xrtm-data 0.7.0__tar.gz → 0.7.2__tar.gz

{xrtm_data-0.7.0/src/xrtm_data.egg-info → xrtm_data-0.7.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: xrtm-data
-Version: 0.7.0
+Version: 0.7.2
 Summary: The Snapshot Vault for XRTM.
 Author-email: XRTM Team <moy@xrtm.org>
 License-Expression: Apache-2.0

{xrtm_data-0.7.0 → xrtm_data-0.7.2}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "xrtm-data"
-version = "0.7.0"
+version = "0.7.2"
 description = "The Snapshot Vault for XRTM."
 readme = "README.md"
 requires-python = ">=3.11,<3.13"

{xrtm_data-0.7.0 → xrtm_data-0.7.2}/src/xrtm/data/corpora/real_binary.py

@@ -33,7 +33,7 @@ from typing import Any, Mapping, Optional, Sequence
 
 from pydantic import BaseModel, ConfigDict, Field, model_validator
 
-from xrtm.data.core import DataSource
+from xrtm.data.core import DataSource, SourceTemporalIntegrityError
 from xrtm.data.core.schemas import ForecastQuestion, MetadataBase
 
 REAL_BINARY_CORPUS_ID = "xrtm-real-binary-v1"
@@ -498,6 +498,11 @@ class RealBinaryCorpusSource(DataSource):
         self, query: Optional[str] = None, limit: int = 5, *, snapshot_time: Optional[datetime] = None
     ) -> list[ForecastQuestion]:
         r"""Fetch deterministic corpus questions, optionally filtering title/content text."""
+        if snapshot_time is not None:
+            raise SourceTemporalIntegrityError(
+                "RealBinaryCorpusSource cannot honour snapshot_time; it returns the embedded corpus "
+                "as-is. Use a time-aware source or omit the snapshot."
+            )
         query_lower = query.lower() if query else None
         matches: list[ForecastQuestion] = []
         for question in self._questions:
@@ -512,6 +517,10 @@ class RealBinaryCorpusSource(DataSource):
         self, question_id: str, *, snapshot_time: Optional[datetime] = None
     ) -> Optional[ForecastQuestion]:
         r"""Retrieve a deterministic corpus question by id."""
+        if snapshot_time is not None:
+            raise SourceTemporalIntegrityError(
+                "RealBinaryCorpusSource cannot honour snapshot_time; it returns the embedded corpus as-is."
+            )
         question = self._questions_by_id.get(question_id)
         if question is None:
             return None

{xrtm_data-0.7.0 → xrtm_data-0.7.2}/src/xrtm/data/providers/online/kalshi.py

@@ -177,7 +177,7 @@ class KalshiSource(DataSource, ResolutionSource):
             title=title[:500] or ticker,
             description=f"{subtitle[:500]}\n\n{price_info}".strip(),
             metadata=MetadataBase(
-                snapshot_time=_parse_ts(item.get("close_time")) if item.get("close_time") else datetime.now(timezone.utc),
+                snapshot_time=datetime.now(timezone.utc),  # fetch time; close_time is in the future
                 source_version="kalshi",
                 tags=["kalshi", "binary", "prediction-market"],
                 raw_data={

{xrtm_data-0.7.0 → xrtm_data-0.7.2}/src/xrtm/data/providers/online/manifold.py

@@ -163,14 +163,13 @@ class ManifoldSource(DataSource, ResolutionSource):
         market_id = str(item.get("id", ""))
         probability = item.get("probability")
         price_info = f"Current probability: {float(probability)*100:.1f}%" if probability is not None else ""
-        close_time = item.get("closeTime")
 
         return ForecastQuestion(
             id=f"manifold-{market_id}",
             title=str(item.get("question", ""))[:500],
             description=f"Manifold market {market_id}. {price_info}".strip(),
             metadata=MetadataBase(
-                snapshot_time=_parse_ts(close_time) if close_time else datetime.now(timezone.utc),
+                snapshot_time=datetime.now(timezone.utc),  # fetch time; close_time is in the future
                 source_version="manifold",
                 tags=["manifold", "binary", "prediction-market"],
                 raw_data={

