scikit-rank 0.1.0__tar.gz

scikit_rank-0.1.0/.editorconfig

@@ -0,0 +1,19 @@
+root = true
+
+[*]
+charset = utf-8
+indent_style = space
+indent_size = 4
+end_of_line = lf
+insert_final_newline = true
+trim_trailing_whitespace = true
+
+[*.{py,jsonnet,json,md}]
+indent_size = 4
+
+[*.{yml,yaml}]
+indent_size = 2
+
+[{Makefile,*.Dockerfile}]
+indent_style = tab
+tab_width = 4

scikit_rank-0.1.0/.gitignore

@@ -0,0 +1,22 @@
+# Python-generated files
+__pycache__/
+*.py[oc]
+build/
+dist/
+wheels/
+*.egg-info
+.ipynb_checkpoints/
+
+# Virtual environments
+.venv
+
+# Devenv
+.devenv*
+devenv.local.nix
+devenv.local.yaml
+
+# Direnv
+.direnv
+
+# Ouptuts
+data

scikit_rank-0.1.0/.pi-subagents/artifacts/8682734e-5509-4c5e-850d-a0a13f741c27_scout_input.md

@@ -0,0 +1,60 @@
+# Task for scout
+
+Audit this repository for the requested cleanup/refactor. Identify package/module layout, build/test/docs configuration, every reference to scikit-rank/current package name and forbidden names beeformer/dcn-exps/tabicl (excluding ipynb), and recommend low-value tests to remove versus keep. Do not edit files. Report concise file paths and commands.
+
+---
+**Output:**
+Write your findings to exactly this path: /Users/a.n.milogradskiy/Developer/personal/scikit-rank/.pi-subagents/artifacts/outputs/8682734e-5509-4c5e-850d-a0a13f741c27/audit-scout.md
+This path is authoritative for this run.
+Ignore any other output filename or output path mentioned elsewhere, including output destinations in the base agent prompt, system prompt, or task instructions.
+
+## Acceptance Contract
+Acceptance level: reviewed
+Completion is not accepted from prose alone. End with a structured acceptance report.
+
+Criteria:
+- criterion-1: Implement the requested change without widening scope
+- criterion-2: Return evidence sufficient for an independent acceptance review
+
+Required evidence: changed-files, tests-added, commands-run, validation-output, residual-risks, no-staged-files
+
+Review gate: required by reviewer.
+
+Finish with a fenced JSON block tagged `acceptance-report` in this shape:
+Use empty arrays when no items apply; array fields contain strings unless object entries are shown.
+```acceptance-report
+{
+  "criteriaSatisfied": [
+    {
+      "id": "criterion-1",
+      "status": "satisfied",
+      "evidence": "specific proof"
+    }
+  ],
+  "changedFiles": [
+    "src/file.ts"
+  ],
+  "testsAddedOrUpdated": [
+    "test/file.test.ts"
+  ],
+  "commandsRun": [
+    {
+      "command": "command",
+      "result": "passed",
+      "summary": "short result"
+    }
+  ],
+  "validationOutput": [
+    "validation output or concise summary"
+  ],
+  "residualRisks": [
+    "none"
+  ],
+  "noStagedFiles": true,
+  "diffSummary": "short description of the diff",
+  "reviewFindings": [
+    "blocker: file.ts:12 - issue found, or no blockers"
+  ],
+  "manualNotes": "anything else the parent should know"
+}
+```

scikit_rank-0.1.0/.pi-subagents/artifacts/8682734e-5509-4c5e-850d-a0a13f741c27_scout_meta.json

@@ -0,0 +1,28 @@
+{
+  "runId": "8682734e-5509-4c5e-850d-a0a13f741c27",
+  "agent": "scout",
+  "task": "Audit this repository for the requested cleanup/refactor. Identify package/module layout, build/test/docs configuration, every reference to scikit-rank/current package name and forbidden names beeformer/dcn-exps/tabicl (excluding ipynb), and recommend low-value tests to remove versus keep. Do not edit files. Report concise file paths and commands.\n\n---\n**Output:**\nWrite your findings to exactly this path: /Users/a.n.milogradskiy/Developer/personal/scikit-rank/.pi-subagents/artifacts/outputs/8682734e-5509-4c5e-850d-a0a13f741c27/audit-scout.md\nThis path is authoritative for this run.\nIgnore any other output filename or output path mentioned elsewhere, including output destinations in the base agent prompt, system prompt, or task instructions.\n\n## Acceptance Contract\nAcceptance level: reviewed\nCompletion is not accepted from prose alone. End with a structured acceptance report.\n\nCriteria:\n- criterion-1: Implement the requested change without widening scope\n- criterion-2: Return evidence sufficient for an independent acceptance review\n\nRequired evidence: changed-files, tests-added, commands-run, validation-output, residual-risks, no-staged-files\n\nReview gate: required by reviewer.\n\nFinish with a fenced JSON block tagged `acceptance-report` in this shape:\nUse empty arrays when no items apply; array fields contain strings unless object entries are shown.\n```acceptance-report\n{\n  \"criteriaSatisfied\": [\n    {\n      \"id\": \"criterion-1\",\n      \"status\": \"satisfied\",\n      \"evidence\": \"specific proof\"\n    }\n  ],\n  \"changedFiles\": [\n    \"src/file.ts\"\n  ],\n  \"testsAddedOrUpdated\": [\n    \"test/file.test.ts\"\n  ],\n  \"commandsRun\": [\n    {\n      \"command\": \"command\",\n      \"result\": \"passed\",\n      \"summary\": \"short result\"\n    }\n  ],\n  \"validationOutput\": [\n    \"validation output or concise summary\"\n  ],\n  \"residualRisks\": [\n    \"none\"\n  ],\n  \"noStagedFiles\": true,\n  \"diffSummary\": \"short description of the diff\",\n  \"reviewFindings\": [\n    \"blocker: file.ts:12 - issue found, or no blockers\"\n  ],\n  \"manualNotes\": \"anything else the parent should know\"\n}\n```",
+  "exitCode": 0,
+  "model": "openai-codex/gpt-5.6-terra:low",
+  "attemptedModels": [
+    "openai-codex/gpt-5.6-terra:low"
+  ],
+  "modelAttempts": [
+    {
+      "model": "openai-codex/gpt-5.6-terra:low",
+      "success": true,
+      "exitCode": 0,
+      "usage": {
+        "input": 64973,
+        "output": 9106,
+        "cacheRead": 388608,
+        "cacheWrite": 0,
+        "cost": 0.3961745,
+        "turns": 10
+      }
+    }
+  ],
+  "transcriptPath": "/Users/a.n.milogradskiy/Developer/personal/scikit-rank/.pi-subagents/artifacts/8682734e-5509-4c5e-850d-a0a13f741c27_scout_transcript.jsonl",
+  "skills": [],
+  "timestamp": 1784053453342
+}

scikit_rank-0.1.0/.pi-subagents/artifacts/8682734e-5509-4c5e-850d-a0a13f741c27_scout_output.md

