Merge pull request #19 from StrategyLogic/dev

Refactor strategic actor simulation and improve scenario validation

Author: RCFans

README.md

@@ -2,7 +2,7 @@
 
 **The Strategic Reasoning Engine.**
 
-[![Pylint](https://github.com/StrategyLogic/omen/actions/workflows/pylint.yml/badge.svg)](https://github.com/StrategyLogic/omen/actions/workflows/pylint.yml) [![Package](https://github.com/StrategyLogic/omen/actions/workflows/python-package.yml/badge.svg)](https://github.com/StrategyLogic/omen/actions/workflows/python-package.yml)
+[![Pylint](https://github.com/StrategyLogic/omen/actions/workflows/pylint.yml/badge.svg)](https://github.com/StrategyLogic/omen/actions/workflows/pylint.yml) [![Package](https://github.com/StrategyLogic/omen/actions/workflows/python-package.yml/badge.svg)](https://github.com/StrategyLogic/omen/actions/workflows/python-package.yml) ![](https://img.shields.io/pypi/l/omenai?timestamp=20260406) ![](https://img.shields.io/pepy/dt/omenai) ![](https://img.shields.io/pypi/v/omenai?label=%20) ![](https://img.shields.io/github/stars/StrategyLogic/omen?style=social) 
 
 > **Simulate the Signs. Reveal the Chaos.**
 

README.zh.md

@@ -2,7 +2,7 @@
 
 **开源战略推演引擎。**
 
-[![Pylint](https://github.com/StrategyLogic/omen/actions/workflows/pylint.yml/badge.svg)](https://github.com/StrategyLogic/omen/actions/workflows/pylint.yml) [![Package](https://github.com/StrategyLogic/omen/actions/workflows/python-package.yml/badge.svg)](https://github.com/StrategyLogic/omen/actions/workflows/python-package.yml)
+[![Pylint](https://github.com/StrategyLogic/omen/actions/workflows/pylint.yml/badge.svg)](https://github.com/StrategyLogic/omen/actions/workflows/pylint.yml) [![Package](https://github.com/StrategyLogic/omen/actions/workflows/python-package.yml/badge.svg)](https://github.com/StrategyLogic/omen/actions/workflows/python-package.yml) ![](https://img.shields.io/pypi/l/omenai?timestamp=20260406) ![](https://img.shields.io/pepy/dt/omenai) ![](https://img.shields.io/pypi/v/omenai?label=%20) ![](https://img.shields.io/github/stars/StrategyLogic/omen?style=social) 
 
 > 模拟征兆，揭示混沌。
 

cases/situations/sap_reltio_acquisition.md

@@ -1,31 +1,31 @@
 # Situation Case: SAP looking to pull more external data into its AI platform with Reltio acquisition
 - Source URL: https://www.theregister.com/2026/03/30/sap_reltio/
-- Capture Date: 2026-04-05
+- Capture Date: 2026-04-06
 - Event Date: Mon 30 Mar 2026
 
 **Context:** SAP is acquiring master data management and data integration specialist Reltio to help integrate data from outside SAP's application portfolio into its AI platform. The move is positioned to boost adoption of SAP's Business Data Cloud (BDC), a service launched in February 2025 in alliance with Databricks.
 
 ## The Story
-SAP announces its acquisition of Reltio, a master data management specialist, positioning it as a way to integrate external data into SAP's AI platform and boost its Business Data Cloud. The article details Reltio's technology and customer base, then contrasts the acquisition with SAP's prior BDC strategy, which faced low awareness and customer difficulties. Analysts frame the Reltio move as an expansion of SAP's data integration focus from sharing SAP data externally to also harmonizing non-SAP data for use within SAP.
+SAP announces its acquisition of Reltio, a master data management firm, positioning it as a way to integrate non-SAP data into its AI platform and boost its Business Data Cloud service. The article notes SAP's previous struggles with BDC adoption, citing user group research showing low familiarity. Analysts frame the acquisition as an extension of SAP's earlier strategy shift with BDC, moving from sharing SAP data externally to also harmonizing external data for use within SAP. A consultant highlights a critical finding from user research about customers stuck in mid-migration paying for unusable systems, noting this line 'is going to haunt SAP's sales deck for a while.'
 
