Tutorial Part 2: Building the Parent SNOs

# Hypothetical CNS 2.0 Tools Library
from cns_tools import StructuredNarrativeObject, ReasoningGraph, EvidenceSet
from cns_tools.utils import get_text_embedding

# We'll also need a unique identifier for our evidence
import hashlib

def hash_source(text):
    return hashlib.sha256(text.encode()).hexdigest()

# --- Mock Evidence Sources ---
# In a real scenario, these would be pointers to actual documents (e.g., DOIs).
# Here, we'll use hashes of hypothetical paper titles as placeholders.

EVIDENCE_HALL_1859 = hash_source("Hall, J. (1859). Palaeontology of New York.")
EVIDENCE_DANA_1873 = hash_source("Dana, J.D. (1873). On the origin of mountains.")
EVIDENCE_DIETZ_1961 = hash_source("Dietz, R.S. (1961). Continent and Ocean Basin Evolution by Spreading of the Sea Floor.")
EVIDENCE_VINE_1963 = hash_source("Vine, F.J. & Matthews, D.H. (1963). Magnetic Anomalies over Oceanic Ridges.")
EVIDENCE_WILSON_1965 = hash_source("Wilson, J.T. (1965). A new class of faults and their bearing on continental drift.")

# 1. Define the Hypothesis Embedding
# In a real system, this would be generated by a sophisticated language model.
hypothesis_geosyncline = "Mountain ranges are formed by the vertical collapse and uplift of large, sediment-filled troughs (geosynclines) on a static, cooling Earth."
H_geosyncline = get_text_embedding(hypothesis_geosyncline)

# 2. Build the Reasoning Graph (G)
G_geosyncline = ReasoningGraph(graph_id="G_Geo_v1")

# Add claims (nodes) to the graph
G_geosyncline.add_claim("c1", "The Earth is a cooling and contracting body.")
G_geosyncline.add_claim("c2", "Thick sedimentary deposits accumulate in large troughs (geosynclines).")
G_geosyncline.add_claim("c3", "The crust buckles under the sediment weight and compressional forces from cooling.")
G_geosyncline.add_claim("c4", "This buckling leads to vertical uplift, forming mountain ranges.")
G_geosyncline.add_claim("c5", "Continents and ocean basins are permanent, fixed features.")

# Add reasoning relationships (edges) between claims
G_geosyncline.add_edge("c1", "c3", "supports") # Cooling earth supports buckling
G_geosyncline.add_edge("c2", "c3", "supports") # Sediment accumulation supports buckling
G_geosyncline.add_edge("c3", "c4", "implies")  # Buckling implies uplift
G_geosyncline.add_edge("c5", "c1", "is_consistent_with") # Fixed continents are consistent with a simple cooling model

# 3. Populate the Evidence Set (E)
E_geosyncline = EvidenceSet(evidence_id="E_Geo_v1")
E_geosyncline.add_evidence(EVIDENCE_HALL_1859, "Supports the existence of thick sedimentary layers in mountain belts.", supports_claims=["c2"])
E_geosyncline.add_evidence(EVIDENCE_DANA_1873, "Provides a mechanism for compression and uplift.", supports_claims=["c3", "c4"])

# 4. Instantiate the SNO
# The Trust Score (T) is initially null, as it will be assigned by the Critic Pipeline.
SNO_geosyncline = StructuredNarrativeObject(
    hypothesis_embedding=H_geosyncline,
    reasoning_graph=G_geosyncline,
    evidence_set=E_geosyncline,
    trust_score=None # To be computed later
)

print("SNO_Geosyncline created successfully.")

# 1. Define the Hypothesis Embedding
hypothesis_tectonics = "The Earth's surface is composed of rigid lithospheric plates that move, and their interactions at boundaries are the primary cause of mountain building, earthquakes, and volcanism."
H_tectonics = get_text_embedding(hypothesis_tectonics)

# 2. Build the Reasoning Graph (G)
G_tectonics = ReasoningGraph(graph_id="G_PT_v1")

# Add claims (nodes)
G_tectonics.add_claim("c1", "The lithosphere is divided into rigid plates.")
G_tectonics.add_claim("c2", "New oceanic crust is generated at mid-ocean ridges (seafloor spreading).")
G_tectonics.add_claim("c3", "Oceanic crust is consumed at subduction zones.")
G_tectonics.add_claim("c4", "Plate motion is driven by mantle convection.")
G_tectonics.add_claim("c5", "Mountain ranges are formed by the collision of continental plates or subduction.")
G_tectonics.add_claim("c6", "The continents are not fixed but drift over time.")

