hypgraph 0.7.4__tar.gz → 0.7.5__tar.gz

{hypgraph-0.7.4 → hypgraph-0.7.5}/CHANGELOG.md

@@ -1,5 +1,15 @@
 # Changelog
  
+## 0.7.5 — 2026-10-28: a cited source is a line of support, and a finding of failure is not a fall
+
+- **A cited source is a line of support, and a succeeded run's "failed" outcome is a finding, not a fall.**
+  A knowledge claim that cites its sources (`cites` → source, or `derived_from` → source) read "no live
+  support · unverified" even after 0.7.4 counted bound results; the source is what such a claim rests on,
+  and its line is `external`. And a result with `outcome: failed` counted as a broken premise whatever its
+  run did: testers write "failed" for the code under test failing its spec (30 results in the challenge
+  records, 13 carrying checked claims that then read UNSUPPORTED — every support has fallen). A result a
+  succeeded run produced did happen; it falls only when there is no run or the run did not succeed.
+
 ## 0.7.4 — 2026-10-28: a bound result is a line of support, and a relation word is read as the relation it names
 
 - **A bound result is a line of support.** A claim's standing counted only `supports` edges into it, so a claim

{hypgraph-0.7.4 → hypgraph-0.7.5}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.5
 Name: hypgraph
-Version: 0.7.4
+Version: 0.7.5
 Summary: The evidence layer for AI research agents: a hash-chained research record that checks every claim as it is written and derives how well each is known.
 Project-URL: Homepage, https://hypgraph.com
 Project-URL: Documentation, https://hypgraph.com/engineers/

{hypgraph-0.7.4 → hypgraph-0.7.5}/src/hypgraph/__init__.py

@@ -16,7 +16,7 @@ Standing is derived on every read, never stored; every write is an event in a SH
 their numbers; reviews pin versions; independence is counted, not assumed. See https://hypgraph.com.
 """
 
-__version__ = "0.7.4"
+__version__ = "0.7.5"
 
 from .bundle import verify_bundle  # noqa: E402
 from .graph import Author, Graph, GraphError, file_sha256  # noqa: E402

{hypgraph-0.7.4 → hypgraph-0.7.5}/src/hypgraph/derive.py

@@ -134,7 +134,11 @@ def fallen(s: Snap, nid: str) -> bool:
     if n["type"] == "run" and n["body"].get("state") in ("failed", "timeout", "cancelled", "interrupted"):
         return True
     if n["type"] == "result" and n["body"].get("outcome") == "failed":
-        return True
+        # "failed" means the measurement didn't happen. A result a succeeded run produced did happen: its author
+        # wrote "failed" for what it measured (the code under test failing its spec — 30 results in the challenge
+        # records, 13 carrying checked claims that then read UNSUPPORTED). That is a finding, not a fall.
+        run = producing_run(s, nid)
+        return not (run and s.nodes[run]["body"].get("state") == "succeeded")
     return False
 
 
@@ -349,8 +353,8 @@ def dependents(s: Snap, nid: str, limit: int = 500) -> list[tuple[str, str, int]
 
 
 def supporters(s: Snap, nid: str) -> list[dict]:
-    """What stands behind a node: every `supports` edge into it, and — for a claim — the results it is derived from
-    and the results its measurements bind (each once). A claim that binds its numbers to a result rests on that
+    """What stands behind a node: every `supports` edge into it, and — for a claim — the results it is derived from,
+    the sources it derives from or cites, and the results its measurements bind (each once). A claim that binds its numbers to a result rests on that
     result whether or not anyone also wrote `supports`; until 0.7.4 only the edge counted, and most honest claims
     read "no live support" with their evidence "unverified"."""
     out = [e for e in s.inn.get(nid, []) if e["rel"] == "supports"]
@@ -358,10 +362,11 @@ def supporters(s: Snap, nid: str) -> list[dict]:
     if n and n["type"] == "claim":
         seen = {e["src"] for e in out}
         for e in s.out.get(nid, []):
-            if e["rel"] == "derived_from" and e["dst"] in s.nodes and s.nodes[e["dst"]]["type"] == "result" \
-                    and e["dst"] not in seen:
+            # a result it derives from, or a source it derives from or cites: what the claim rests on
+            if e["rel"] in ("derived_from", "cites") and e["dst"] in s.nodes and e["dst"] not in seen \
+                    and s.nodes[e["dst"]]["type"] in (("result", "source") if e["rel"] == "derived_from" else ("source",)):
                 seen.add(e["dst"])
-                out.append({"src": e["dst"], "rel": "derived_from", "dst": nid, "data": {}, "by": n["created_by"]})
+                out.append({"src": e["dst"], "rel": e["rel"], "dst": nid, "data": {}, "by": n["created_by"]})
         for m in n["body"].get("measurements") or []:
             rid = m.get("result") if isinstance(m, dict) else None
             if rid and rid in s.nodes and s.nodes[rid]["type"] == "result" and rid not in seen:

