Update app.py

app.py

@@ -11,6 +11,25 @@ from sentence_transformers import SentenceTransformer
 MODEL_BASE = "Qwen/Qwen2.5-1.5B-Instruct"
 DOCS_DIR = "docs"
 
+# ---------- Metni parçalara bölme (RAG chunking) ----------
+def split_text(text: str, chunk_size: int = 800, overlap: int = 100):
+    chunks = []
+    start = 0
+    length = len(text)
+
+    while start < length:
+        end = min(start + chunk_size, length)
+        chunk = text[start:end].strip()
+        if chunk:
+            chunks.append(chunk)
+        start = end - overlap
+        if start < 0:
+            start = 0
+
+    return chunks
+# ---------------------------------------------------------
+
+
 print("Loading tokenizer…")
 tokenizer = AutoTokenizer.from_pretrained(MODEL_BASE)
 
@@ -34,24 +53,11 @@ def load_docs():
             pass
     return texts
 
-def split_text(text: str, chunk_size: int = 800, overlap: int = 100):
-    chunks = []
-    start = 0
-    n = len(text)
-    while start < n:
-        end = start + chunk_size
-        chunk = text[start:end]
-        chunks.append(chunk.strip())
-        start = end - overlap
-    return chunks
-
 print("Loading RAG docs…")
-raw_docs = load_docs()
-
-# chunk'lanmış dokümanlar
-docs: list[str] = []
+raw_docs = load_docs()          # tam dokümanlar
+docs = []                       # chunk'lar buraya
 for d in raw_docs:
-    docs.extend(split_text(d))
+    docs.extend(split_text(d))  # her dokümanı parçalara böl
 
 if docs:
     embed_model = SentenceTransformer("sentence-transformers/all-MiniLM-L6-v2")
@@ -65,9 +71,10 @@ else:
     print("RAG: no docs found, context will be empty.")
 
 def retrieve_context(query: str, k: int = 3) -> str:
-    if index is None or embed_model is None:
+    if index is None or embed_model is None or not docs:
         return ""
     q_emb = embed_model.encode([query], convert_to_numpy=True)
+    k = min(k, len(docs))
     D, I = index.search(q_emb, k)
     parts = []
     for i in I[0]:
@@ -122,7 +129,7 @@ chatbot = gr.ChatInterface(
     respond,
     type="messages",
     additional_inputs=[
-        gr.Textbox("Merhaba", label="System"),
+        gr.Textbox("You are a scientific assistant.", label="System"),
         gr.Slider(1, 4096, 1024, step=1, label="Max tokens"),
         gr.Slider(0.1, 2.0, 0.7, step=0.1, label="Temperature"),
         gr.Slider(0.1, 1.0, 0.95, step=0.05, label="Top-p"),