From b757ec6149bf567ada1a84f7f11cdbb5452bd37e Mon Sep 17 00:00:00 2001
From: zeeoale <zeooale@gmail.com>
Date: Sat, 14 Feb 2026 00:43:52 +0100
Subject: [PATCH] Add JSON name to input

---
 PromptTagsExtractorNode.py | 71 ++++++++++++++++++++++++++++++++++++--
 1 file changed, 68 insertions(+), 3 deletions(-)

diff --git a/PromptTagsExtractorNode.py b/PromptTagsExtractorNode.py
index 4704f5c..3dc53dd 100644
--- a/PromptTagsExtractorNode.py
+++ b/PromptTagsExtractorNode.py
@@ -1,4 +1,5 @@
 import re
+import os
 from collections import Counter
 
 
@@ -19,6 +20,14 @@ class PromptTagsExtractorNode:
                 # whitelist manuale, es: "nun, veil, latex, ritual"
                 "keep_csv": ("STRING", {"default": ""}),
                 "min_len": ("INT", {"default": 4, "min": 3, "max": 12}),
+                # ✅ nuovo: nome world/json per includere tag “identitari”
+                # esempi:
+                # "PFN_Tokyo_Street_Style_HipHop.json"
+                # "PFN_Tokyo_Dance_Crew_Studio"
+                "json_name": ("STRING", {"default": ""}),
+                # se vuoi controllare quanto “forzare” i token del nome
+                "force_name_tokens": ("BOOLEAN", {"default": True}),
+                "name_tokens_max": ("INT", {"default": 12, "min": 3, "max": 24}),
             }
         }
 
@@ -39,7 +48,7 @@ class PromptTagsExtractorNode:
 
     _PROMPT_GARBAGE = set("""
     masterpiece best quality amazing 4k ultra detailed absurdres newest scenery depth field volumetric lighting
-    high resolution aesthetic cinematic detailed ultra highres sharp focus 
+    high resolution aesthetic cinematic detailed ultra highres sharp focus
     """.split())
 
     # -------------------------
@@ -56,6 +65,46 @@ class PromptTagsExtractorNode:
         text = re.sub(r"\s+", " ", text).strip().lower()
         return text
 
+    def _tokenize_json_name(self, json_name: str, min_len: int, max_tokens: int):
+        if not json_name:
+            return []
+
+        # se arriva un path o un filename, prendiamo solo il basename
+        base = os.path.basename(json_name.strip())
+
+        # togli estensione .json se presente
+        base = re.sub(r"\.json$", "", base, flags=re.IGNORECASE)
+
+        # normalizza separatori: underscore, dash, spazi
+        base = base.replace("-", " ").replace("_", " ")
+        base = re.sub(r"\s+", " ", base).strip().lower()
+
+        # tokenizza
+        raw = base.split()
+
+        out = []
+        for w in raw:
+            if len(w) < min_len:
+                continue
+            # filtra roba troppo “tecnica”
+            if w in {"json"}:
+                continue
+            # se vuoi tenere "pfn" come tag identitario, NON filtrarlo.
+            # Se invece non lo vuoi, aggiungilo qui:
+            # if w == "pfn": continue
+
+            out.append(w)
+
+        # de-dup preservando ordine
+        seen = set()
+        uniq = []
+        for w in out:
+            if w not in seen:
+                seen.add(w)
+                uniq.append(w)
+
+        return uniq[:max_tokens]
+
     def _extract_tags(self, prompt: str, top_k: int, min_len: int, keep_csv: str):
         keep_set = set(
             w.strip().lower()
@@ -89,8 +138,24 @@ class PromptTagsExtractorNode:
     # Nodo
     # -------------------------
 
-    def run(self, prompt, top_k, keep_csv="", min_len=4):
-        tags = self._extract_tags(prompt, top_k, min_len, keep_csv)
+    def run(self, prompt, top_k, keep_csv="", min_len=4, json_name="", force_name_tokens=True, name_tokens_max=12):
+        base_tags = self._extract_tags(prompt, top_k, min_len, keep_csv)
+
+        name_tokens = self._tokenize_json_name(json_name, min_len=min_len, max_tokens=name_tokens_max)
+
+        if name_tokens:
+            if force_name_tokens:
+                # ✅ forzali in testa (identità sempre presente)
+                tags = name_tokens + [t for t in base_tags if t not in set(name_tokens)]
+            else:
+                # alternativa: aggiungili solo se non ci sono già
+                tags = base_tags[:]
+                for t in name_tokens:
+                    if t not in tags:
+                        tags.append(t)
+        else:
+            tags = base_tags
+
         tags_csv = ", ".join(tags)
         tags_space = " ".join(tags)
         return (tags_csv, tags_space)