Skip to content

Conversation

rtobar
Copy link

@rtobar rtobar commented May 20, 2020

Various improvements to the benchmarking code. Please check each commit message for further explanation and better breakup of the changes.

rtobar added 3 commits May 20, 2020 14:33
The table showing the results for the different facilities had a column
named "Type", which in the case of the jsonslicer tests corresponded to
the text type of the produced output. To clarify this I've changed the
name to "Output type", and also removed the information redundancy in
the jsonslicer test case names, where the output type was mentioned too.

I have also corrected the output type for the ijson tests, as ijson
always outputs str.

Signed-off-by: Rodrigo Tobar <rtobar@icrar.org>
This change brings all the assertion code together into a common
function that all test cases use. This way it's easier to add new
assertions.

Signed-off-by: Rodrigo Tobar <rtobar@icrar.org>
The latter always take paths as str, not as bytes, and no internal check
is done (yet) to double-check this. As a result the comparisons with
ijson were bogus, as ijson.items was yielding no results (with python
3.8.2; in python 2.7 they would have yielded the correct results, but
jsonslicer doesn't support 2.7).

To avoid bogus comparisons in the future a new assertion in the common
assertion code now verifies that the expected number of objects were
yielded by the different methods.

Signed-off-by: Rodrigo Tobar <rtobar@icrar.org>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant