Quant-Strategy-Tokenizer

KO | EN

gitlite — search

#bash #proxy #data-science #llm #mysql #database #json #cli #react-native #ios

Quant-Strategy-Tokenizer

(★ 9)

Agent-oriented modules for decomposing quantitative trading strategies into reusable strategy tokens.

파일 탐색기

ZIP 다운로드

- - ci.yml
- - QLIB_ADAPTER_BOUNDARY.md
  - QLIB_ADAPTER_GUIDE.md
  - README.md
- - gkr-artifact-naming.md
  - token-first-after-core.md
- - AGENT_PLAYBOOK.md
  - AGENT_TAKEOVER_PROMPT.md
  - conformance.md
  - CUSTOM_TOKEN_GUIDE.md
  - methodology.md
  - README.md
  - RECIPE_AUTHORING_GUIDE.md
  - SECONDARY_DEVELOPMENT_GUIDE.md
  - task_contract.md
  - TOKEN_REGISTRATION_GUIDE.md
  - USAGE_GUIDE.md
  - workflow.md
- - coverage_taxonomy.md
  - custom_token_route_policy.md
  - false_supported_protocol.md
  - kernel_gap_decision_protocol.md
  - market_weight_protocol.md
  - README.md
- - performance-90d.png
  - PROJECT_EXPERIENCE.md
- - custom_token_governance_review.md
  - dogfood_target_set.md
  - external_benchmark_sources.md
  - kernel_gap_review.md
  - original_failure_strategy_dogfood.md
  - reserved_non_goal_boundary_review.md
  - stage_3b_decision.md
  - strategy_coverage_frontier.md
  - strategy_coverage_matrix.yaml
  - strategy_coverage_public_statement.md
  - strategy_coverage_report.md
  - token_contract_audit.md
  - token_coverage_review.md
  - token_gap_report.md
  - token_surface_acceptance.md
- - adapter_manifest-0.4.schema.json
  - approval_record-0.4.schema.json
  - artifact_base-0.4.schema.json
  - backtest_evidence-0.4.schema.json
  - custom_token_audit-0.4.schema.json
  - decision-0.4.schema.json
  - decision_aggregator-0.4.schema.json
  - decision_fold_policy-0.4.schema.json
  - decision_monoid-0.4.schema.json
  - execution_grant-0.4.schema.json
  - execution_report-0.4.schema.json
  - implementation_ref-0.4.schema.json
  - ir-0.4.schema.json
  - lock-0.4.schema.json
  - numeric_policy-0.4.schema.json
  - panel_group_spec-0.4.schema.json
  - panel_missing_policy-0.4.schema.json
  - panel_representation-0.4.schema.json
  - panel_selection_weight-0.4.schema.json
  - panel_temporal_state-0.4.schema.json
  - panel_universe_mask-0.4.schema.json
  - port_spec-0.4.schema.json
  - port_temporal_rule-0.4.schema.json
  - portfolio_snapshot-0.4.schema.json
  - profile_policy-0.4.schema.json
  - runtime_environment_ref-0.4.schema.json
  - state_fsm-0.4.schema.json
  - state_policy-0.4.schema.json
  - token_evolution_policy-0.4.schema.json
  - token_pack-0.4.schema.json
  - token_spec-0.4.schema.json
  - token_surface-0.4.schema.json
  - type_spec-0.4.schema.json
- architecture.md
- FINAL_ACCEPTANCE.md
- FINAL_HANDOFF.md
- FINAL_REPORT.md
- FINAL_SCOPE.md
- NON_GOALS.md
- reference.md
- ROADMAP_CLOSED.md
- security.md
- token_coverage.md
- token_family_registry.md
- - kdj_cross_basic.gkr.yaml
  - kdj_with_ema_filter.gkr.yaml
  - README.md
- - __init__.py
- - __init__.py
  - adapter_manifest.py
  - artifact_id.py
  - backtest_evidence.py
  - base.py
  - decimal_string.py
  - execution_report.py
  - portfolio_snapshot.py
  - safety.py
- - __init__.py
  - context.py
  - errors.py
  - output.py
  - result.py
- - __init__.py
  - audit.py
  - implementation.py
  - models.py
  - pack_io.py
  - reference_validation.py
  - service.py
- - __init__.py
  - model.py
  - reference.py
  - token_pack.py
- - __init__.py
  - base.py
  - feature_frame.py
  - market_frame.py
  - signal_frame.py
  - trace_log.py
- - __init__.py
  - audit_chain_hash.py
  - behavior_hash.py
  - common.py
  - expected_artifact_hash.py
  - graph_hash.py
  - implementation_ref_hash.py
  - instance_hash.py
  - param_hash.py
  - runtime_environment_hash.py
  - signature_hash.py
  - token_pack_hash.py
  - token_spec_hash.py
- - __init__.py
  - canonical.py
  - hash.py
  - loader.py
  - panel_validation.py
  - pathing.py
  - schema.py
  - temporal_validation.py
  - validator.py
- - __init__.py
  - stateless.py
- - __init__.py
  - policy.py
- - __init__.py
  - model.py
  - operators.py
  - reference_validation.py
  - token_pack.py
- - __init__.py
  - input_spec.py
  - output_spec.py
  - port_spec.py
  - temporal.py
- - __init__.py
  - policy.py
- - __init__.py
  - fsm.py
  - policy.py
  - reducers.py
  - reference.py
  - reference_validation.py
  - token_pack.py
  - trace.py
- - __init__.py
  - policy.py
- - __init__.py
  - builtin.py
  - lock.py
  - maturity.py
  - pack.py
  - package_policy.py
  - reference.py
  - registry.py
  - spec.py
  - surface.py
- - __init__.py
  - type_spec.py
  - value_type.py
- - __init__.py
  - models.py
  - registry.py
- __init__.py
- __main__.py
- canonical_json.py
- cli.py
- py.typed
- - __init__.py
  - test_prompt_artifact_format.py
  - test_prompt_semantic_anchors.py
  - test_validate_prompt_set.py
  - test_verify_prompt_remote_artifacts.py
- - schema_helpers.py
  - test_adapter_manifest_pep440.py
  - test_artifact_id_determinism.py
  - test_backtest_evidence_schema.py
  - test_decimal_string_strict_canonical.py
  - test_execution_report_schema.py
  - test_portfolio_snapshot_schema.py
  - test_provenance_chain.py
  - test_raw_payload_rules.py
- - test_stable_json_bytes_public_api.py
- - test_gkr_suffix_gate.py
- - test_core_rule_token_batch.py
  - test_custom_token_governance.py
  - test_dogfood_target_set.py
  - test_frontier_gate.py
  - test_kernel_gap_review.py
  - test_original_failure_dogfood.py
  - test_panel_factor_weight_batch.py
  - test_reserved_non_goal_boundaries.py
  - test_state_gate_risk_batch.py
  - test_strategy_coverage_matrix_validator.py
  - test_strategy_coverage_report_tool.py
- - test_service.py
- - test_aggregators.py
  - test_decision_model.py
  - test_fold_policies.py
  - test_monoids.py
- - test_final_closure_docs.py
- - test_reference_custom_token_v04.py
  - test_reference_panel_v04.py
  - test_reference_state_v04.py
  - test_reference_strategy_hashes.py
  - test_reference_temporal_v04.py
  - test_stage3a_demo_hashes.py
- - test_csv_io_roundtrip.py
  - test_feature_frame_model.py
  - test_frame_hash_determinism.py
  - test_frame_uses_decimal_string_canonical.py
  - test_json_io_roundtrip.py
  - test_long_format_multi_symbol.py
  - test_market_frame_model.py
  - test_signal_frame_model.py
  - test_trace_log_model.py
- - test_behavior_hash.py
  - test_behavior_material_wp4.py
  - test_hash_kinds.py
  - test_signature_hash.py
- - test_canonical_v04.py
  - test_capabilities_and_token_ref.py
  - test_gkr_paths.py
  - test_ir_shell.py
  - test_temporal_validator_v04.py
  - test_type_shorthand_canonicalization.py
- - test_numeric_policy.py
- - test_wp8a_design_gate.py
  - test_wp8b_panel_type_layer.py
  - test_wp8c_panel_operators.py
  - test_wp8d_weight_operators.py
  - test_wp8e_reference.py
- - test_port_spec.py
  - test_temporal_rule.py
- - test_profile_policy.py
- - test_fsm.py
  - test_policy.py
  - test_reducers.py
  - test_reference_state.py
  - test_reference_validation_state.py
- - __init__.py
  - _helpers.py
  - test_demo_hash_sentinels.py
  - test_hash_registry_stability.py
  - test_profile_gate_matrix.py
  - test_reserved_design_rejection.py
  - test_token_contract_completeness.py
  - test_token_surface_acceptance.py
- - test_pr6_core_rule_tokens.py
  - test_pr8_panel_factor_weight_tokens.py
  - test_pr9_state_gate_risk_tokens.py
  - test_stage3a_acceptance.py
  - test_token_surface.py
  - test_wp3a1_primitives.py
  - test_wp3a2_series_tokens.py
  - test_wp3a3_decision_state_gate.py
  - test_wp3a4_panel_weight_risk_optimizer.py
  - test_wp3a5_reserved_design.py
- - test_token_evolution_policy.py
- - test_decision_algebra_pack.py
  - test_lock_integration.py
  - test_registry_v2.py
  - test_state_basic_pack.py
  - test_state_fsm_pack.py
  - test_token_pack_v2.py
  - test_token_spec_v2.py
- - test_panel_type_shell.py
  - test_timeseries_type.py
  - test_type_spec_parsing.py
- - test_validation_models.py
  - test_validator_registry.py
- __init__.py
- helpers.py
- __init__.py
- report_strategy_coverage.py
- validate_prompt_set.py
- validate_strategy_coverage_matrix.py
- verify_prompt_remote_artifacts.py
.gitattributes
.gitignore
CHANGELOG.md
CONTRIBUTING.md
LICENSE
pyproject.toml
README.md
requirements.txt