 ### 1. Core Facts (The Facts)
 - SAP is to acquire master data management and data integration specialist Reltio.
-- The acquisition aims to help integrate data from outside SAP's application portfolio into its AI platform.
+- The acquisition aims to help integrate data from outside SAP's broad application portfolio into its AI platform.
 - The move may be aimed at boosting adoption of SAP's Business Data Cloud (BDC).
 - BDC was launched in February 2025 in an alliance with data lake and AI platform vendor Databricks.
-- Reltio was founded in 2011 and aimed to create a 'cloud-native' approach to data integration and MDM.
-- In 2024, Reltio migrated its underlying database from a self-managed instance of Cassandra to Spanner on Google Cloud.
+- Reltio uses an AI-based entity resolution system to find and merge related records into a 'golden record'.
+- SAP expects to use Reltio's system to integrate data across SAP and non-SAP applications.
 - Reltio customers include Pfizer, Radisson, and Warner Bros.
-- SAP expects to use Reltio's AI-based entity resolution system to integrate data across SAP and non-SAP applications.
+- Reltio migrated its underlying database from Cassandra to Spanner on Google Cloud in 2024.
 
 ### 2. Existing Data and Research Findings (Non-opinion)
-- Research from DSAG in December found that 83 percent of its members were only slightly familiar with BDC or not familiar with it at all.
+- In December, research from DSAG found that 83 percent of its members were only slightly familiar with BDC or not familiar with it at all.
 - DSAG found that BDC customers stuck mid-migration end up paying full cloud fees for a system they cannot use.
 
 ### 3. Raw Context
 - SAP argues the move will help make BDC 'fully interoperable' with other stores of enterprise data to support the development of AI agents by cleansing and harmonizing data.
-- Reltio uses an AI-based entity resolution system to find and merge related records from different formats and applications into a 'golden record' of data about customers, products, suppliers, locations, and employees.
-- SAP has ERP, HR, CRM, supply chain, and procurement systems.
+- BDC was set up to build 'insight apps' that use analytics tools and AI models connected to real-time data from applications.
+- Reltio aimed to create a 'cloud-native' approach to data integration and MDM.
 - SAP has so far struggled to grab users' attention with BDC.
 
 ### 4. Direct Quotes (Speaker / Role)
@@ -37,7 +37,7 @@ SAP announces its acquisition of Reltio, a master data management specialist, po
 - SAP has so far struggled to grab users' attention with BDC.
 - 83 percent of DSAG members were only slightly familiar with BDC or not familiar with it at all.
 - BDC customers stuck mid-migration end up paying full cloud fees for a system they cannot use.
-- The DSAG finding about mid-migration fees 'is going to haunt SAP's sales deck for a while.'
+- The finding about customers paying for unusable systems is noted as a line that 'is going to haunt SAP's sales deck for a while.'
 
 ### 6. Unknowns / Open Questions
 - Whether the move will successfully boost adoption of SAP's Business Data Cloud.
@@ -46,7 +46,7 @@ SAP announces its acquisition of Reltio, a master data management specialist, po
 
 > SAP has so far struggled to grab users' attention with BDC. In December, research from DSAG, the German-speaking user group, found that 83 percent of its members were only slightly familiar with BDC or not familiar with it at all.
 
-> While BDC was mainly about sharing SAP data externally, the Reltio move is about harmonizing data from non-SAP systems and sharing it with SAP, he said. Binoy James, SVP applications at SAP consultancy Protera, noted DSAG had found that BDC customers stuck mid-migration end up paying full cloud fees for a system they cannot use. 'That line is going to haunt SAP's sales deck for a while,' he said.
+> Binoy James, SVP applications at SAP consultancy Protera, noted DSAG had found that BDC customers stuck mid-migration end up paying full cloud fees for a system they cannot use. 'That line is going to haunt SAP's sales deck for a while,' he said.
 
 ---
 

config/prompts/base.yaml

@@ -250,15 +250,66 @@ prompts:
     - version must be '0.1.0'
     - id must equal provided situation_id
     - context must preserve all provided context fields
-    - signals must be non-empty
-    - uncertainty_space.overall_confidence must be in [0,1]
-    - uncertainty_space.assumptions_explicit should include assumption objects with:
-      target_unknown, assumption_text, quality_score (0..1), coverage_type
-    - assumptions_explicit should map to known_unknowns when possible
+
+    === SIGNAL EXTRACTION RULES (CRITICAL) ===
+    
+    A "signal" is an independent environmental factor, NOT a restatement of the subject's actions or status.
+    
+    Independence Test (must pass for every signal):
+    - If the subject (e.g., SAP) disappeared tomorrow, would this factor still exist?
+    - If YES → it's a signal. If NO → it's context or event, not a signal.
+    
+    Examples of what signals ARE:
+    - "Enterprise AI procurement shifting from model-first to data-first" (market trend)
+    - "AI-native MDM replacing traditional MDM" (tech evolution)
+    - "Customer tolerance for 'pay-during-migration' fees decreasing" (market sentiment shift)
+    
+    Examples of what signals ARE NOT:
+    - "SAP's BDC has low customer awareness" (subject's product status → put in context)
+    - "SAP acquired Reltio" (subject's action → put in events)
+    - "Analyst said X about SAP" (commentary on subject → not independent)
+    
+    signals_template_json is a schema contract, not a prompt to summarize.
+    
+    Each signal must include:
+    - id: unique identifier (e.g., sig_tech_001, sig_market_002)
+    - name: concise, factor-focused (not subject-focused)
+    - domain: one of [tech, market, capital, standard, policy]
+    - strength: numeric in [0,1] (confidence that this signal is real and impactful)
+    - direction: one of [up, down, mixed] (direction of the signal's trend)
+    - mapped_targets: array of objects, each with:
+        - space: [TechSpace, MarketSpace]
+        - element_key: string (what this signal affects)
+        - impact_type: [driver, constraint, amplifier, dampener]
+        - impact_strength: numeric in [0,1]
+        - mechanism_conditions: object with impact_type-required fields:
+            - driver: activation_condition, expected_effect
+            - constraint: binding_condition, release_condition
+            - amplifier/dampener: modulation_target, modulation_condition, modulation_factor
+    - cascade_rules: array of {trigger_condition, next_signal_id, expected_lag} OR provide no_cascade_reason
+    - market_constraints: array of {constraint_key, binding_strength in [0,1]}
+    - evidence_basis: brief note on what evidence supports this signal (not the subject's response to it)
+    
+    Signal Quantity: Prefer 3 to 7 high-quality, independent signals over many weak summary bullets.
+    Signal Independence: Two signals that describe the same underlying factor from different angles should be merged.
+    
+    === CONTEXT PRESERVATION ===
+    - context.current_state: describe subject's situation, NOT signals
+    - context.known_unknowns: what subject doesn't know about the signals
+    - signals: what is happening in the environment, independent of subject
+    
+    === UNCERTAINTY SPACE ===
+    - overall_confidence: numeric in [0,1] (confidence in the situation assessment)
+    - assumptions_explicit: array of {target_unknown, assumption_text, quality_score, coverage_type}
+    - assumptions should map to known_unknowns when possible
+    
+    === OUTPUT ===
+    Return ONLY valid JSON object matching the structure defined in signals_template_json.
 
     situation_id: [[situation_id]]
     source_path: [[source_path]]
     context_json: [[context_json]]
+    signals_template_json: [[signals_template_json]]
 
     source_document:
     [[source_text]]
@@ -269,13 +320,18 @@ prompts:
 
