Update app.py

app.py

@@ -6,38 +6,45 @@ For more information on `huggingface_hub` Inference API support, please check th
 """
 client = InferenceClient("HuggingFaceH4/zephyr-7b-beta")
 
+def respond(message, history, system_message, max_tokens, temperature, top_p):
+    import requests
+    from pdf.loader import PyPDFLoader
+    
+    URL = "https://www.esmo.org/content/download/6594/114963/1/ES-Cancer-de-Mama-Guia-para-Pacientes.pdf"
+    response = requests.get(URL)
+    open("ES-Cancer-de-Mama-Guia-para-Pacientes.pdf", "wb").write(response.content)
+    loader = PyPDFLoader("ES-Cancer-de-Mama-Guia-para-Pacientes.pdf")
+    documents = loader.load()
+    
+    text_splitter = RecursiveCharacterTextSplitter(chunk_size=1000, chunk_overlap=20)
+    all_splits = text_splitter.split_documents(documents)
+    
+    model_name = "sentence-transformers/paraphrase-multilingual-mpnet-base-v2"
+    model_kwargs = {"device": "cuda"}
+    embeddings = HuggingFaceEmbeddings(model_name=model_name, model_kwargs=model_kwargs)
+    
+    vectordb = Chroma.from_documents(documents=all_splits, embedding=embeddings, persist_directory="chroma_db")
+
+    query = message
+    docs = vectordb.similarity_search_with_score(query)
+    context = []
+    for doc, score in docs:
+        if score < 7:
+            doc_details = doc.to_json()['kwargs']
+            context.append(doc_details['page_content'])
+    if len(context) != 0:
+        messages = [
+            {"role": "user", "content": "Basándote en la siguiente información: " + "\n".join(context) + "\n Responde en castellano a la pregunta: " + query}]
+        prompt = pipeline.tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
+        outputs = pipeline(prompt, max_new_tokens=max_tokens, do_sample=True, temperature=temperature, top_k=50,
+                           top_p=top_p)
+        answer = outputs[0]["generated_text"]
+        return answer[answer.rfind("[/INST]") + 8:], docs
+    else:
+        return "No tengo información para responder a esta pregunta", docs
+
+
 
-def respond(
-    message,
-    history: list[tuple[str, str]],
-    system_message,
-    max_tokens,
-    temperature,
-    top_p,
-):
-    messages = [{"role": "system", "content": system_message}]
-
-    for val in history:
-        if val[0]:
-            messages.append({"role": "user", "content": val[0]})
-        if val[1]:
-            messages.append({"role": "assistant", "content": val[1]})
-
-    messages.append({"role": "user", "content": message})
-
-    response = ""
-
-    for message in client.chat_completion(
-        messages,
-        max_tokens=max_tokens,
-        stream=True,
-        temperature=temperature,
-        top_p=top_p,
-    ):
-        token = message.choices[0].delta.content
-
-        response += token
-        yield response
 
 """
 For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface