patryk025/Aidem-Media-DLL-Analysis
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Body normalisation: per-method similarity score + leaf delta

Browse Source 
Turns the dispatch axis from a binary changed/unchanged into a "how much" measure
of code change — the original goal. ams.normalize compares two body fingerprints
(the ordered leaf-call anchors) with difflib after collapsing consecutive-duplicate
anchors (a load-twice codegen artefact), yielding a 0-100 similarity and the exact
leaves that appeared/vanished.

Every dispatch `changed` entry now carries body={similarity, added, removed}, and the
block carries a summary={shared, identical, changed, mean_similarity}.

Golden pair (cross-compiler): 470 shared bodies, 131 identical, mean 66% similar;
Animo SHOW/HIDE/PAUSE/RESUME come out 100% despite MSVC6 vs MSVC8, LOAD 50% with the
swapped leaves spelled out.

- normalize.py: canonical / body_similarity / body_delta
- diff: _dispatch_diff enriches changed with body + adds summary
- render: METHOD BODIES shows %, leaf delta, summary line
- UI: similarity % + leaf delta + axis summary
- tests: 5 new -> 34/34

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
This commit is contained in:
Patryk Gensch
2026-05-31 13:23:15 +02:00
parent 27399a52b1
commit b0d3d22445
7 changed files with 198 additions and 18 deletions
Download Patch File Download Diff File Expand all files Collapse all files

36
ams/render.py
View File

@@ -82,11 +82,23 @@ def _dispatch_name(r: dict) -> str:
return r.get("name") or "id {0}".format(r.get("id"))
def _leaves(items: list, cap: int = 4) -> str:
shown = items[:cap]
extra = "…+{0}".format(len(items) - cap) if len(items) > cap else ""
return ", ".join(shown) + extra
def _section_dispatch(out: list[str], block: dict) -> None:
"""Method-body fingerprints (per owner+id). `calls` deltas are summarised by length so the
line stays readable; the full anchor lists live in the JSON."""
"""Method bodies (per owner+id), normalised. Each changed entry shows a similarity score and
the leaf-level delta (which calls appeared/vanished); a summary line gives the overall drift."""
out.append("")
out.append("{0:<16} {1}".format("METHOD BODIES", _counts(block)))
summ = block.get("summary")
head = "METHOD BODIES"
if summ:
head = "{0} (shared {1}, ~{2} changed, mean {3}% similar)".format(
"METHOD BODIES", summ["shared"], summ["changed"], summ["mean_similarity"])
out.append("{0}".format(head))
out.append("{0:<16} {1}".format("", _counts(block)))
owner_of = lambda r: r["owner"]
added = _group_by(block["added"], owner_of)
removed = _group_by(block["removed"], owner_of)
@@ -99,14 +111,18 @@ def _section_dispatch(out: list[str], block: dict) -> None:
for it in sorted(removed.get(owner, []), key=_dispatch_name):
out.append(" - {0}".format(_dispatch_name(it)))
for it in sorted(changed.get(owner, []), key=_dispatch_name):
ch = change_by_id[id(it)]["changes"]
ch = change_by_id[id(it)]
body = ch.get("body", {})
sim = body.get("similarity")
bits = []
if "impl" in ch:
bits.append("impl {0} -> {1}".format(ch["impl"][0], ch["impl"][1]))
if "calls" in ch:
a, b = ch["calls"]
bits.append("calls {0} -> {1}".format(len(a or []), len(b or [])))
out.append(" ~ {0:<22} {1}".format(_dispatch_name(it), "; ".join(bits)))
if body.get("added"):
bits.append("+[{0}]".format(_leaves(body["added"])))
if body.get("removed"):
bits.append("-[{0}]".format(_leaves(body["removed"])))
if not bits and "impl" in ch["changes"]:
bits.append("impl {0} -> {1}".format(*ch["changes"]["impl"]))
label = "{0} {1}%".format(_dispatch_name(it), sim) if sim is not None else _dispatch_name(it)
out.append(" ~ {0:<26} {1}".format(label, " ".join(bits)))
_EMPTY = {"added": [], "removed": [], "changed": []}