     Scenario requirements:
     - Must include exactly A, B, C each once
-    - A intent: Aggressive/Alternative challenger, optimistic positive-signal assumptions
-    - B intent: Baseline/Conservative maintainer, linear extrapolation from current state
-    - C intent: Collapse/Contingency extreme-risk, negative-signal breakout
+    - A intent: Offense (Breakthrough action), proactive rule-changing and asymmetric advantage capture
+    - B intent: Defense (Protective action), survival and compliance under hard external constraints
+    - C intent: Confrontation action, direct rivalry under rule-bounded resource competition
     - each scenario must include goal, target, objective, variables, constraints, tradeoff_pressure, resistance_assumptions, modeling_notes
+    - constraints and tradeoff_pressure must reflect constraint-network and rival-network pressure when present
+    - planning_template_json is a rule template input; do not treat it as a prompt
+    - raw_prior_scores should be provided as array with A/B/C scenario_key and score values when available
 
     pack_id: [[pack_id]]
     pack_version: [[pack_version]]
+    planning_template_json: [[planning_template_json]]
+    planning_query_json: [[planning_query_json]]
     situation_artifact_json: [[situation_artifact_json]]
 
   situation_source_to_case_prompt: |

config/templates/planning.yaml

@@ -0,0 +1,33 @@
+template_id: scenario_planning_template
+template_version: 0.1.0
+required_query_fields:
+  - situation_id
+  - actor_ref
+  - space_inputs
+  - constraint_signals
+  - similarity_scores
+slot_policy:
+  - scenario_key: A
+    label: Offense
+    intent: Breakthrough action, proactive rule-changing and asymmetric advantage capture
+    default_prior: 0.4
+  - scenario_key: B
+    label: Defense
+    intent: Protective action, survival and compliance under hard external constraints
+    default_prior: 0.35
+  - scenario_key: C
+    label: Confrontation
+    intent: Confrontation action, direct rivalry under rule-bounded resource competition
+    default_prior: 0.25
+prompt_contract:
+  objective: Keep deterministic A/B/C planning with stable priors
+  notes:
+    - This file is a planning rule template, not an LLM prompt.
+    - Use it to constrain planning slots and prior defaults.
+  output_keys:
+    - pack_id
+    - pack_version
+    - derived_from_situation_id
+    - ontology_version
+    - scenarios
+    - raw_prior_scores

config/templates/signals.yaml

@@ -0,0 +1,103 @@
+template_id: situation_signals_v0_1_0
+template_version: 0.1.0
+description: Minimal mechanism-bearing signal schema for Situation_v0_1_0 artifacts.
+domain_values:
+  - tech
+  - market
+  - capital
+  - standard
+  - policy
+direction_values:
+  - up
+  - down
+  - mixed
+impact_types:
+  - driver
+  - constraint
+  - amplifier
+  - dampener
+expected_lag_values:
+  - short
+  - medium
+  - long
+space_values:
+  - TechSpace
+  - MarketSpace
+signal_schema:
+  required_fields:
+    - id
+    - name
+    - domain
+    - strength
+    - direction
+    - mapped_targets
+    - cascade_rules
+    - market_constraints
+    - mechanism_note
+  defaults:
+    domain: market
+    strength: 0.5
+    impact_strength: 0.5
+    direction_by_impact_type:
+      driver: up
+      constraint: down
+      amplifier: up
+      dampener: down
+    market_binding_strength: 0.6
+    expected_lag: medium
+    mapped_target_space: MarketSpace
+    no_cascade_reason: direct local effect in current horizon
+  mapped_target_shape:
+    required_fields:
+      - space
+      - element_key
+      - impact_type
+      - impact_strength
+      - mechanism_conditions
+    mechanism_conditions_shape:
+      required_by_impact_type:
+        driver:
+          - activation_condition
+          - expected_effect
+        constraint:
+          - binding_condition
+          - release_condition
+          - expected_effect
+        amplifier:
+          - modulation_target
+          - modulation_condition
+          - modulation_factor
+          - expected_effect
+        dampener:
+          - modulation_target
+          - modulation_condition
+          - modulation_factor
+          - expected_effect
+  cascade_rule_shape:
+    required_fields:
+      - trigger_condition
+      - next_signal_id
+      - expected_lag
+  market_constraint_shape:
+    required_fields:
+      - constraint_key
+      - binding_strength
+authoring_rules:
+  - Every signal must encode a mechanism, not a summary headline.
+  - strength and impact_strength must be numeric values in [0,1].
+  - direction must be one of up, down, mixed.
+  - mapped_targets must use space plus element_key and classify impact_type as driver, constraint, amplifier, or dampener.
+  - mapped_targets.mechanism_conditions must satisfy required fields by impact_type.
+  - cascade_rules must include trigger_condition, next_signal_id, and expected_lag unless no_cascade_reason is provided.
+  - constraints impact_type must link at least one market_constraints.constraint_key.
+  - market_constraints must capture explicit external frictions or gating conditions and binding_strength in [0,1].
+  - mechanism_note should explain why the signal matters to scenario planning.
+fallbacks:
+  activation_condition: Active under the current situation context and source-grounded constraints.
+  binding_condition: Constraint binds when external pressure dominates internal optionality.
+  release_condition: Constraint relaxes when ecosystem leverage or policy pressure shifts.
+  modulation_condition: Modulation appears when adjacent adoption and coordination signals move together.
+  expected_effect: Moves the mapped target trajectory in the declared direction.
+  modulation_target: primary_driver
+  cascade_trigger: If the signal intensifies in the next planning window.
+  market_constraint: Current market adoption and coordination frictions remain binding.

