args-me-model

The data model of the args.me search engine (work-in-progress).

Claims are extracted from or generated by one or more Sources and connected via Support relations. Two claims can reference each other as their counter if they state the opposite of each other. Each of Claim, Source, and Support have additional fields to store information and allow for arbitrary annotations.

Install

pip install args-me-model

Quickstart

from args_me_model import Claim, Source

# Creating claims
premise1 = Claim.from_source(
        Source(name="common-knowledge", text="There are no clouds")
    )
premise2 = Claim.from_source(
        Source(name="common-knowledge", text="A clear sky is blue")
    )
claim = Claim.from_source(
        Source(name="common-knowledge", text="The sky is blue"),
        support = [[premise1, premise2]] # linked support relations
    )
counterclaim = Claim.from_source(
        Source(name="uncommon-knowledge", text="The sky is not blue"),
        counter = claim # also sets claim.counter
    )

# Writing claims to a file
Claim.write_ndjson(
        [premise1, premise2, claim, counterclaim],
        "claims.ndjson"
    )

# Iterating over claims from a file
for claim in Claim.read_ndjson("claims.ndjson"):
    print(claim.text)

Contributing

See the development notes.