{xrtm_data-0.7.0 → xrtm_data-0.7.2}/src/xrtm/data/providers/online/polymarket.py

@@ -79,8 +79,9 @@ class PolymarketSource(DataSource, ResolutionSource):
         try:
             data = self._get_json(url)
         except DataSourceError:
-            # Gamma API may return a list directly
-            url = f"{self.api_base}/events?limit={min(limit, 50)}&closed=false"
+            # Fallback: same markets endpoint without the volume ordering,
+            # which some Gamma deployments reject.
+            url = f"{self.api_base}/markets?limit={min(limit, 100)}&closed=false"
             data = self._get_json(url)
 
         # Gamma API may wrap in "markets" key or return list directly

{xrtm_data-0.7.0 → xrtm_data-0.7.2/src/xrtm_data.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: xrtm-data
-Version: 0.7.0
+Version: 0.7.2
 Summary: The Snapshot Vault for XRTM.
 Author-email: XRTM Team <moy@xrtm.org>
 License-Expression: Apache-2.0

{xrtm_data-0.7.0 → xrtm_data-0.7.2}/tests/test_online_sources.py

@@ -28,6 +28,28 @@ def test_sources_implement_resolution_interface():
         assert isinstance(source, ResolutionSource)
 
 
+@pytest.mark.asyncio
+async def test_manifold_snapshot_time_is_fetch_time(monkeypatch):
+    from datetime import datetime, timedelta, timezone
+
+    source = ManifoldSource()
+    monkeypatch.setattr(
+        source,
+        "_get_json",
+        lambda url: [
+            {"id": "abc", "question": "Will X?", "isResolved": False, "closeTime": 4102444800000},  # year 2100
+        ],
+    )
+
+    questions = await source.fetch_questions(limit=1)
+
+    assert questions
+    snapshot = questions[0].metadata.snapshot_time
+    assert snapshot is not None
+    # Snapshot must be the fetch time, never the (future) close time.
+    assert snapshot <= datetime.now(timezone.utc) + timedelta(seconds=5)
+
+
 @pytest.mark.asyncio
 async def test_manifold_fetch_resolved_maps_yes_no(monkeypatch):
     source = ManifoldSource()
@@ -172,3 +194,38 @@ async def test_kalshi_fetch_resolved_paginates_cursor(monkeypatch):
 
     assert len(resolved) == 250
     assert calls["count"] == 2
+
+
+@pytest.mark.asyncio
+async def test_polymarket_fetch_questions_fallback_uses_markets(monkeypatch):
+    from xrtm.data.core import DataSourceError
+
+    source = PolymarketSource()
+    calls = {"urls": []}
+
+    def _get_json(url):
+        calls["urls"].append(url)
+        if "order=volume24hr" in url:
+            raise DataSourceError("volume ordering unsupported")
+        return [
+            {
+                "id": "m1",
+                "question": "Will X happen?",
+                "description": "d",
+                "outcomes": '["Yes", "No"]',
+                "outcomePrices": '["0.6", "0.4"]',
+                "volume": 1000,
+                "liquidity": 500,
+                "endDate": "2026-12-31T00:00:00Z",
+                "slug": "will-x",
+                "closed": False,
+            }
+        ]
+
+    monkeypatch.setattr(source, "_get_json", _get_json)
+
+    questions = await source.fetch_questions(limit=1)
+
+    assert len(calls["urls"]) == 2
+    assert "/markets?" in calls["urls"][1]
+    assert questions and questions[0].id == "polymarket-m1"

{xrtm_data-0.7.0 → xrtm_data-0.7.2}/tests/test_real_binary_corpus.py

@@ -26,6 +26,19 @@ from xrtm.data.corpora import (
 )
 
 
+@pytest.mark.asyncio
+async def test_real_binary_corpus_source_rejects_snapshot_time():
+    from datetime import datetime
+
+    from xrtm.data import SourceTemporalIntegrityError
+
+    source = RealBinaryCorpusSource()
+    with pytest.raises(SourceTemporalIntegrityError):
+        await source.fetch_questions(limit=1, snapshot_time=datetime(2026, 1, 1))
+    with pytest.raises(SourceTemporalIntegrityError):
+        await source.get_question_by_id("any", snapshot_time=datetime(2026, 1, 1))
+
+
 def test_real_binary_corpus_shape_and_resolution_coverage():
     records = load_real_binary_corpus()