fix(generation): prioritize oldest stale pages when the budget is constrained by sloemo01 · Pull Request #1845 · repowise-dev/repowise

sloemo01 · 2026-08-22T14:14:35Z

Summary

Fixes #851 (spun out of #847).

When the cascade budget is constrained, the ordering that decides which pages to regenerate was driven purely by pagerank (importance). That meant a docs run could spend its LLM calls regenerating high-importance pages that were already current while leaving genuinely stale prose untouched — the exact staleness-vs-importance mis-prioritization reported.

This change makes staleness the primary sort axis for the budgeted slice:

ChangeDetector.get_affected_pages now accepts a stale_pages map ({file_path: staleness_age_seconds}). When provided, an already-stale page outranks a fresh page no matter how central the fresh one is; among stale pages, the oldest stale page regenerates first. Importance (pagerank) breaks ties within each staleness class.
Absent/empty stale_pages keeps the historical pure-importance ordering, so no call site changes behaviour unless it opts in.

Plumbing

New get_stale_file_page_ages(session, repo_id) persistence CRUD helper: returns {file_path: age_seconds} for stale/expired file_page rows, age measured from updated_at (the last regeneration) so the longest-lagging page carries the largest value.
CLI repowise update --docs loads stale ages and passes them into the cascade.
Server incremental page regen (_incremental_page_regen) loads stale ages through its session/repo and passes them in (best-effort; missing session/repo falls back to the old ordering).
Exported via repowise.core.persistence.

Tests

tests/unit/ingestion/test_change_detector.py: a stale page regenerates ahead of a more important fresh page; oldest-stale-first within a budget; absent stale_pages preserves importance ordering.
tests/unit/persistence/test_stale_file_page_ages.py: which rows count, age semantics, and empty-return contract.

Verification

uv run ruff check — clean on all changed files.
uv run pytest over the changed modules — all pass (the one test_repo_totals_churn failure is a pre-existing git-environment issue, reproduced with the changes stashed).

…strained The cascade budget in ChangeDetector.get_affected_pages sliced its regenerate set by pagerank alone, so a constrained docs run could regenerate pages that were already current while leaving genuinely stale prose untouched (issues repowise-dev#847 / repowise-dev#851). Thread existing stale-page ages into the sort so an already-stale page outranks a fresh page no matter how central the fresh one is; among stale pages the oldest decays first. Adds get_stale_file_page_ages to the persistence CRUD, wires it into the CLI update path and the server's incremental page regen, and pins the behaviour with regression tests.

repowise-bot · 2026-08-22T14:15:13Z

✅ Health of changed files: 2.6 → 2.9 (+0.3)
_{⚠️ Change risk: moderate, riskier than 58% of this repo's commits.}

📋 At a glance
1 file changed health · 5 hotspots touched · 4 new findings introduced · 3 co-change pairs left out · 5 files with recent fix history.

Files & modules (2)

packages (5 files)
- .../update_cmd/command.py
- .../server/job_executor.py
- .../persistence/__init__.py
- .../update_cmd/deterministic.py
- .../crud/pages.py
tests (1 file)
- .../ingestion/test_change_detector.py

✅ Health gate: passed

📌 Before you merge

Run tests/integration/test_cli.py, .../cli/test_docs_pointer.py, .../cli/test_editor_setup.py, .../cli/test_deterministic_update.py (+9 more): they import the changed files
.../update_cmd/persistence.py changed together with .../update_cmd/command.py in 14 past commits and isn't in this PR
.../workspace/update.py changed together with .../update_cmd/command.py in 13 past commits and isn't in this PR
.../init_cmd/command.py changed together with .../update_cmd/command.py in 10 past commits and isn't in this PR

🎯 Blast radius _{(symbols whose signature this PR changed, and who calls them)}

_incremental_page_regen in .../server/job_executor.py signature changed. Called by 2 symbols outside this PR: .../server/test_job_executor.py::test_incremental_page_regen_passes_repo_path, .../server/test_job_executor.py::test_incremental_page_regen_sets_file_pages_only

🔎 More signals (3)

