Skip to content

One name (verdi-agents), tidy repo, plain-language docs; 0.3.0 - #18

Merged
SushantGautam merged 4 commits into
mainfrom
chore/unify-name-verdi-agents
Nov 4, 2026
Merged

SushantGautam merged 4 commits into
mainfrom
chore/unify-name-verdi-agents

Conversation

@SushantGautam

Copy link
Copy Markdown
Collaborator

Summary

Makes Verdi read like a small, understandable library for ALIFE researchers.

One name everywhere

  • PyPI verdi-agents, command verdi-agents, Python import verdi_agents (hyphens are not valid in imports). The old verdi command and import verdi are gone.
  • New subcommands verdi-agents pair and verdi-agents compare, so every documented workflow is one verdi-agents <command>.

Tidier repository

  • Removed internal working notes (AUDIT.md, findings.md, docs/concordia_eval.md) and a 4k-line generated example report.
  • Top-level experiments/ (one recorded test scenario) moved to tests/fixtures/reference_run/.
  • Modules named after internal milestone codes renamed: pilot → reference_run, g5_pairs → counterfactual_pairs, g5_close → pair_audit, g5_judge_live → pair_judge, interview_experiment → interviews, judge_live → judge, live_bridge → trace_audit, assembled → assembly; regime files and test files renamed to match.
  • Docstrings and comments no longer use internal codes (ADR-0xx, G4/G5, P0–P2, WS-n, Lane X, L1/L2, Society V2) or point at documents that do not exist. Only text changed in those files (AST with strings masked is identical).
  • Machine-specific absolute paths removed from code and tests; example scenarios point at http://localhost:8000/v1 instead of a personal endpoint.

Plain-language docs

  • README rewritten: what Verdi does, a first run, commands, a glossary of terms, what each output file is, and what Verdi does not claim.
  • Guides and example headers use verdi-agents commands and plain wording.

Unchanged on purpose: data identifiers stored in run files and recorded inputs (verdi.social/v1.0, pilot_v1, regime labels G0–G4, prompt versions), and VERDI_* environment variables, so existing runs and recorded inputs still load.

Version bumped to 0.3.0 (breaking: package and command renamed).

Test plan

  • ruff check src tests clean
  • pytest: full suite passes (1 skipped, as before)
  • Built wheel + sdist; fresh venv install: verdi-agents help, verdi-agents init, import verdi_agents → 0.3.0
  • sdist contains no AUDIT/findings/notes files

🤖 Generated with Claude Code

Install, command and import now share one name:
- PyPI: verdi-agents
- CLI: verdi-agents (the old 'verdi' command is removed)
- Python: import verdi_agents (hyphens are not valid in imports)

Data identifiers such as 'verdi.social/v1.0' are unchanged so existing
run artifacts and recorded tapes still load.
- Remove internal working notes (AUDIT.md, findings.md,
  docs/concordia_eval.md) and a generated example report.
- Move the recorded reference scenario from the top-level experiments/
  folder into tests/fixtures/reference_run/.
- Rename modules named after internal milestone codes:
  pilot -> reference_run, g5_pairs -> counterfactual_pairs,
  g5_close -> pair_audit, g5_judge_live -> pair_judge,
  interview_experiment -> interviews, judge_live -> judge,
  live_bridge -> trace_audit, assembled -> assembly.
- Rename regime files and test files to say what they cover.
Docstrings, comments and messages no longer refer to internal milestone
codes (ADR-0xx, G4/G5, P0-P2, WS-n, Lane X, L1/L2, Society V2) or to
documents that are not in the repository. Only text changed: an AST
comparison with string constants masked is identical for every file
except the path fixes below.

Also remove machine-specific absolute paths: the default origin run for
counterfactual pairs is now inside the checkout and can be overridden
with VERDI_ORIGIN_RUN_DIR; tests locate optional recorded runs relative
to the repository.
- New subcommands: 'verdi-agents pair' (fixed vs changeable rules) and
  'verdi-agents compare' (measure a paired run). Both accept the input
  path positionally.
- README rewritten for researchers new to the tool: what Verdi does, a
  first run, the commands, a glossary, and what each output file is.
- Guides and example headers use verdi-agents commands instead of
  python -m module paths, and no longer use internal terms.
- Example scenarios point at a localhost model server instead of a
  personal endpoint.
- Bump version to 0.3.0.
Copilot AI balanced review requested due to automatic review settings November 4, 2026 10:01

Copilot AI left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Copilot was unable to review this pull request because the user who requested the review has insufficient remaining quota. Try requesting a review at a lower effort level.

An organization owner can choose the repository organization as the billing source for reviews requested by members with a paid Copilot license. Learn about Copilot code review billing.

@SushantGautam
SushantGautam merged commit 3093813 into main Nov 4, 2026
1 check passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants