rbiswas4 / benchmarkinstancecatalogs

benchmarks for instance catalogs

BenchmarkInstanceCatalogs

benchmarks for instance catalogs

Adding examples

benchmarkinstancecatalogs's Issues

Allow overplotting in plotResults

This means providing kwargs to plotResults to set the format/color of points, and plotting on a old similar figure object.
While at it, also add a grid to the results.

Get Size of data packet

For the formats added to figure plots, Come up with saner names

Enable more random timings for queries per boundLen

To make the randomness more wieldly, rewrite boundLens using repeats before shuffle. We may also think about doing the sam to mjds

Add a timestamp to each query in the queries dataframe

Two things:

• Make self.df a more intuitive name like queries
• Convert time stamps at the start of a query to a string, and create a column in the dataframe holding this string
• list these in the results dataframe
• Also add numSamps to results

Fix bug on constraints=None

constraints is being set to None, in spite of passing in constraints as a parameter.

Include provenance information

It would be nice to include information about the exact instance Catalogs queried and from where and serialize that information. Use this thread to include desired properties.

• user machine information
• Date and times of queries
• Constraints used on the catalog
• CatalogDBObject type
• Host database

Fix axes sharing in plots.

Add Chunksize parameter to use catalog generation branch update

Write out all output in a directory, fail on start if directory exists

• Copy run script to directory