NER

[hrshdhgd]

• Added runner into requirements.txt
• Dynamically creates settings.ini based on provided information (path, ontology etc.)
• Tests added

Addresses #3

[cmungall]

add a Makefile step for the code maintainer to update the envo termlist

[cmungall]

see how the mixs json is loaded

The general pattern is

• allow code user to specify their own paths
• get the right defaults with zero configuration

[hrshdhgd]

see how the mixs json is loaded

Could you please guide me to a sample mixs.json file?

[cmungall]

sample-annotator/sample_annotator/__init__.py

Lines 3 to 5 in bc17c12

	MAIN_MODEL_DIR = os.path.join(os.path.abspath(os.path.dirname(__file__)), 'model')
	MAIN_SCHEMA_DIR = os.path.join(MAIN_MODEL_DIR, 'schema')
	MIXS_SCHEMA = os.path.join(MAIN_SCHEMA_DIR, 'mixs.json')

[hrshdhgd]

add a Makefile step for the code maintainer to update the envo termlist

So if I get this correct, you want me to add a line in the Makefile to generate envo_termlist.tsv ?

[cmungall]

So if I get this correct, you want me to add a line in the Makefile to generate envo_termlist.tsv ?

Yes

The guiding principle for all repos is that derived files must have provenance and have transparent portable means to recreate

[cmungall]

can you also add a text. See the geo example for guidance.

Note that after this PR is merged we will integrate the functionality here:

sample-annotator/sample_annotator/sample_annotator.py

Line 206 in bc17c12

def perform_text_mining(self, sample: SAMPLE, report: AnnotationReport):

and add a test to test_sample_info

 - description: easy to pass TM text
    sample:
      id: TEST:2
      env_broad_scale: terrestrial biome
      env_local_scale: ...
    must_pass: false
    output:
       env_broad_scale: ENVO:nnn....
       env_local_scale: ...
    expected_failures: {}

and another harder one:

 - description: slightly harder TM test
    sample:
      id: TEST:2
      description: my soil sample
    must_pass: false
    output:
       env_medium_scale: ENVO:<ID FOR SOIL>
    expected_failures: {}

[hrshdhgd]

made a few edits, thoughts @cmungall ?

[cmungall]

need to await runner on pypi

[cmungall]

1. We need a pypi release (for which I need to decide on name)
2. if Pipfile is SOT then this needs to go in Pipfile