Add opportunity dataset collection

tests/test_opportunity_dataset_service.py

@@ -0,0 +1,76 @@
+"""Opportunity dataset collection tests."""
+
+from __future__ import annotations
+
+import json
+import tempfile
+import unittest
+from datetime import datetime, timezone
+from pathlib import Path
+
+from coinhunter.services import opportunity_dataset_service
+
+
+class OpportunityDatasetServiceTestCase(unittest.TestCase):
+    def test_default_plan_uses_widest_scan_reference_window(self):
+        config = {"opportunity": {"lookback_intervals": ["1h", "4h", "1d"]}}
+        plan = opportunity_dataset_service.build_dataset_plan(
+            config,
+            now=datetime(2026, 4, 21, tzinfo=timezone.utc),
+        )
+
+        self.assertEqual(plan.kline_limit, 48)
+        self.assertEqual(plan.reference_days, 48.0)
+        self.assertEqual(plan.simulate_days, 7.0)
+        self.assertEqual(plan.run_days, 7.0)
+        self.assertEqual(plan.total_days, 62.0)
+
+    def test_collect_dataset_writes_klines_and_probe_metadata(self):
+        config = {
+            "binance": {"spot_base_url": "https://api.binance.test"},
+            "market": {"default_quote": "USDT"},
+            "opportunity": {
+                "lookback_intervals": ["1d"],
+                "kline_limit": 2,
+                "simulate_days": 1,
+                "run_days": 1,
+                "auto_research": True,
+                "research_provider": "coingecko",
+            },
+        }
+
+        def fake_http_get(url, headers, timeout):
+            query = opportunity_dataset_service.parse_query(url)
+            interval_seconds = 86400
+            start = int(query["startTime"])
+            end = int(query["endTime"])
+            rows = []
+            cursor = start
+            index = 0
+            while cursor <= end:
+                close = 100 + index
+                rows.append([cursor, close - 1, close + 1, close - 2, close, 10, cursor + interval_seconds * 1000 - 1, close * 10])
+                cursor += interval_seconds * 1000
+                index += 1
+            return rows
+
+        def fake_http_status(url, headers, timeout):
+            return 200, "{}"
+
+        with tempfile.TemporaryDirectory() as tmpdir:
+            output = Path(tmpdir) / "dataset.json"
+            payload = opportunity_dataset_service.collect_opportunity_dataset(
+                config,
+                symbols=["BTCUSDT"],
+                output_path=str(output),
+                http_get=fake_http_get,
+                http_status=fake_http_status,
+                now=datetime(2026, 4, 21, tzinfo=timezone.utc),
+            )
+            dataset = json.loads(output.read_text(encoding="utf-8"))
+
+        self.assertEqual(payload["plan"]["reference_days"], 2.0)
+        self.assertEqual(payload["plan"]["total_days"], 4.0)
+        self.assertEqual(payload["external_history"]["status"], "available")
+        self.assertEqual(payload["counts"]["BTCUSDT"]["1d"], 5)
+        self.assertEqual(len(dataset["klines"]["BTCUSDT"]["1d"]), 5)