From ca3e460e033f0119395d7f385f36b844a2b3532f Mon Sep 17 00:00:00 2001
From: digital-garbage <133010426+digital-garbage@users.noreply.github.com>
Date: Thu, 24 Sep 2026 02:27:35 +0300
Subject: [PATCH] fix: Enhance reference finds a shortcut by any meaningful
 word of its content

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---
 conditioning.py               | 29 +++++++++++++++++++++++++++--
 tests/test_prompt_enhancer.py | 15 +++++++++++++++
 2 files changed, 42 insertions(+), 2 deletions(-)

diff --git a/conditioning.py b/conditioning.py
index b11b43f8..8eb3e8b3 100644
--- a/conditioning.py
+++ b/conditioning.py
@@ -3659,6 +3659,24 @@ Rules:
 #: The prompt enhancer's default instructions. Separate from the repair advisor above: that
 #: one is triggered by a rating and fixes a named fault, this one runs on every generation and
 #: only elaborates. Kept short so a user editing it can actually read it.
+# Filler that says nothing about which shortcut a prompt means. Without this, every shortcut
+# containing "the" or "with" would match every prompt.
+_ENHANCER_STOPWORDS = frozenset("""
+the and with from into onto over under for but not are was were been being has have had
+its his her hers their theirs them they she him you your yours our ours this that these
+those there here then than very just only also some any all each every both few more most
+other such own same too can will would could should may might must does did doing done
+while when where which who whom whose what why how about above after again against
+before below between during out off once further through until upon via per
+""".split())
+
+
+def _enhancer_vocab(text):
+    """Meaningful words (letters only, 3+ long, not filler) of `text`, lowercased."""
+    return {w for w in re.findall(r"[^\W\d_]{3,}", str(text or "").lower())
+            if w not in _ENHANCER_STOPWORDS}
+
+
 V2_PROMPT_ENHANCER_SYSTEM_PROMPT = """You expand short video prompts into detailed ones for a text-to-video model.
 
 Rules:
@@ -8988,7 +9006,9 @@ class FunPackVideoRefinerV2(FunPackVideoRefiner):
             lines += reps if len(reps) == 1 else [f"Variant {i + 1}: {r}" for i, r in enumerate(reps)]
             # triggers are already expanded away, so the content counts as a mention too
             return {"block": "\n".join(lines),
-                    "words": [sc.get("name", "")] + list(sc.get("triggers") or []) + reps}
+                    "words": [sc.get("name", "")] + list(sc.get("triggers") or []) + reps,
+                    # any meaningful word of its name or content, typed in the prompt, counts
+                    "vocab": _enhancer_vocab(" ".join([str(sc.get("name", ""))] + reps))}
 
         groups = []
         keys = [str(k) for k in (shortcut_keys or []) if str(k).strip()]
@@ -9027,7 +9047,8 @@ class FunPackVideoRefinerV2(FunPackVideoRefiner):
     def _v2_enhancer_reference(text, sources):
         """Reference appended after the prompt the enhancer rewrites. Per group: the entries
         the prompt mentions, or the whole group when it mentions none (the model may need
-        what the prompt does not name). Shortcuts are mentioned by name, trigger or content;
+        what the prompt does not name). Shortcuts are mentioned by name, trigger, content, or
+        any meaningful word of their name or content;
         lore entries by their keywords, constant entries joining any match."""
         low = str(text or "").lower()
 
@@ -9045,7 +9066,11 @@ class FunPackVideoRefinerV2(FunPackVideoRefiner):
             # whole words only: "rain" is not in "train", "sea" is not in "season"
             return re.search(r"(?<!\w)" + re.escape(word.lower()) + r"(?!\w)", low) is not None
 
+        typed = _enhancer_vocab(low)
+
         def mentioned(e):
+            if e.get("vocab") and typed & e["vocab"]:
+                return True
             if "keys" in e:
                 keys = e["keys"]
                 if isinstance(keys, str):
diff --git a/tests/test_prompt_enhancer.py b/tests/test_prompt_enhancer.py
index 051b56a4..3589a552 100644
--- a/tests/test_prompt_enhancer.py
+++ b/tests/test_prompt_enhancer.py
@@ -553,3 +553,18 @@ def test_names_and_keywords_match_whole_words_only(tmp_path, monkeypatch):
     assert "grey sea" not in ref                                      # "season" is not "sea"
     assert "old port" in ref                                          # entry opted out of whole words
     assert "black tower" in ref                                       # regex key
+
+
+def test_a_word_from_inside_a_shortcut_finds_it_even_when_another_matches(tmp_path):
+    import json as _json
+    pack = tmp_path / "library.json"
+    pack.write_text(_json.dumps({"shortcuts": {
+        "a": {"name": "Mira", "triggers": ["/mira"], "replacements": ["woman in a crimson trench coat"]},
+        "b": {"name": "Night", "triggers": ["/night"], "replacements": ["moonlit street"]},
+        "c": {"name": "Rain", "triggers": ["/rain"], "replacements": ["heavy rain"]}}}))
+    src = C.FunPackVideoRefinerV2._v2_enhancer_sources([], [str(pack)])
+    ref = C.FunPackVideoRefinerV2._v2_enhancer_reference("a figure in a trench coat at night", src)
+    assert "[Shortcut] Mira" in ref and "[Shortcut] Night" in ref      # "trench" + name "night"
+    assert "heavy rain" not in ref
+    # filler words never count: "with the" alone mentions nothing, so the whole file goes
+    assert "heavy rain" in C.FunPackVideoRefinerV2._v2_enhancer_reference("with the", src)