps, respectively.
  • Finally, remember that:

a. Whenever you need to read either the main or science headers of a file, you can always use the read_hdr.py script located in the utils directory of the testing tool.

b. If you need to change/add a keyword value to a specific extension of a file, you can use the change_keywd.py script, also located at the utils directory.

POSSIBLE OUTCOMES OF THE PYTESTS

  • Passed = the assertion was true, so the test condition was met.
  • Failed = the assertion was false, the test condition was NOT met (there will be an AssertionError on-screen and in the html file, with a clear PTT customized message of what happened).
  • Skipped = the test was skipped (there will be a message on the html report and on-screen explaining why the test was skipped).
  • Error = this is a coding error, or a bug in either the pipeline or the PTT. Please follow the appropriate steps described in the "For Testers" section of our testing campaign Confluence page: https://confluence.stsci.edu/display/JWST/NIRSpec+Pipeline+Testing+Build+7.1+part+1

ADDING TESTING ROUTINES

You can add additional testing routines. We prefer if these are written in python 3.5, however, due to time constraints, we devised a method for including scripts in other programing languages. For this testing campaign, the guidelines for auxiliary code are the following:

  1. Programing language = preferably Python 3.5, however, other languages (e.g. IDL, C) are OK.

  2. Please comment as much as you can so that it is easy for other testers to follow your code.

  3. At the end of your script you must create a pass or fail test, e.g. after you calculate a root-mean-square check if that value is within some range. If instead you have a threshold value, please make the threshold an input variable with a default value.

  4. Your script must produce a text file named scriptname_result.txt at the end, which will only contain one word: True or False.

  5. Create a directory named scriptname with the following contents: a. A text file named scriptname_init.txt, and in there type in line 1 the pipeline step after which your code should be run, in line 2 the language you used, and in line 3 the exact command you would use to run the script. If this is an IDL script, line 2 should be the exact call from within the IDL session.

b. The script you wrote.

c. One level up from where you created the script directory, create a new directory named YourName_PTT_auxiliary_code. In this directory, please place the directory scriptname, and a create a text file named YourName_PTT_auxiliary_code.txt. This text file will only contain the following path: /grp/jwst/wit4/nirspec_vault/pipe_testing_tool/auxiliary_code

  1. Now, copy the YourName_PTT_auxiliary_code directory in the NIRSpec vault, and please send an email to Gray Kanarek ([email protected]). The vault's path is: /grp/jwst/wit4/nirspec_vault/staging.

  2. Your code will be implemented for the next testing campaign and if there is a threshold associated with it, it will appear as a variable in the PTT configuration file. For the current testing campaign, please place it in the "PTT Comments and/or errors" column of
    Table 2 of our progress keeping Confluence page: https://confluence.stsci.edu/display/JWST/NIRSpec+Pipeline+Testing+Build+7.1+part+1 You will be asked to write a description of your code for the final testing report of the current testing campaign.

Enjoy your pipeline testing!

If you have any question of what a specific step does, you can get a description at: http://jwst-pipeline.readthedocs.io/en/latest/jwst/pipeline/description.html#stage-2-spectroscopic-pipeline-step-flow-calwebb-spec2