← All workshopsRead Where Did This Maintenance Claim Come From?

Graph Databases

Select the evidence available for this decision

The decision reaches the exclusive expiry hour

Two permitted sources disagree about P7. Botie needs to retain their claim identities while excluding expired, future or unapproved evidence.

BOTIE

C1 is available from hour 2 until hour 8, with the end excluded. C2 is available from hour 4 with no stated end. Both sources are permitted. At hour 8, which remains?
The inputs we are comparing

Claim and provenance records

Same in both cases

[
  {
    "id": "C1",
    "subject": "P7",
    "statement": "inspection overdue",
    "source": "manual-A",
    "observed_at": 2,
    "valid_until": 8
  },
  {
    "id": "C2",
    "subject": "P7",
    "statement": "inspection current",
    "source": "log-B",
    "observed_at": 4,
    "valid_until": null
  },
  {
    "id": "C3",
    "subject": "P7",
    "statement": "inspection current",
    "source": "manual-A",
    "observed_at": 9,
    "valid_until": null
  },
  {
    "id": "C4",
    "subject": "P7",
    "statement": "inspection overdue",
    "source": "unknown",
    "observed_at": 1,
    "valid_until": null
  }
]

Permitted source IDs

Same in both cases

  1. 0manual-A
  2. 1log-B

Decision hour

Two current permitted sources disagree

6

The older claim reaches its exclusive end

8
Result for two current permitted sources disagree
  1. 0C1
  2. 1C2

For a list use brackets, such as [2, 4]. For a matrix put each row in brackets. Text answers can stay as text.

One claim starts as another expires

This check uses the pieces from “Apply the decision-time interval”.

BOTIE

At hour twelve, use sources log-C and inspection-D. Q1 expires exactly at twelve; Q2 becomes available then; Q3 is already available and disagrees with Q2. Q4 is from another source, and Q5 arrives at thirteen. Predict the eligible claim IDs while retaining any disagreement.
Claim and provenance records
[
  {
    "id": "Q1",
    "subject": "R8",
    "statement": "needs inspection",
    "source": "log-C",
    "observed_at": 7,
    "valid_until": 12
  },
  {
    "id": "Q2",
    "subject": "R8",
    "statement": "inspection current",
    "source": "inspection-D",
    "observed_at": 12,
    "valid_until": null
  },
  {
    "id": "Q3",
    "subject": "R8",
    "statement": "needs inspection",
    "source": "log-C",
    "observed_at": 11,
    "valid_until": 15
  },
  {
    "id": "Q4",
    "subject": "R8",
    "statement": "inspection current",
    "source": "unreviewed-E",
    "observed_at": 9,
    "valid_until": null
  },
  {
    "id": "Q5",
    "subject": "R8",
    "statement": "inspection current",
    "source": "log-C",
    "observed_at": 13,
    "valid_until": null
  }
]
Permitted source IDs
  1. 0log-C
  2. 1inspection-D
Decision hour
12

Use a number, text, or an array matching the requested result. For a matrix, put each row in brackets. Saving your prediction runs your current construction on these inputs and checks the earlier examples too.

Build a construction in the workspace above first. This case will test those pieces.

What information would help check a model-extracted claim against its actual source passage?

Retain enough provenance to investigate and correct a result, while preserving uncertainty.

These components operate on the supplied records in your browser. They do not run a graph database or interpret Cypher.

Work through the Python challenges →
Continue the courseWhich Connected Records Can the Assistant Use? →
← Can Two Teams Describe the Same Resource?