[WIP] added push_datapackage and pull_datapackage

[roll]

Work in progress will wait a discussion and https://github.com/frictionlessdata/testsuite-py tests added and passed.

[coveralls]

Coverage decreased (-12.0%) to 86.754% when pulling 63cce7a on feature/push-pull-datapackage into cf80599 on master.

[coveralls]

Coverage decreased (-11.6%) to 87.116% when pulling 1d8e7e5 on feature/push-pull-datapackage into cf80599 on master.

[coveralls]

Coverage decreased (-11.7%) to 87.007% when pulling 104b3ac on feature/push-pull-datapackage into cf80599 on master.

[coveralls]

Coverage decreased (-12.1%) to 86.628% when pulling 63c152a on feature/push-pull-datapackage into cf80599 on master.

[coveralls]

Coverage decreased (-12.5%) to 86.252% when pulling 8889a3d on feature/push-pull-datapackage into cf80599 on master.

[coveralls]

Coverage decreased (-12.5%) to 86.252% when pulling 50d19e3 on feature/push-pull-datapackage into cf80599 on master.

[coveralls]

Coverage decreased (-8.7%) to 90.029% when pulling 6fadf35 on feature/push-pull-datapackage into cf80599 on master.

[coveralls]

Coverage decreased (-0.5%) to 98.266% when pulling 004235d on feature/push-pull-datapackage into cf80599 on master.

[coveralls]

Coverage increased (+0.3%) to 98.988% when pulling 2516f10 on feature/push-pull-datapackage into cf80599 on master.

[roll]

@pwalsh @vitorbaptista
Please take a look.

It works with 9 datasets (push/pull to SQL/BigQuery) from testsuite-py:
https://travis-ci.org/frictionlessdata/testsuite-py/jobs/119227783

[pwalsh]

@roll 🎱

[pwalsh]

@roll also, really cool using behave for the integration tests!

[pwalsh]

@roll please get final signoff on this from @vitorbaptista

[vitorbaptista]

The quote is in the wrong place

[vitorbaptista]

@roll The code looks good, it just needs a few tweaks to be more consistent with the overall style of this package. For instance, instead of having pull_datapackage and push_datapackage, we'd have DataPackage.pull() as a class method which returns a DataPackage instance, and datapackage.push() as an instance method.

To be honest, I prefer DataPackage.from_storage() and datapackage.to_storage(), but that might be because "push" is a tricky word for me. It sounds the same as the Portuguese word "puxe", which means "pull". It takes me a second to translate, which makes pulling doors with "push" signs in them 😅

Push/pull are very common for devs, though, so I'm happy to keep it.

[roll]

@vitorbaptista
It's following of our prev discussion - frictionlessdata/tableschema-py#51 (comment)

I think we've decided for now to have temp but working tooling:

- push_resource (filesystem -> db)
- pull_resource (db -> filesystem)
- push_datapackage (filesystem -> db)
- pull_datapackage (db -> filesystem)

So I've intentionally done it without touching DataPackage API and OOP-approach. I think we could integrate it with the main API after jsontableschm.Resource will be designed (we need to see how it will be looking) and then deprecate API from pushpull.py.

cc @pwalsh

[coveralls]

Coverage increased (+0.3%) to 98.988% when pulling 27a12c5 on feature/push-pull-datapackage into cf80599 on master.

[vitorbaptista]

Is this a common pattern? This feels like it should belong in the README instead, maybe in a "Running the tests" subsection inside "Developer notes".

[roll]

Ohh my bad system .gitignore was broken

[vitorbaptista]

You're right. Just check a few typos on the README (like exmples) and feel free to merge 👍

[coveralls]

Coverage increased (+0.3%) to 98.988% when pulling b17a2ea on feature/push-pull-datapackage into cf80599 on master.

[coveralls]

Coverage increased (+0.3%) to 98.988% when pulling 5f198e9 on feature/push-pull-datapackage into cf80599 on master.

[roll]

Great!

• typos fixed
• tested against testsuite-py again
• merged