Update app.py

app.py

@@ -1,14 +1,104 @@
-import socket
-
-def test_dns_resolution():
-    try:
-        # Telegram API hostname
-        host = 'api.telegram.org'
-        print(f"Resolving DNS for {host}")
-        ip = socket.gethostbyname(host)
-        print(f"DNS resolution successful: {ip}")
-    except socket.gaierror as e:
-        print(f"DNS resolution failed: {e}")
-
-if __name__ == "__main__":
-    test_dns_resolution()
+import numpy as np
+import pandas as pd
+
+from sentence_transformers import SentenceTransformer
+
+# from sklearn.decomposition import PCA
+from sklearn.metrics import DistanceMetric
+
+# import matplotlib.pyplot as plt
+# import matplotlib as mpl
+import warnings
+warnings.filterwarnings("ignore")
+import gradio as gr
+# from transformers.utils.hub import move_cache
+#
+# move_cache()
+# from mailersend import emails
+
+import asyncio
+from telegram import Bot
+
+token = '7370905765:AAHvmlw68cW7RxWzsJE1yxTzgf3xQFiokDo'
+chat_id = '7431171535'
+pd.set_option('display.max_colwidth', None)
+df_resume = pd.read_csv('QA.csv',encoding='latin-1')
+# df_resume['role'][df_resume['role'].iloc[-1] == df_resume['role']] = "Other" # relabel random role as "other"
+print(df_resume.head())
+model = SentenceTransformer("all-MiniLM-L6-v2")
+# file_name = 'questions.txt'
+
+async def send_telegram_message(token, chat_id, message):
+    bot = Bot(token=token)
+    await bot.send_message(chat_id=chat_id, text=message)
+    print("Message sent successfully.")
+
+
+def savevectorstore():
+    # store embed vectors
+    embedding_arr = model.encode(df_resume['question'])
+    print(embedding_arr.shape)
+    np.save('embeddings.npy', embedding_arr)
+# savevectorstore()
+# load embed vectors
+def rag_chain(question,name):
+    embedding_arr = np.load('embeddings.npy')
+    # print(embedding_arr.shape)
+
+    # pca = PCA(n_components=2).fit(embedding_arr)
+    # print(pca.explained_variance_ratio_)
+
+    query = question
+
+    query_embedding = model.encode(query)
+
+    dist = DistanceMetric.get_metric('euclidean') # other distances: manhattan, chebyshev
+
+    # compute pair wise distances between query embedding and all resume embeddings
+    dist_arr = dist.pairwise(embedding_arr, query_embedding.reshape(1, -1)).flatten()
+    # sort results
+    idist_arr_sorted = np.argsort(dist_arr)
+
+    # print(df_resume['name'].iloc[idist_arr_sorted[:1]].to_string(index=False))
+    # print(df_resume['name'].iloc[idist_arr_sorted[:1]])
+
+    que = f"**Most relevant question.**<br><br>{str(df_resume['question'].iloc[idist_arr_sorted[:1]].to_string(index=False)).replace('||','<br>')}"
+    # que = df_resume['question'].iloc[idist_arr_sorted[:1]]
+    ans = f"**Answer ideas.**<br><br>{str(df_resume['answer'].iloc[idist_arr_sorted[:1]].to_string(index=False)).replace('||','<br>')}"
+    # ans = df_resume['answer'].iloc[idist_arr_sorted[:1]]
+    # profit = df_resume['profit'].iloc[idist_arr_sorted[:1]].to_string(index=False)
+    # product = df_resume['product'].iloc[idist_arr_sorted[:1]].to_string(index=False)
+#     que1 = que
+#     ans1 = ans
+#     print(que1)
+#     print(ans1)
+#     que2= f'''Most relevant question
+#
+#
+# {que1}'''
+#     ans2= f'''Answer ideas
+#
+#
+# {ans1}'''
+#     print(que2)
+#     print(ans2)
+    # with open('questions.txt', 'a',encoding='utf-8') as file:
+    #     file.write(f"\n\n{question}\n\n{country}\n\n{whatsapp}\n\n{name}||{sales}||{profit}||{product}")
+    message = f"{name}\n\n{question}\n\n{ans}"
+    asyncio.run(send_telegram_message(token, chat_id, message))
+    return que,ans
+# rag_chain('I am very hungry.')
+
+desc = "This is an awesome ML App.  I'm really excited to show you"
+long_desc = "如果我没有回答你的问题，把问题发给Eric吧。"
+search_interface = gr.Interface(
+    fn=rag_chain,
+    inputs=[gr.Textbox(label="Question"),gr.Textbox(label="Name")],
+    outputs=[gr.Markdown(label="Most relevant question"),gr.Markdown(label="Answer ideas")],
+    title="Ask Eric",
+    description="Hi,我是数字分身，欢迎提问！",
+    # theme=gr.themes.Glass
+    article=long_desc
+)
+
+search_interface.launch(share=True,debug=True)