# Add reasoning relationships (edges)
G_tectonics.add_edge("c2", "c1", "supports")
G_tectonics.add_edge("c3", "c1", "supports")
G_tectonics.add_edge("c1", "c5", "implies")
G_tectonics.add_edge("c4", "c1", "provides_mechanism_for")
G_tectonics.add_edge("c2", "c6", "implies") # Seafloor spreading implies continental drift

# This is a key point of conflict with the other SNO
G_tectonics.add_claim("c7_conflict", "Continents and ocean basins are NOT permanent, fixed features.")
G_tectonics.add_edge("c6", "c7_conflict", "implies")

# 3. Populate the Evidence Set (E)
E_tectonics = EvidenceSet(evidence_id="E_PT_v1")
E_tectonics.add_evidence(EVIDENCE_DIETZ_1961, "Proposes the mechanism of seafloor spreading.", supports_claims=["c2"])
E_tectonics.add_evidence(EVIDENCE_VINE_1963, "Symmetrical magnetic stripes around mid-ocean ridges provide strong proof of seafloor spreading.", supports_claims=["c2"])
E_tectonics.add_evidence(EVIDENCE_WILSON_1965, "Identifies transform faults, a necessary component of plate boundary interactions.", supports_claims=["c1", "c5"])

# 4. Instantiate the SNO
SNO_plate_tectonics = StructuredNarrativeObject(
    hypothesis_embedding=H_tectonics,
    reasoning_graph=G_tectonics,
    evidence_set=E_tectonics,
    trust_score=None # To be computed later
)

print("SNO_PlateTectonics created successfully.")

# DSPy signature for systematic SNO generation
class StatisticalSNOGenerator(dspy.Signature):
    """Generate high-quality opposing SNOs for statistical synthesis validation."""
    
    debate_specification = dspy.InputField(desc="Scientific debate with documented resolution and primary sources")
    quality_requirements = dspy.InputField(desc="Evidence standards, reasoning complexity, hypothesis precision")
    validation_framework = dspy.InputField(desc="Ground truth criteria and success metrics")
    
    sno_historical = dspy.OutputField(desc="SNO representing historical/minority position")
    sno_modern = dspy.OutputField(desc="SNO representing accepted/majority position") 
    quality_metrics = dspy.OutputField(desc="Evidence count, reasoning depth, source authenticity scores")
    validation_criteria = dspy.OutputField(desc="Measurable synthesis success criteria")

# Quality control parameters derived from manual prototype:
QUALITY_STANDARDS = {
    'min_evidence_sources': 3,  # Based on manual SNO construction
    'min_reasoning_nodes': 5,   # Complexity threshold from prototype
    'hypothesis_precision': 0.9, # Semantic clarity requirement
    'source_authenticity': 0.95, # Historical accuracy standard
    'dialectical_opposition': 0.8 # CScore threshold for valid pairs
}

# Domain expansion for statistical validation:
VALIDATION_DOMAINS = [
    {'domain': 'geology', 'debate': 'plate_tectonics_vs_geosyncline', 'prototype': True},
    {'domain': 'biology', 'debate': 'darwin_vs_lamarck_evolution'},
    {'domain': 'physics', 'debate': 'wave_vs_particle_light'},
    {'domain': 'chemistry', 'debate': 'atomic_vs_continuous_matter'},
    {'domain': 'cosmology', 'debate': 'big_bang_vs_steady_state'},
    {'domain': 'medicine', 'debate': 'germ_vs_miasma_theory'},
    {'domain': 'astronomy', 'debate': 'heliocentric_vs_geocentric'},
    {'domain': 'genetics', 'debate': 'mendelian_vs_blending_inheritance'}
]