Prepare testing corpus about log-anomaly-detector

[MichaelClifford]

PR #38 adds an additional directory validation_data that provides a script and README instructions for users to prepare a testing corpus with data from their own Elasticsearch instance and how to run the tool locally.

The main script generate_validation_data.py randomly replaces some of the log entries with random strings and generates a corresponding labels file.

The next step would be to include an automated validation checker that outputs the results of a training testing loop. (current prototype of this is in process in a jupyter notebook)

[zak-hassan]

I think the goal is to have more unit test to test those use cases @vpavlin instead of generating so much random data. For now the goal is to have 1 example of sample data we can utilize to test that this code works.

[MichaelClifford]

@zmhassan, I agree. But I think we also want a single testing courpus and automated validation tests as a means of measuring the performance of models as we tune parameters, or implement different encoding and classification approaches.

[zak-hassan]

@MichaelClifford I believe you added some test data into validation_data/Hadoop_2k.json. So I'm assuming we can close this issue.