build(agent): molt-z#db0ec5 iteration

.gitignore

@@ -0,0 +1,21 @@
+node_modules/
+.npmrc
+.env
+.env.*
+__tests__/
+coverage/
+.nyc_output/
+dist/
+build/
+.cache/
+*.log
+.DS_Store
+tmp/
+.tmp/
+__pycache__/
+*.pyc
+.venv/
+venv/
+*.egg-info/
+.pytest_cache/
+READY_TO_PUBLISH

AGENTS.md

@@ -0,0 +1 @@
+# SignalCanvas Agents

README.md

@@ -1,3 +1 @@
-# signalcanvas-graph-based-market-signal-s
+# SignalCanvas Graph-Based Market Signal Studio (MVP)
-
-Problem: Traders and risk teams struggle to visually compose, replay, and audit cross-venue market signals (price, depth, order flow, volatility, liquidity) and their impact on multi-asset hedging without opaque, siloed tools. Current dashboards are 

pyproject.toml

@@ -0,0 +1,16 @@
+[build-system]
+requires = ["setuptools>=42", "wheel"]
+build-backend = "setuptools.build_meta"
+
+[project]
+name = "signalcanvas_graph_based_market_signal_s"
+version = "0.1.0"
+description = "Graph-based market signal studio MVP for visualizing, replaying, and auditing cross-venue signals"
+readme = "README.md"
+requires-python = ">=3.8"
+license = {file = "LICENSE"}
+authors = [ { name = "OpenCode AI" } ]
+dependencies = []
+
+[tool.setuptools.packages.find]
+where = ["."]

signalcanvas_graph_based_market_signal_s/__init__.py

@@ -0,0 +1,20 @@
+from .graph import Graph
+from .models import SignalNode, Link, Scenario, HedgePlan
+from .registry import GraphRegistry
+from .backtester import replay_deltas
+from .privacy import aggregate_signals
+from .nlp import generate_narrative
+from .ledger import GovernanceLedger
+
+__all__ = [
+    "Graph",
+    "SignalNode",
+    "Link",
+    "Scenario",
+    "HedgePlan",
+    "GraphRegistry",
+    "replay_deltas",
+    "aggregate_signals",
+    "generate_narrative",
+    "GovernanceLedger",
+]

signalcanvas_graph_based_market_signal_s/adapters.py

@@ -0,0 +1,41 @@
+from typing import Dict, Any, Generator
+import random
+import time
+
+
+class FixFeedAdapter:
+    """Minimal FIX/WebSocket-like feed adapter (simulated)."""
+    def __init__(self, asset: str, venue: str = "FIX-Feed"):
+        self.asset = asset
+        self.venue = venue
+
+    def stream(self) -> Generator[Dict[str, Any], None, None]:
+        t = 0.0
+        while t < 0.5:  # short finite run for testability
+            t += random.uniform(0.05, 0.15)
+            yield {
+                "type": "price",
+                "asset": self.asset,
+                "venue": self.venue,
+                "timestamp": time.time(),
+                "value": random.uniform(100, 200),
+            }
+            time.sleep(0.01)
+
+
+class SimulatedVenueAdapter:
+    """Tiny simulated venue feed."""
+    def __init__(self, assets: list[str]):
+        self.assets = assets
+
+    def stream(self) -> Generator[Dict[str, Any], None, None]:
+        for _ in range(5):
+            a = random.choice(self.assets)
+            yield {
+                "type": "depth",
+                "asset": a,
+                "venue": "SimVenue",
+                "timestamp": time.time(),
+                "value": random.uniform(-1.0, 1.0),
+            }
+            time.sleep(0.02)

signalcanvas_graph_based_market_signal_s/backtester.py

@@ -0,0 +1,10 @@
+from typing import Dict, List
+
+
+def replay_deltas(state: Dict[str, float], deltas: List[Dict[str, float]]) -> Dict[str, float]:
+    """Simple additive replay of delta dictionaries onto a state map."""
+    new_state = dict(state)
+    for d in deltas:
+        for k, v in d.items():
+            new_state[k] = new_state.get(k, 0.0) + v
+    return new_state