data/scenarios/nokia_v1/nokia-elop-2010_generation.json

@@ -0,0 +1,51 @@
+{
+  "artifact_type": "situation_generation_trace",
+  "situation_id": "nokia-elop-2010",
+  "situation_artifact_path": "data/scenarios/nokia_v1/nokia-elop-2010_situation.json",
+  "source_path": "cases/situations/nokia-elop-2010.md",
+  "pack_id": "nokia_v1",
+  "pack_version": "1.0.0",
+  "generated_at": "2026-04-06T16:45:04.903200",
+  "validation_passed": true,
+  "validation_issues": [],
+  "confidence": {
+    "confidence_risk": 0.4,
+    "confidence_overall": 0.6,
+    "overall_confidence": 0.6,
+    "coverage_ratio": 0.7125,
+    "guardrail_applied": false
+  },
+  "metrics": {
+    "known_unknowns_count": 4,
+    "assumptions_filled_count": 4,
+    "assumptions_quality_avg": 0.7125,
+    "cognitive_coverage": 0.7125,
+    "confidence_delta_cap": 0.5
+  },
+  "assumptions_explicit": [
+    {
+      "target_unknown": "The extent and nature of structural conflict between incumbent and transition teams.",
+      "assumption_text": "Internal organizational conflict is significant and will create substantial friction during any strategic pivot.",
+      "quality_score": 0.7,
+      "coverage_type": "partial"
+    },
+    {
+      "target_unknown": "The full impact of resource reallocation drag across platform programs.",
+      "assumption_text": "Dividing resources between Symbian and MeeGo is causing execution delays and quality issues in both programs.",
+      "quality_score": 0.8,
+      "coverage_type": "high"
+    },
+    {
+      "target_unknown": "The degree of cultural mismatch between historical operating model and new execution cadence.",
+      "assumption_text": "Nokia's traditional hardware-focused, long-cycle development culture is incompatible with the rapid software iteration required for modern smartphone competition.",
+      "quality_score": 0.75,
+      "coverage_type": "high"
+    },
+    {
+      "target_unknown": "The specific veto-node risk in key governance and partnership decisions.",
+      "assumption_text": "Key stakeholders within Nokia's governance structure could block or significantly delay critical partnership decisions needed for platform strategy.",
+      "quality_score": 0.6,
+      "coverage_type": "partial"
+    }
+  ]
+}