biocommons / hgvs-eval Goto Github PK

Following the current README, I got various errors when running make devready, the first being
a problem with install pysam.

I ended up manually running:
pip install -r
on etc/develop.reqs and etc/install.reqs
and then make devready

after that I could run
python app.py &
successfully.

Has anyone else had success with the README in a fresh virtualenv

While previously it worked for me, on both Master and nicole_wor1 branches
currently when I run the setup instructions in the README'and run
python app.y &

I get the below

(venv) paschallj@jplap:~/hgvs-eval7/hgvs-eval$ python app.py 
Traceback (most recent call last):
  File "app.py", line 9, in <module>
    from hgvseval.testservice.biocommonsService import BiocommonsService
  File "/home/paschallj/hgvs-eval7/hgvs-eval/hgvseval/testservice/biocommonsService.py", line 10, in <module>
    class BiocommonsService(HGVSTestService):
  File "/home/paschallj/hgvs-eval7/hgvs-eval/hgvseval/testservice/biocommonsService.py", line 12, in BiocommonsService
    hdp = hgvs.dataproviders.uta.connect()
  File "/home/paschallj/hgvs-eval7/hgvs-eval/venv/local/lib/python2.7/site-packages/hgvs/dataproviders/uta.py", line 108, in connect
    conn = UTA_postgresql(url=url, pooling=pooling, application_name=application_name)
  File "/home/paschallj/hgvs-eval7/hgvs-eval/venv/local/lib/python2.7/site-packages/hgvs/dataproviders/uta.py", line 473, in __init__
    super(UTA_postgresql, self).__init__(url)
  File "/home/paschallj/hgvs-eval7/hgvs-eval/venv/local/lib/python2.7/site-packages/hgvs/dataproviders/uta.py", line 183, in __init__
    self._connect()
  File "/home/paschallj/hgvs-eval7/hgvs-eval/venv/local/lib/python2.7/site-packages/hgvs/dataproviders/uta.py", line 488, in _connect
    self._conn = psycopg2.connect(**conn_args)
  File "/home/paschallj/hgvs-eval7/hgvs-eval/venv/local/lib/python2.7/site-packages/psycopg2/__init__.py", line 130, in connect
    conn = _connect(dsn, connection_factory=connection_factory, **kwasync)
psycopg2.OperationalError: invalid port number: "None"

Any ideas @reece or @nicoleruiz ?

Summary

HGVS Nomenclature is the lingua franca for presenting sequence variants to humans in databases, scientific literature, and clinical reports. Unfortunately, some implementations of the standard are buggy and/or incomplete. The purpose of this project is to develop an framework for evaluating implementations of HGVS Nomenclature to help users understand the limitations of packages, to help package developers improve their packages, and ultimately to increase the fidelity of data sharing. As a byproduct of the proposed plan, an OpenAPI specification would be developed and, if adopted, enable easier migration between tools.

Community Benefits

The primary beneficiaries of this work would be the users of specific implementations who would be able to make well-informed decisions about gaps or issues. Since these users often work with clinicians, scientists, or adjacent professionals, it is expected that they would develop greater confidence that HGVS variant descriptions and manipulations were being performed consistently.

Required and Desired Skills

• Python: very good-excellent
• OpenAPI: some experience is ideal

Expected Results / Deliverables

• An OpenAPI specification for HGVS Nomenclature variant descriptions and manipulation
• A set of classified tests that exercise all aspects of HGVS Nomenclature
• A framework for executing/evaluating packages, storing and publishing results, and comparing within and across implementations
• Demonstration of the framework using at least one publicly available implementation (ideally more)
• Package the project so that it can be used in CI/CD testing and by a public evaluation server
• Develop a full-stack app to store and display evaluation results.

Benefits to Intern

This project is well-suited to individuals who have experience with REST interface design and testing methodology. Implementation will be in Python, and applicants should have strong Python skills. Exposure to HGVS and genetic variants is desirable but not required. This project has enormous opportunities to learn and to have significant impact in the scientific community. At a minimum, the intern would:

• become familiar with HGVS Nomenclature and some of the challenges when writing standards
• learn how to design a REST interface and use it to create a façade over different software packages
• develop software architecture skills

This project could be simple, but also offers many avenues for extensions and learning. For example, the student might seek to build a web interface with authentication to store results.

How to apply

Students applying to this project should briefly describe their vision for this project, highlight their existing skills and the skills they would need to learn, and estimate an implementation timeline.

investigate protobuf -> swagger documentation

I noticed that setup.py may not be up to date. Maybe @reece can update it? Not sure if there is some kind of biocommons package naming scheme you want to go with.

Some services (e.g. Counsyl and SolveBio) lack a function that maps a genomic HGVS to a specific transcript passed in as a parameter, but rather return a list of all transcripts which a genomic position maps to.

There are two ways to deal with this in context of a test defined for a g. to c. expecting a specific transcript:

1. Require that in implementing the REST API all services implement such an accession parameterized g to c. function

2. Allow "set membership" matches where a test will succeed if any of a list of returned HGVS expressions matches the expected output.