Python SDK

# List all projects (returns paginated iterator)
projects = client.projects.list(status="active")
for project in projects:
    print(f"{project.id}: {project.name}")

# Create a new project with governance settings
project = client.projects.create(
    name="Fluid Dynamics Study",
    description="Discover governing equations for turbulent flow",
    settings={
        "default_mode": "neuro_symbolic",
        "governance": {
            "promotion_ceiling": "validate",
            "require_negative_controls": True,
            "truth_dial_threshold": 0.85
        }
    }
)

# Retrieve a specific project
project = client.projects.get("proj_tds_8f2a1b")

# Update project settings
project = client.projects.update(
    "proj_tds_8f2a1b",
    description="Updated description",
    settings={"governance": {"truth_dial_threshold": 0.90}}
)

# Upload a dataset from a local file (parquet, CSV, HDF5)
dataset = client.datasets.upload(
    project_id="proj_tds_8f2a1b",
    file_path="turbulence_re5000.parquet",
    name="Turbulence Re=5000",
    variables=["velocity_x", "velocity_y", "pressure", "time"]
)
print(f"Uploaded: {dataset.id} ({dataset.size_bytes} bytes, {dataset.rows} rows)")

# Get automated profile (statistics, quality, distributions)
profile = client.datasets.profile(
    project_id="proj_tds_8f2a1b",
    dataset_id="ds_turb_001"
)
for col in profile.columns:
    print(f"  {col.name}: mean={col.mean:.3f}, std={col.std:.3f}, nulls={col.null_count}")

# List datasets in a project
datasets = client.datasets.list(project_id="proj_tds_8f2a1b", status="ready")

# Create a campaign with budget and governance
campaign = client.campaigns.create(
    project_id="proj_tds_8f2a1b",
    name="Reynolds Number Sweep",
    budget={"compute_units": 2500, "max_runs": 50},
    governance={
        "promotion_ceiling": "validate",
        "require_negative_controls": True,
        "auto_promote_to": "explore"
    }
)

# List campaigns with filtering
campaigns = client.campaigns.list(
    project_id="proj_tds_8f2a1b",
    status="active"
)

# Get campaign details including budget usage
campaign = client.campaigns.get(
    project_id="proj_tds_8f2a1b",
    campaign_id="camp_rns_4e8f2a"
)
print(f"Budget: {campaign.budget.compute_units_used}/{campaign.budget.compute_units} CU")

# Symbolic discovery — find closed-form governing equations
run_sym = client.runs.create(
    project_id="proj_tds_8f2a1b",
    campaign_id="camp_rns_4e8f2a",
    mode="symbolic",
    dataset_id="ds_turb_001",
    parameters={"constraints": ["conservation_of_energy"], "max_complexity": 5}
)

# Neural discovery — learn latent representations
run_neural = client.runs.create(
    project_id="proj_tds_8f2a1b",
    campaign_id="camp_rns_4e8f2a",
    mode="neural",
    dataset_id="ds_turb_001",
    parameters={"config": "default"}
)

# Neuro-Symbolic — combine neural embeddings with symbolic regression
run_ns = client.runs.create(
    project_id="proj_tds_8f2a1b",
    campaign_id="camp_rns_4e8f2a",
    mode="neuro_symbolic",
    dataset_id="ds_turb_001",
    parameters={
        "config": "default",
        "max_complexity": 5,
        "cross_validate": True
    }
)

# Causal mode — causal dynamics discovery
run_cde = client.runs.create(
    project_id="proj_tds_8f2a1b",
    campaign_id="camp_rns_4e8f2a",
    mode="causal",
    dataset_id="ds_turb_001",
    parameters={"config": "default"}
)

# Wait for a run to complete (blocks with polling)
completed = client.runs.wait(run_sym.id, poll_interval=5.0, timeout=3600.0)
print(f"Status: {completed.status}, Claims: {completed.claims_count}")

# Check status without blocking
status = client.runs.status(run_ns.id)
print(f"Stage: {status.pipeline.current_stage}, Progress: {status.pipeline.progress}")

# Cancel a queued or running run
client.runs.cancel(run_neural.id)

# List runs in a campaign
runs = client.runs.list(
    project_id="proj_tds_8f2a1b",
    campaign_id="camp_rns_4e8f2a",
    status="completed",
    mode="symbolic"
)

# List claims with filters
claims = client.claims.list(
    project_id="proj_tds_8f2a1b",
    run_id="run_ns_7a3b2c",
    type="law",                  # "law", "causal", "conservation"
    tier="explore",              # "explore", "validate", "publish"
    min_score=0.8,
    sort="score",
    order="desc"
)

for claim in claims:
    print(f"[{claim.type}] {claim.summary}")
    print(f"  Score: {claim.score}, Tier: {claim.tier}")
    print(f"  Scope: {claim.scope}")

# Get a specific claim with full detail
claim = client.claims.get(project_id="proj_tds_8f2a1b", claim_id="clm_law_9d4e5f")
print(f"Expression: {claim.expression}")
print(f"Evidence: r²={claim.evidence.r_squared}, controls={claim.evidence.negative_controls_passed}")

# Promote a claim through Truth Dial tiers
promoted = client.claims.promote(
    project_id="proj_tds_8f2a1b",
    claim_id="clm_law_9d4e5f",
    target_tier="validate",
    rationale="Passed all negative controls and cross-dataset validation."
)
print(f"Promoted to {promoted.tier} (from {promoted.previous_tier})")

# Export claims to JSON or CSV
client.claims.export(
    project_id="proj_tds_8f2a1b",
    run_id="run_ns_7a3b2c",
    format="csv",
    output_path="claims_export.csv"
)

# Query ledger entries for a specific claim
entries = client.ledger.query(
    project_id="proj_tds_8f2a1b",
    claim_id="clm_law_9d4e5f",
    event_type="claim_promoted"
)
for entry in entries:
    print(f"[{entry.timestamp}] {entry.event_type} by {entry.actor}")
    print(f"  Details: {entry.details}")

# Verify ledger integrity (chain validation)
verification = client.ledger.verify(project_id="proj_tds_8f2a1b")
print(f"Integrity: {verification.status}")  # "valid" or "corrupted"
print(f"Entries verified: {verification.entries_checked}")

# Export full ledger for external auditing
client.ledger.export(
    project_id="proj_tds_8f2a1b",
    format="json",
    output_path="ledger_audit.json"
)