hypgraph-0.7.5/tests/test_cited_and_failed_support.py

@@ -0,0 +1,60 @@
+"""The last claims reading 'no live support' after bound results counted: a knowledge claim resting on the sources it
+cites, and claims on results whose author wrote outcome 'failed' for the code under test failing its spec, on runs
+that succeeded (E4: 13 checked claims read UNSUPPORTED — every support has fallen)."""
+import pytest
+
+from hypgraph import Graph, Policy
+from hypgraph import derive as D
+
+
+@pytest.fixture
+def env():
+    g = Graph(policy=Policy(reviewers={"critic"}, instrument="lab"))
+    inv = g.author("lead").investigation("Support", "Find out what a claim rests on.")["id"]
+    return g, inv
+
+
+def _state(g, nid):
+    return D.state(g.snapshot(), nid)
+
+
+def test_a_cited_source_is_a_line_of_support_rated_external(env):
+    g, inv = env
+    r = g.author("researcher")
+    s1 = r.add("source", inv, citation="config.py in the 2.4.0 release", url="file:sources/config.py",
+               passage='"retry_backoff": 30,       # seconds; 60 before 2.3.0')["id"]
+    s2 = r.add("source", inv, citation="The 2.4 configuration guide", url="file:sources/config-guide.md",
+               passage="retry_backoff defaults to 30 seconds")["id"]
+    c = r.add("claim", inv, statement="The default retry_backoff in the current release is 30 seconds.",
+              scope="The 2.4.0 release's config module and its configuration guide.",
+              links=[{"rel": "cites", "to": s1}, {"rel": "cites", "to": s2}])
+    st = _state(g, c["id"])
+    assert st["independent_lines"] == 2 and st["evidence"] == "external" and not st["unsupported"]
+
+
+def test_a_failed_outcome_on_a_succeeded_run_is_a_finding_not_a_fall(env):
+    g, inv = env
+    t = g.author("tester")
+    run = t.add("run", inv, kind="agent", name="spec cases", state="succeeded")["id"]
+    r = t.add("result", inv, kind="computed", outcome="failed",
+              summary="Of the 8 valid-input cases, parse_duration gets exactly 1 wrong: case 8.",
+              values={"valid_wrong": {"k": 1, "n": 8}}, links=[{"rel": "produced", "from": run}])["id"]
+    assert not D.fallen(g.snapshot(), r)
+    c = t.add("claim", inv, statement="parse_duration gets exactly 1 of the 8 valid-input cases wrong.",
+              scope="SPEC.md's 8 valid-input cases on the current diff.",
+              measurements=[{"name": "valid wrong", "result": r, "key": "valid_wrong", "k": 1, "n": 8}],
+              links=[{"rel": "derived_from", "to": r}])
+    st = _state(g, c["id"])
+    assert st["independent_lines"] == 1 and not st["unsupported"] and "UNSUPPORTED" not in st["sentence"]
+
+
+def test_a_failed_outcome_still_falls_when_the_run_failed_or_there_is_none(env):
+    g, inv = env
+    t = g.author("tester")
+    bad = t.add("run", inv, kind="agent", name="spec cases", state="failed")["id"]
+    r1 = t.add("result", inv, kind="computed", outcome="failed", summary="The harness crashed before the cases ran.",
+               values={"n_cases": 12}, links=[{"rel": "produced", "from": bad}])["id"]
+    r2 = t.add("result", inv, kind="computed", outcome="failed", summary="pytest could not import the module.",
+               values={"n_collected": 0})["id"]
+    s = g.snapshot()
+    assert D.fallen(s, r1) and D.fallen(s, r2)