# CDN으로 사용하기

jsDelivr

jsDelivr는 공개 GitHub 리포지토리를 별도 설정 없이 CDN으로 즉시 서빙합니다. 버전과 파일을 고르면 웹페이지에 바로 붙일 수 있는 링크와 예시 코드가 만들어집니다.

버전

파일

링크

예시

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

Archived agent-ready research prototype for typed quantitative strategy records.

QST turns strategy intent into deterministic Graph Kernel Records (GKR): typed
`.gkr.yaml` files that can be validated, canonicalized, hashed, reviewed, and
handed off to agents or downstream systems.

It is not a trading bot, broker adapter, exchange adapter, backtester, optimizer
runtime, or production execution engine.

## Status

This repository is closed as an **archived agent-ready research prototype**.

The final tree includes:

- typed `qst-ir/0.4` strategy records
- deterministic `qst-canonical/0.4` canonical JSON
- graph, parameter, and instance hashes
- token surface governance and conformance tests
- Coverage Frontier v0.3 evidence and publication gate
- active agent prompt pack and handoff docs
- a Qlib partial workflow adapter proof

## Install

```bash
pip install -e ".[dev]"
```

## Quick Check

```bash
python -m qst.cli vocabulary --check
python -m qst.cli validate examples/strategies/kdj_cross_basic.gkr.yaml
python -m qst.cli hash examples/strategies/kdj_cross_basic.gkr.yaml
python -m qst.cli canonicalize examples/strategies/kdj_cross_basic.gkr.yaml --output .local_audit/kdj.canonical.json
```

## Qlib Adapter Proof

QST includes a partial Qlib workflow YAML importer:

```bash
python -m qst.cli adapter qlib import examples/adapters/qlib/workflow_config_lightgbm_alpha158.yaml --output .local_audit/qlib_lightgbm_alpha158.gkr.yaml --coverage .local_audit/qlib_lightgbm_alpha158.coverage.json
python -m qst.cli validate .local_audit/qlib_lightgbm_alpha158.gkr.yaml
python -m qst.cli hash .local_audit/qlib_lightgbm_alpha158.gkr.yaml
python -m qst.cli canonicalize .local_audit/qlib_lightgbm_alpha158.gkr.yaml --output .local_audit/qlib_lightgbm_alpha158.canonical.json
```

The adapter is record-layer evidence only. It does not import Qlib, run qrun,
train models, run inference, execute backtests, connect to brokers or exchanges,
or claim lossless Qlib conversion.

## What It Provides

- **Record layer:** stable `.gkr.yaml` strategy records with typed nodes,
  token refs, capabilities, and port signatures.
- **Identity layer:** canonical JSON plus deterministic graph, parameter, and
  instance hashes.
- **Token layer:** accepted, experimental, reserved, and custom-token surfaces
  with maturity and execution-support metadata.
- **Evidence layer:** public demos, reference diagnostics, hash sentinels,
  coverage matrix, coverage report, and dogfood cases.
- **Agent layer:** active prompt pack, takeover prompt, playbook, usage guide,
  and secondary development guides.

## Boundaries

QST deliberately does not provide:

- live trading
- broker or exchange integration
- full backtesting
- Qlib runtime replacement
- qrun execution
- model training or inference execution
- lossless Qlib conversion
- arbitrary Python strategy parsing
- production portfolio optimization
- profitability claims

Custom-token execution is explicit and approval-bound. Verification does not
import or execute user code; execution requires integrity verification, local
approval, an execution grant, and output validation.

## Handoff

Start here:

- [Final Handoff](docs/FINAL_HANDOFF.md)
- [Final Scope](docs/FINAL_SCOPE.md)
- [Final Acceptance](docs/FINAL_ACCEPTANCE.md)
- [Agent Takeover Prompt](docs/agent/AGENT_TAKEOVER_PROMPT.md)
- [Agent Playbook](docs/agent/AGENT_PLAYBOOK.md)
- [Usage Guide](docs/agent/USAGE_GUIDE.md)

## Key Paths

```text
qst/                         Python package and CLI
examples/strategies/         12 public GKR strategy examples
examples/adapters/qlib/      Qlib partial workflow adapter examples
examples/custom_token/       Custom-token reference example
tests/reference/             Deterministic reference fixtures and traces
tests/adapters/qlib/         Qlib adapter proof tests
docs/agent/                  Agent handoff and prompt guidance
docs/adapters/               Adapter boundary and Qlib adapter guide
docs/reports/                Coverage Frontier and acceptance reports
```

## Documentation

- [Architecture](docs/architecture.md)
- [Security](docs/security.md)
- [Reference](docs/reference.md)
- [Token Family Registry](docs/token_family_registry.md)
- [Token Coverage](docs/token_coverage.md)
- [Coverage Report](docs/reports/strategy_coverage_report.md)
- [Qlib Adapter Boundary](docs/adapters/QLIB_ADAPTER_BOUNDARY.md)
- [Agent Guidance](docs/agent/README.md)

## Project Identity

- Python import package: `qst`
- CLI: `qst`
- Distribution name: `quant-strategy-tokenizer`
- Editable strategy source: `.gkr.yaml`
- Packaged record suffix: `.gkr`
- IR: `qst-ir/0.4`
- Canonical schema: `qst-canonical/0.4`

Quant-Strategy-Tokenizer

파일 탐색기

# CDN으로 사용하기

Was this content helpful?

Recent Feedback

Download README