Various utils.

1 files changed, 74 insertions, 0 deletions

diff --git a/src/eval_tune/compare_ecm_shareddepth.py b/src/eval_tune/compare_ecm_shareddepth.py
new file mode 100644
index 0000000..2194f02
--- /dev/null
+++ b/src/eval_tune/compare_ecm_shareddepth.py
@@ -0,0 +1,74 @@
+#!/usr/bin/env python3
+"""Pair up per-problem node counts at the DEEPEST depth BOTH binaries
+completed within their sn=N node budget (not the final/achieved depth,
+which is too coarse -- see compare_ecm_depth.py). Continuous signal,
+still fully bounded since we only ever compare depths both sides
+actually finished.
+"""
+import re
+import statistics
+import sys
+
+DEPTH_LINE_RE = re.compile(r"^\s*(\d+)[+]?\s+(\S+)\s+[\d:.]+\s+(\d+)\s")
+ROOT_POS_RE = re.compile(r"^The root position is:")
+
+
+def parse(path):
+    """Return list of {depth: node_count} dicts, one per problem, in order."""
+    problems = []
+    cur = {}
+    with open(path) as f:
+        for line in f:
+            if ROOT_POS_RE.match(line):
+                if cur:
+                    problems.append(cur)
+                cur = {}
+                continue
+            m = DEPTH_LINE_RE.match(line)
+            if m:
+                depth = int(m.group(1))
+                nodes = int(m.group(3))
+                cur[depth] = nodes  # later (deeper/re-searched) lines overwrite
+    if cur:
+        problems.append(cur)
+    return problems
+
+
+def main():
+    base_path, cand_path = sys.argv[1], sys.argv[2]
+    base = parse(base_path)
+    cand = parse(cand_path)
+    n = min(len(base), len(cand))
+
+    ratios = []
+    wins = losses = ties = 0
+    skipped = 0
+
+    for i in range(n):
+        b, c = base[i], cand[i]
+        shared_depths = set(b) & set(c)
+        if not shared_depths:
+            skipped += 1
+            continue
+        d = max(shared_depths)
+        bn, cn = b[d], c[d]
+        if bn == 0:
+            skipped += 1
+            continue
+        ratio = cn / bn
+        ratios.append(ratio)
+        if cn < bn: wins += 1
+        elif cn > bn: losses += 1
+        else: ties += 1
+
+    print(f"Paired problems: {n}  (usable: {len(ratios)}, skipped: {skipped})")
+    if ratios:
+        print(f"  median candidate/baseline ratio at shared depth: {statistics.median(ratios):.4f}")
+        print(f"  geomean ratio: {statistics.geometric_mean(ratios):.4f}")
+        print(f"  wins (candidate fewer nodes): {wins}")
+        print(f"  losses (candidate more nodes): {losses}")
+        print(f"  ties: {ties}")
+
+
+if __name__ == "__main__":
+    main()