@@ -0,0 +1,59 @@
+# Code Context
+
+## Files Retrieved
+1. `pyproject.toml` (lines 1-112) — package name, Hatch target, deps, Ruff.
+2. `README.md` (lines 1-189) — primary docs; install/test/lint commands.
+3. `src/scikit_rank/__init__.py` (lines 1-53) — public API exports.
+4. `src/scikit_rank/sklearn/estimator.py` (lines 1-900) — `DCNBase`, classifier, regressor, ranker.
+5. `src/scikit_rank/{data.py,preprocessing.py,run.py,factories.py}`, `modules/`, `train/`, `utils/`, `sklearn/` — implementation.
+6. `tests/conftest.py` (lines 1-19), `tests/test_*.py` — 24 tests / 3,855 LOC.
+7. `.pre-commit-config.yaml` (lines 1-29), `devenv.nix` (lines 1-35), `devenv.yaml`, `uv.lock:2093` — tooling/lock.
+8. `exps/{README.md,bars/README.md,temporal_article/README.md}` — experiment docs.
+
+## Key Code
+Distribution is `scikit-rank` (`pyproject.toml:2`); source/import package is `scikit_rank` (`pyproject.toml:48`). Hatch builds only `src/scikit_rank`. Public flow is `__init__` → sklearn estimator → `DataRouter`/`TabularPreprocessor` → factories/`TrainingRun` → `Trainer`; `modules/` holds DCNv2/losses/reducers.
+
+A rename must update both forms, the lockfile, Hatch/Ruff config, docs/scripts/tests/docstrings, IPC prefix at `src/scikit_rank/data.py:376`, and temporary router columns at `src/scikit_rank/sklearn/_data_router.py:40-41` if internal residue is forbidden. `exps/bars/` is separate config-driven benchmark infrastructure; untracked `exps/temporal_article/run_final.sh` and FuxiCTR vendor files are not committed scope.
+
+## Architecture
+sklearn wrappers share preprocessing, factory-built PyTorch DCNv2, and Accelerate/Ignite training. `DataRouter` supports eager and LazyFrame input; lazy data streams through Arrow IPC. Experiments are separate from the installable package.
+
+## Build/Test/Docs
+* Build: `uv build`, Hatchling (`pyproject.toml:43-48`), Python `>=3.12,<3.14` (`:6`).
+* Test: no pytest config/markers/CI found; `uv run python -m pytest -q` (`README.md:184-188`), pytest in dev group (`pyproject.toml:23-32`).
+* Lint/hooks: Ruff (`pyproject.toml:50-112`); pre-commit excludes tests/exps/notebooks from Ruff (`.pre-commit-config.yaml:21-29`).
+* Dev: devenv provides Python 3.12, uv, Ruff, pre-commit, OpenMP (`devenv.nix:19-34`).
+* Docs: Markdown README/experiment READMEs; no Sphinx/MkDocs. Notebook excluded.
+
+## Name-reference Inventory
+```bash
+git grep -inE 'scikit[ _-]?rank|beeformer|dcn-exps|tabicl' -- ':!*.ipynb'
+git grep -ilE 'beeformer|dcn-exps|tabicl' -- ':!*.ipynb'
+git ls-files | grep -iE 'scikit|beeformer|dcn-exps|tabicl'
+```
+* `beeformer`, `dcn-exps`, `tabicl`: **zero tracked matches** (case-insensitive, notebook excluded). Ignored `.pytest_cache/v/cache/lastfailed` has historic `tabicl/...`; clear caches for whole-worktree acceptance scans.
+* Main distribution/docs refs: `pyproject.toml:2,48,104`; `uv.lock:2093`; `README.md:1,5,12,45,51,60,83,95`; `exps/README.md:3,21,107`; `exps/bars/README.md:4,104,167`; `exps/bars/{_adapters.py,_data.py,train.py}`; `exps/scripts/{compare_predictions_auc.py,evaluate_ranking.py,train_temporal_dcn.py:25,27}`; `exps/temporal_article/{README.md,final_metrics.json}`.
+* Source: all `src/scikit_rank/` paths plus intra-package imports/docstrings in `__init__.py`, `data.py`, `factories.py`, `preprocessing.py`, `run.py`, `modules/{__init__,dcn}.py`, `sklearn/{__init__,_data_router,_types,estimator}.py`, `train/{__init__,optimizers,options}.py`, `utils/{__init__,ipc_materializer}.py`.
+* Tests: every test except `tests/conftest.py` references `scikit_rank`; exact inventory: `git grep -l scikit_rank -- tests`.
+
+## Test Triage
+**Keep:** sklearn/loss contracts (`test_sklearn_compliance`, `test_estimator_specs`, `test_loss_model_alignment`, `test_ranker_eval_metrics`); data/lazy tests (`test_data_router`, `test_lazy_arrow_source`, `test_lazy_dataframe_support`, `test_lazy_eager_parity`); feature tests (`test_dcn_features`, `test_encoder_specs`, `test_embedding_features`, `test_multihash_features`, `test_ple_encoder`, `test_quantile_normalize`, `test_inner_cross_layers`, `test_extra_layers`); training tests (`test_group_batch_sampler`, `test_optimizers`, `test_ema`). Each covers a distinct advertised feature.
+
+**Lowest-value consolidation only if DDP is deliberately de-scoped:** `test_accelerate_multiworker.py`, `test_ddp_loss_reduce.py`, `test_ddp_metric_gather.py`, `test_ddp_ranking_training.py`, `test_ddp_sharding.py` (1,113 LOC, overlapping subprocess setup). Do not remove while README promises multi-GPU/DDP. For faster retained-DDP CI, keep `test_ddp_ranking_training` and `test_group_batch_sampler`, consolidate unique reduce/gather/sharding assertions before removing probes.
+
+## Start Here
+Open `pyproject.toml`, then update all `git grep` results, move `src/scikit_rank`, and validate install/import/full tests.
+
+## Commands Run
+```bash
+git status --short; git ls-files
+git grep -inE 'scikit[ _-]?rank|beeformer|dcn-exps|tabicl' -- ':!*.ipynb'
+git grep -ilE 'beeformer|dcn-exps|tabicl' -- ':!*.ipynb'
+git diff --cached --name-only
+wc -l tests/test_*.py
+```
+No tracked source/config was edited. Five pre-existing temporal experiment changes are unstaged; index is empty.
+
+```acceptance-report
+{"criteriaSatisfied":[{"id":"criterion-1","status":"satisfied","evidence":"Read-only audit; bounded cleanup targets."},{"id":"criterion-2","status":"satisfied","evidence":"Paths, inventory commands, test triage, and Git state supplied."}],"changedFiles":[],"testsAddedOrUpdated":[],"commandsRun":[{"command":"git grep -inE 'scikit[ _-]?rank|beeformer|dcn-exps|tabicl' -- ':!*.ipynb'","result":"passed","summary":"Current-name refs found; no tracked forbidden-name match."},{"command":"git status --short && git diff --cached --name-only","result":"passed","summary":"Five pre-existing unstaged changes; index empty."},{"command":"wc -l tests/test_*.py","result":"passed","summary":"24 tests, 3,855 lines."}],"validationOutput":["Tracked forbidden names: zero matches excluding notebooks.","No staged files."],"residualRisks":["Ignored pytest cache contains tabicl paths.","Do not overwrite existing temporal experiment edits.","Removing DDP tests while retaining DDP claims risks regressions."],"noStagedFiles":true,"diffSummary":"No repository diff; audit artifact only.","reviewFindings":["No blockers: rename both scikit-rank and scikit_rank, lockfile, and all grep references."],"manualNotes":"No tests run: read-only audit."}
+```