Small bug fixes

[yarikoptic]

• [x] now seems to have to use incremental=True while aggregating metadata for the subdataset all the way to the top. @mih please verify that this is expected (was not needed before). Without it it would drop metadata for any other present subdataset in the dataset's superdatasets
• [x] use legacy. of openfmri for crawling (although nothing new is expected there)
• [x] fix metadata fetching for CRCNS (Fixes #15 )

[codecov-io]

Codecov Report

Merging #10 into master will decrease coverage by 0.02%. The diff coverage is 90%.

@@            Coverage Diff             @@
##           master      #10      +/-   ##
==========================================
- Coverage   86.76%   86.74%   -0.03%     
==========================================
  Files          50       50              
  Lines        4111     4112       +1     
==========================================
  Hits         3567     3567              
- Misses        544      545       +1

Impacted Files	Coverage Δ
datalad_crawler/pipelines/openfmri.py	85% <100%> (ø)	:arrow_up:
datalad_crawler/nodes/annex.py	81.66% <100%> (ø)	:arrow_up:
datalad_crawler/pipelines/tests/test_openfmri.py	91.38% <100%> (ø)	:arrow_up:
datalad_crawler/pipelines/crcns.py	73.11% <75%> (-0.8%)	:arrow_down:

---

Continue to review full report at Codecov.

Legend - Click here to learn more Δ = absolute <relative> (impact), ø = not affected, ? = missing data Powered by Codecov. Last update d5f34fa...3f30bfd. Read the comment docs.

[yarikoptic]

will be merged if tests pass - need other fixes on top.

[yarikoptic]

ha ha -- I thought to submit crcns fix in a different PR but apparently has to come here, so here it comes!

[yarikoptic]

Foi I had to restart python 3.6 run because of that issue with attr and converter being undefined. It is appearing from time to time while importing scrapy - may be there is some late bindings and race condition