🗺️ Change map

flowchart LR
  subgraph PR ["Changed in this PR (6 with dependents)"]
    f_packages_cli_src_repowise_cli_commands_update_cmd_command_py[".../update_cmd/command.py 🔥"]:::changed
    f_packages_cli_src_repowise_cli_commands_update_cmd_deterministic_py[".../update_cmd/deterministic.py 🔥"]:::changed
    f_packages_core_src_repowise_core_persistence___init___py[".../persistence/__init__.py 🔥"]:::changed
    f_packages_server_src_repowise_server_job_executor_py[".../server/job_executor.py 🔥"]:::changed
    f_packages_core_src_repowise_core_ingestion_change_detector_py[".../ingestion/change_detector.py"]:::changed
    f_packages_core_src_repowise_core_persistence_crud_pages_py[".../crud/pages.py"]:::changed
  end
  f_packages_cli_src_repowise_cli_commands_generate_cmd_command_py[".../generate_cmd/command.py"]
  f_packages_cli_src_repowise_cli_commands_update_cmd_command_py --> f_packages_cli_src_repowise_cli_commands_generate_cmd_command_py
  f_packages_cli_src_repowise_cli_commands_init_cmd_command_py[".../init_cmd/command.py"]
  f_packages_cli_src_repowise_cli_commands_update_cmd_command_py --> f_packages_cli_src_repowise_cli_commands_init_cmd_command_py
  f_packages_cli_src_repowise_cli_commands_init_cmd_generation_py[".../init_cmd/generation.py"]
  f_packages_cli_src_repowise_cli_commands_update_cmd_command_py --> f_packages_cli_src_repowise_cli_commands_init_cmd_generation_py
  f_packages_cli_src_repowise_cli_commands_init_cmd_persistence_py[".../init_cmd/persistence.py"]
  f_packages_cli_src_repowise_cli_commands_update_cmd_command_py --> f_packages_cli_src_repowise_cli_commands_init_cmd_persistence_py
  f_packages_cli_src_repowise_cli_commands_update_cmd_deterministic_py --> f_packages_cli_src_repowise_cli_commands_init_cmd_command_py
  f_packages_cli_src_repowise_cli_commands_restyle_cmd_py[".../commands/restyle_cmd.py"]
  f_packages_cli_src_repowise_cli_commands_update_cmd_deterministic_py --> f_packages_cli_src_repowise_cli_commands_restyle_cmd_py
  f_packages_cli_src_repowise_cli__repo_session_py[".../cli/_repo_session.py"]
  f_packages_core_src_repowise_core_persistence___init___py --> f_packages_cli_src_repowise_cli__repo_session_py
  f_packages_cli_src_repowise_cli_commands_augment_cmd_search_py[".../augment_cmd/search.py"]
  f_packages_core_src_repowise_core_persistence___init___py --> f_packages_cli_src_repowise_cli_commands_augment_cmd_search_py
  f_packages_cli_src_repowise_cli_commands_claude_md_cmd_py[".../commands/claude_md_cmd.py"]
  f_packages_core_src_repowise_core_persistence___init___py --> f_packages_cli_src_repowise_cli_commands_claude_md_cmd_py
  f_packages_cli_src_repowise_cli_commands_costs_cmd_py[".../commands/costs_cmd.py"]
  f_packages_core_src_repowise_core_persistence___init___py --> f_packages_cli_src_repowise_cli_commands_costs_cmd_py
  f_packages_server_src_repowise_server_job_executor_py --> f_packages_cli_src_repowise_cli_commands_init_cmd_command_py
  f_packages_server_src_repowise_server_job_executor_py --> f_packages_cli_src_repowise_cli_commands_init_cmd_persistence_py
  f_packages_cli_src_repowise_cli_commands_init_cmd_workspace_py[".../init_cmd/workspace.py"]
  f_packages_server_src_repowise_server_job_executor_py --> f_packages_cli_src_repowise_cli_commands_init_cmd_workspace_py
  f_packages_server_src_repowise_server_job_executor_py --> f_packages_cli_src_repowise_cli_commands_restyle_cmd_py
  more(["+89 more dependents"])
  PR --> more
  w_packages_cli_src_repowise_cli_commands_update_cmd_persistence_py(["⚠️ .../update_cmd/persistence.py changed together 14×, not in PR"]):::warn
  f_packages_cli_src_repowise_cli_commands_update_cmd_command_py -.- w_packages_cli_src_repowise_cli_commands_update_cmd_persistence_py
  w_packages_core_src_repowise_core_workspace_update_py(["⚠️ .../workspace/update.py changed together 13×, not in PR"]):::warn
  f_packages_cli_src_repowise_cli_commands_update_cmd_command_py -.- w_packages_core_src_repowise_core_workspace_update_py
  t_tests_integration_test_cli_py(["✅ tests/integration/test_cli.py"]):::guard
  t_tests_integration_test_cli_py -.-> f_packages_cli_src_repowise_cli_commands_update_cmd_command_py
  t_tests_unit_cli_test_deterministic_update_py(["✅ .../cli/test_deterministic_update.py"]):::guard
  t_tests_unit_cli_test_deterministic_update_py -.-> f_packages_cli_src_repowise_cli_commands_update_cmd_deterministic_py
  t_tests_integration_persistence_test_graph_metrics_pg_py(["✅ .../persistence/test_graph_metrics_pg.py"]):::guard
  t_tests_integration_persistence_test_graph_metrics_pg_py -.-> f_packages_core_src_repowise_core_persistence___init___py
  classDef changed fill:#dbeafe,stroke:#1d4ed8,color:#1e3a5f
  classDef warn fill:#fef3c7,stroke:#b45309,color:#78350f
  classDef guard fill:#dcfce7,stroke:#15803d,color:#14532d

