Use data registry to access paths and save pipeline runs

[joezuntz]

This adds an interface to the data registry. We want to:

• Get paths to overall_input in yaml files using a direct lookup of a file by ID or name.
• Support looking up a file from a specific user or project.
• Support lookup by alias when ready.
• Support getting an input using its tag and a previous execution lookup.
• Add a way to save an execution entry from an entire pipeline run and its outputs.

[codecov]

Codecov Report

Attention: 4 lines in your changes are missing coverage. Please review.

Comparison is base (25def53) 89.81% compared to head (a7f921d) 89.66%.

Additional details and impacted files

@@            Coverage Diff             @@
##           master      #93      +/-   ##
==========================================
- Coverage   89.81%   89.66%   -0.16%     
==========================================
  Files          16       16              
  Lines        1709     1722      +13     
==========================================
+ Hits         1535     1544       +9     
- Misses        174      178       +4     

Files	Coverage Δ
ceci/pipeline.py	84.88% <73.33%> (-0.35%)	⬇️

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[stuartmcalpine]

I have updated the registry setup to the latest dataregistry interface syntax.

One technicality is during data_registry_lookup you can search for a dataset by name, and check it is unique. names dont have to be unique in the registry, name + version does, so we can come back to that