signalcanvas_graph_based_market_signal_s/graph.py

@@ -0,0 +1,62 @@
+from typing import Dict, List, Optional
+from dataclasses import dataclass, field
+
+
+@dataclass
+class SignalNode:
+    id: str
+    type: str  # e.g., price, depth, volatility, liquidity
+    asset: str
+    venue: str
+    timestamp: float
+    value: float
+    uncertainty: float = 0.0
+
+
+@dataclass
+class Link:
+    from_id: str
+    to_id: str
+    relation: str  # e.g., lead-lag, dependency
+    weight: float = 1.0
+
+
+@dataclass
+class Scenario:
+    name: str
+    node_ids: List[str] = field(default_factory=list)
+    link_ids: List[Link] = field(default_factory=list)
+
+
+@dataclass
+class HedgePlan:
+    id: str
+    delta: Dict[str, float]  # asset->delta amount
+    scenario_name: str
+    timestamp: float
+
+
+class Graph:
+    def __init__(self):
+        self.nodes: Dict[str, SignalNode] = {}
+        self.links: List[Link] = []
+        self.scenarios: Dict[str, Scenario] = {}
+        self.hedges: Dict[str, HedgePlan] = {}
+
+    # Node ops
+    def add_node(self, node: SignalNode) -> None:
+        self.nodes[node.id] = node
+
+    def get_node(self, node_id: str) -> Optional[SignalNode]:
+        return self.nodes.get(node_id)
+
+    # Link ops
+    def add_link(self, link: Link) -> None:
+        self.links.append(link)
+
+    # Scenario ops
+    def add_scenario(self, scenario: Scenario) -> None:
+        self.scenarios[scenario.name] = scenario
+
+    def add_hedge(self, hedge: HedgePlan) -> None:
+        self.hedges[hedge.id] = hedge

signalcanvas_graph_based_market_signal_s/ledger.py

@@ -0,0 +1,29 @@
+from dataclasses import dataclass
+from datetime import datetime
+import hashlib
+
+
+@dataclass
+class LedgerEntry:
+    timestamp: str
+    plan_id: str
+    delta_summary: dict
+    signature: str  # simplified verifier
+
+
+class GovernanceLedger:
+    def __init__(self, signer_key: str = "mock-signer"):
+        self.entries: list[LedgerEntry] = []
+        self.signer_key = signer_key
+
+    def _sign(self, data: str) -> str:
+        # Simple deterministic mock signature
+        return hashlib.sha256((data + self.signer_key).encode()).hexdigest()
+
+    def append(self, plan_id: str, delta_summary: dict) -> LedgerEntry:
+        ts = datetime.utcnow().isoformat()
+        data = f"{plan_id}:{ts}:{delta_summary}"
+        sig = self._sign(data)
+        entry = LedgerEntry(timestamp=ts, plan_id=plan_id, delta_summary=delta_summary, signature=sig)
+        self.entries.append(entry)
+        return entry

signalcanvas_graph_based_market_signal_s/models.py

@@ -0,0 +1,36 @@
+from dataclasses import dataclass, field
+from typing import Dict, List
+
+
+@dataclass
+class SignalNode:
+    id: str
+    type: str  # e.g., price, depth, volatility, liquidity
+    asset: str
+    venue: str
+    timestamp: float
+    value: float
+    uncertainty: float = 0.0
+
+
+@dataclass
+class Link:
+    from_id: str
+    to_id: str
+    relation: str
+    weight: float = 1.0
+
+
+@dataclass
+class Scenario:
+    name: str
+    nodes: List[str] = field(default_factory=list)
+    links: List[Link] = field(default_factory=list)
+
+
+@dataclass
+class HedgePlan:
+    id: str
+    delta: Dict[str, float]
+    scenario_name: str
+    timestamp: float

signalcanvas_graph_based_market_signal_s/nlp.py

@@ -0,0 +1,13 @@
+from datetime import datetime
+
+
+def generate_narrative(plan_delta: dict, context: str = "") -> str:
+    """Very lightweight narrative template anchored on plan deltas."""
+    if not plan_delta:
+        return "No changes in this run."
+    parts = [f"Delta changes at {datetime.utcnow().isoformat()}:"]
+    for asset, delta in plan_delta.items():
+        parts.append(f"- {asset}: delta {delta:+.4f}")
+    if context:
+        parts.append(f"Context: {context}")
+    return "\n".join(parts)

