Skip to content

Latest commit

 

History

History
49 lines (43 loc) · 2.75 KB

TESTING.md

File metadata and controls

49 lines (43 loc) · 2.75 KB
Raw

Testing is Good

We can't know that our software works unless we test it thoroughly, so let us make sure this indexing service is well tested. For this particular tool, testing is rather difficult, however, because of the need for manual database interactions. To ensure that testing is still properly carried out, we have compiled a script that can be manually run to test all the parts of the system.

There are three scripts which facilitate manual testing.

  1. reset_indexing_state.py: Removes all occurrences, words, and relationships from the database.
  2. show_words_in_document.py: Prints the list of each word in the database.
  3. sanity_check.py: Compares the word count and line count of files (as calculated by the importer and tokenizer) with their corresponding entries in the occurrences database (line count uses the e_eol field). Using this tool, you can verify that the values are the same, which means that the indexer is likely performing correctly.

How to test

  1. Upload the small-* files from the test_files folder to the database (you can actually upload any files you like - the more testing the merrier). This will populate the database with files so that the indexer will actually have work to do.
  2. python reset_indexing_state.py - resets the database and removes the index_events file.
  3. python main.py - runs the indexer. This process does not exit, so you will need to have another terminal session to continue the testing procedure.
  4. python show_words_in_document.py in another terminal. This should output all the different words in each of the files you uploaded. Using small test files makes this verification feasible. This step will be useful when improving the tokenization part of the system.
  5. python sanity_check.py - To check that the indexing did what it was supposed to do.
  6. Now, to make sure that updates work well, delete one of the files you uploaded, and rerun the sanity check. Note that the words inside the database are not removed currently. This may be the wrong behavior, but it is not too significant since the number of unique words is dwarfed by the number of occurrences of words. After deletion of the file, the main process should have logged a message saying that a particular document was removed.
  7. Re-upload the deleted file. This should cause the main process to spend some time indexing a new document. Run the sanity check again.
  8. Now modify one of the files from the Kardia CRM UI. The main process should reindex the file. Run the sanity check and make sure that the numbers of occurrences and lines updated how you edited.

This script should hopefully not take too much time to complete. You can modify it and cut out some parts when you only want to test certain features.