Fix for #1334: Refactor transcript annotation

[andrewjesaitis]

• Places extra data in attributes field
• Simplify code path for annotation parsing

[david4096]

Removes code and adds tests. 💯

[david4096]

I aspire to have these tuples as part of the registry editor.

It should be possible to easily write a VCF->ANN plugin that accepts a tuple or dictionary that can effectively parse a ANN field.

[andrewjesaitis]

Yeah, I thought the same thing since we are mixing data with code here. But, we'll save it for an annotation plugin refactor.

[david4096]

These could probably go into protocol in schemas. I think this pattern of dot access for dictionaries is a sane way to work with data when you don't want to make a new class in python, and mocks the protocol buffers and peewee models well.

This is what I did for the peer service

class Struct:
    def __init__(self, **entries):
        self.__dict__.update(entries)

###

map(lambda x: Struct(**x), self._peers[offset:offset + limit])

Maybe you could do this recursively?

https://github.com/ga4gh/server/pull/1556/files#diff-02d20abc7e4d68c8ae177064ca45176aR49

[andrewjesaitis]

Cool. I'm moving these over to schemas then.

Here's the recursive version:

def deepGetAttr(obj, path):
    first, sep, rest = path.partition(".")
    obj = getattr(obj, first)
    if rest:
        obj = deepGetAttr(obj, rest)
    return obj

Seems like the recursive version is a little bit uglier. Any specific benefit you were thinking of?

[david4096]

Sorry, failed to communicate myself properly how you've approached it is fine. Sending in dot notation as a string scares me a little bit :).

I was thinking something like this, but don't let it hold up this PR.

[andrewjesaitis]

Also, how does it work when a PR in server depends on a PR in schemas? Should I commit my constraints.txt file that points to my version of schemas that has he edited protocol.py?

[dcolligan]

Here's how I would advise doing it:

• push a branch to ga4gh/schemas and open a PR on it
• open a PR with your contstraints.txt file line for schemas pointing to the remote schemas branch
• presumably both the schemas and the server changes are going in, and the server depends on schemas, so merge the schemas PR first, once it is approved
• change your server PR to use the default version of constraints.txt (that is, the one that points to all the master branches)
• merge the server PR

[david4096]

Hey we'll merge this next @andrewjesaitis !

[codecov-io]

Codecov Report

Merging #1562 into master will decrease coverage by -0.02%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##           master    #1562      +/-   ##
==========================================
- Coverage   85.61%   85.59%   -0.02%     
==========================================
  Files          33       33              
  Lines        7188     7152      -36     
  Branches      898      894       -4     
==========================================
- Hits         6154     6122      -32     
+ Misses        851      848       -3     
+ Partials      183      182       -1

Impacted Files	Coverage Δ
ga4gh/server/datamodel/variants.py	94.27% <100%> (+0.26%)	✅

---

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 cdfd148...5792e83. Read the comment docs.