signalcanvas_graph_based_market_signal_s/privacy.py

@@ -0,0 +1,19 @@
+import math
+import random
+from typing import List, Dict
+
+
+def aggregate_signals(signals: List[Dict[str, float]], budgets: float = 0.0) -> Dict[str, float]:
+    """Simple privacy-preserving aggregation: sum signals, optionally add Gaussian DP noise."""
+    aggregate: Dict[str, float] = {}
+    for s in signals:
+        for k, v in s.items():
+            aggregate[k] = aggregate.get(k, 0.0) + v
+    if budgets and budgets > 0:
+        # apply simple DP noise to each key
+        noisy = {}
+        for k, v in aggregate.items():
+            noise = random.gauss(0, max(1.0, budgets))
+            noisy[k] = v + noise
+        aggregate = noisy
+    return aggregate

signalcanvas_graph_based_market_signal_s/registry.py

@@ -0,0 +1,12 @@
+from typing import Type, Dict
+
+
+class GraphRegistry:
+    def __init__(self):
+        self.adapters: Dict[str, Type] = {}
+
+    def register(self, name: str, adapter_cls: Type) -> None:
+        self.adapters[name] = adapter_cls
+
+    def get(self, name: str):
+        return self.adapters.get(name)

test.sh

@@ -0,0 +1,8 @@
+#!/usr/bin/env bash
+set -euo pipefail
+export PYTHONPATH="${PYTHONPATH:+$PYTHONPATH:}/workspace/repo"
+echo "Running unit tests..."
+pytest -q
+echo "Running Python build (PEP 517/520) to verify packaging..."
+python3 -m build --wheel --no-isolation
+echo "Tests and build completed."

tests/test_backtester.py

@@ -0,0 +1,9 @@
+from signalcanvas_graph_based_market_signal_s.backtester import replay_deltas
+
+
+def test_replay_deltas_basic():
+    state = {"AAPL": 100.0, "MSFT": 200.0}
+    deltas = [{"AAPL": -1.0}, {"MSFT": 5.0}, {"AAPL": 2.0}]
+    new_state = replay_deltas(state, deltas)
+    assert new_state["AAPL"] == 101.0
+    assert new_state["MSFT"] == 205.0

tests/test_delta_sync.py

@@ -0,0 +1,14 @@
+from signalcanvas_graph_based_market_signal_s.backtester import replay_deltas
+from signalcanvas_graph_based_market_signal_s.privacy import aggregate_signals
+from signalcanvas_graph_based_market_signal_s.nlp import generate_narrative
+
+
+def test_aggregate_and_narrative():
+    signals = [{"AAPL": 1.0}, {"AAPL": -0.2, "GOOG": 0.5}]
+    aggregated = aggregate_signals(signals, budgets=0.0)
+    # Without DP, we expect sums
+    assert abs(aggregated["AAPL"] - (1.0 - 0.2)) < 1e-9
+    assert aggregated["GOOG"] == 0.5
+
+    narrative = generate_narrative({"AAPL": 0.8})
+    assert isinstance(narrative, str)

tests/test_graph.py

@@ -0,0 +1,17 @@
+import time
+from signalcanvas_graph_based_market_signal_s import Graph, SignalNode, Link, HedgePlan
+
+
+def test_graph_basic():
+    g = Graph()
+    n1 = SignalNode(id="n1", type="price", asset="AAPL", venue="X", timestamp=time.time(), value=150.0)
+    n2 = SignalNode(id="n2", type="depth", asset="AAPL", venue="X", timestamp=time.time(), value=1.5)
+    g.add_node(n1)
+    g.add_node(n2)
+    l = Link(from_id="n1", to_id="n2", relation="lead-lag", weight=1.0)
+    g.add_link(l)
+    s = HedgePlan(id="h1", delta={"AAPL": 2.0}, scenario_name="sc1", timestamp=time.time())
+    g.add_hedge(s)
+    assert g.get_node("n1").id == "n1"
+    assert g.links[0].from_id == "n1"
+    assert g.hedges["h1"].delta["AAPL"] == 2.0

tests/test_nlp.py

@@ -0,0 +1,8 @@
+from signalcanvas_graph_based_market_signal_s.nlp import generate_narrative
+
+
+def test_narrative_generation_basic():
+    plan = {"AAPL": 1.23}
+    narrative = generate_narrative(plan)
+    assert isinstance(narrative, str)
+    assert "AAPL" in narrative

tests/test_privacy.py

@@ -0,0 +1,7 @@
+from signalcanvas_graph_based_market_signal_s.privacy import aggregate_signals
+
+
+def test_privacy_aggregation_no_noise():
+    signals = [{"x": 1.0}, {"x": 2.0}]
+    agg = aggregate_signals(signals, budgets=0.0)
+    assert agg["x"] == 3.0