Handle self-edges and multi-edges

Also clarify that we're computing log(n)

strider/query_planner.py

@@ -247,41 +247,44 @@ def get_next_qedge(qgraph):
 def get_pinnedness(qgraph, qnode_id):
     """Get pinnedness of each node."""
     adjacency_mat = get_adjacency_matrix(qgraph)
-    return -compute_expected_n(
+    num_ids = get_num_ids(qgraph)
+    return -compute_log_expected_n(
         adjacency_mat,
+        num_ids,
         qnode_id,
     )
 
 
-def compute_expected_n(adjacency_mat, qnode_id, last=None, level=0):
-    """Compute the expected number of unique knodes bound to the specified qnode in the final results."""
-    pinnedness = math.log(adjacency_mat[qnode_id][qnode_id])
+def compute_log_expected_n(adjacency_mat, num_ids, qnode_id, last=None, level=0):
+    """Compute the log of the expected number of unique knodes bound to the specified qnode in the final results."""
+    log_expected_n = math.log(num_ids[qnode_id])
     if level < 10:
-        for neighbor in adjacency_mat[qnode_id]:
-            if neighbor in (qnode_id, last):
+        for neighbor, num_edges in adjacency_mat[qnode_id].items():
+            if neighbor == last:
                 continue
-            pinnedness += min(max(compute_expected_n(
+            # ignore contributions >0 - edges should only _further_ constrain n
+            log_expected_n += num_edges * min(max(compute_log_expected_n(
                 adjacency_mat,
+                num_ids,
                 neighbor,
                 qnode_id,
                 level + 1,
             ), 0) + math.log(R / N), 0)
-    return pinnedness
+    return log_expected_n
 
 
 def get_adjacency_matrix(qgraph):
     """Get adjacency matrix."""
-    A = defaultdict(lambda: defaultdict(bool))
-    for qnode_id, qnode in qgraph["nodes"].items():
-        ids = qnode.get("ids")
-        if ids is None:
-            num_ids = N
-        elif isinstance(ids, list):
-            num_ids = len(ids)
-        else:
-            num_ids = ids
-        A[qnode_id][qnode_id] = num_ids
+    A = defaultdict(lambda: defaultdict(int))
     for qedge in qgraph["edges"].values():
-        A[qedge["subject"]][qedge["object"]] = True
-        A[qedge["object"]][qedge["subject"]] = True
+        A[qedge["subject"]][qedge["object"]] += 1
+        A[qedge["object"]][qedge["subject"]] += 1
     return A
+
+
+def get_num_ids(qgraph):
+    """Get the number of ids for each node."""
+    return {
+        qnode_id: qnode["ids"]
+        for qnode_id, qnode in qgraph["nodes"].items()
+    }