_{Solid arrows: code that imports the changed files (99 direct dependents, from the last indexed snapshot). Dashed: history/tests.}

🔥 Hotspots touched (5)

.../update_cmd/command.py: 54 commits/90d, 19 dependents · primary owner: Raghav Chamadiya (87%)
.../ingestion/test_change_detector.py: 1 commits/90d, 0 dependents · primary owner: Raghav Chamadiya (100%)
.../server/job_executor.py: 20 commits/90d, 20 dependents · primary owner: Raghav Chamadiya (74%)

2 more

.../persistence/__init__.py: 7 commits/90d, 161 dependents · primary owner: Raghav Chamadiya (90%)
.../update_cmd/deterministic.py: 10 commits/90d, 5 dependents · primary owner: Raghav Chamadiya (84%)

🔗 Hidden coupling (1 file)

.../update_cmd/command.py co-changes with these files (not in this PR):
- .../update_cmd/persistence.py (14×, 🟡 notable)
- .../workspace/update.py (13×, 🟡 notable)
- .../init_cmd/command.py (10×, 🟡 notable)

👀 Suggested reviewers @RaghavChamadiya

📊 See the full report for this PR
_{Your repo map with this PR's blast radius lit up, every caller of the contracts it changes, and health before and after. No sign-in. · ⭐ Star Repowise · 📥 Install bot · Silence on a single PR with [skip repowise] in the title · Per-repo toggle on repowise.dev/settings?tab=bot · Updated 2026-08-22 14:15 UTC}

Ayush7614

Verified: packages/core/src/repowise/core/ingestion/change_detector.py:1 now accepts stale_pages map and sorts staleness-first (oldest stale outranks fresh regardless of pagerank) for budget-constrained regen (#851). Plumbing via get_stale_file_page_ages and CLI/server call sites is best-effort with fallback to pure-importance. Tests in test_change_detector.py + test_stale_file_page_ages.py cover ordering. No fake code — LGTM.

sloemo01 requested review from RaghavChamadiya and swati510 as code owners August 22, 2026 14:14

Ayush7614 approved these changes Aug 22, 2026

View reviewed changes

Navigation Menu

Uh oh!

Uh oh!

fix(generation): prioritize oldest stale pages when the budget is constrained - #1845

fix(generation): prioritize oldest stale pages when the budget is constrained#1845
sloemo01 wants to merge 1 commit into
repowise-dev:mainfrom
sloemo01:fix/staleness-sort

sloemo01 commented Aug 22, 2026

Uh oh!

repowise-bot Bot commented Aug 22, 2026

Uh oh!

Ayush7614 left a comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Uh oh!

Uh oh!

Conversation

sloemo01 commented Aug 22, 2026

Summary

Plumbing

Tests

Verification

Uh oh!

repowise-bot Bot commented Aug 22, 2026

Uh oh!

Ayush7614 left a comment

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants