Skip to content

Define trajectory JSON schema v0 #2

Description

@rmems

Goal

Define the first JSON schema for software-engineering trajectory records.

Context

The core data unit should describe one useful engineering trajectory, such as:

issue/review signal → before state → patch/fix → validation → outcome

Tasks

  • Add schemas/pr_trajectory.schema.json.
  • Add fields for repository, PR number, source URL, language, domain, task type, review signal, before context, after patch, validation events, outcome, and training use.
  • Add enum-like conventions for task_type, outcome, and training_use.
  • Add a tiny example record under datasets/examples/.

Proposed top-level fields

  • id
  • repo
  • pr_number
  • source_urls
  • language
  • domain
  • task_type
  • issue_context
  • review_signals
  • before_context
  • patch
  • validation
  • outcome
  • training_use
  • quality_score

Acceptance Criteria

  • Schema validates the example JSON file.
  • Example is small enough to commit safely.
  • README documents this as schema v0, not final.

Metadata

Metadata

Assignees

Labels

No labels
No labels

Projects

No projects

Relationships

None yet

Development

No branches or pull requests

Issue actions