Add app.py

app.py

@@ -0,0 +1,56 @@
+import os
+import gradio as gr
+import cohere
+
+
+COHERE_KEY = os.getenv('COHERE_KEY')
+co = cohere.Client(COHERE_KEY)
+
+#list_history = [["question", "answer"], ["how", "how what..."]]
+def convert_history(list_history):
+    """
+    Applies the prompt.
+    Converts the chat history structure taken by Gradio to the structure suitable for Cohere.
+    """
+    chat_history = [
+        {"role": "SYSTEM", "text": open("prompt.md","r",encoding="UTF-8").read()}
+    ]
+    for item in list_history:
+        dict_chat = {"role": "USER", "text": item[0]}
+        chat_history.append(dict_chat)
+        dict_chat = {"role": "CHATBOT", "text": item[1]}
+        chat_history.append(dict_chat)
+    return chat_history
+        
+
+
+
+def reply(message:str, history:list):
+    """
+    Takes the input message of the user and chat history and streams the reply of the chatbot.
+    """
+    chat_history = convert_history(history)
+    response = co.chat_stream(
+        message=message,
+        chat_history=chat_history,
+        model="command-nightly",
+        temperature=0.25
+    )
+    text_so_far = ""
+    for event in response:
+        if event.event_type == 'text-generation':
+            text_so_far += event.text
+            yield text_so_far
+
+
+description = """
+Hello! 
+💬 Use the text box below to ask questions the Technorama 2024 conference!
+<nobr> 🗣️ Talk to me in English, Dutch, or French. </nobr>
+<nobr> 🔗 [This chatbot was made by Alfiya  Khabibullina](https://www.linkedin.com/in/alfiya-khabibullina-7b13131b8/) </nobr>
+"""
+
+gr.ChatInterface(reply,
+    title="Technorama Assistant",
+    description=description
+).launch()

scraping_agenda.py

@@ -1,30 +1,13 @@
 import playwright
+import playwright.sync_api
 from playwright.sync_api import sync_playwright
 import json
-import playwright.sync_api
+
 
 
 url = "https://techorama.be/agenda/"
 selector_links = "div.m-subject__container-inner.a-box-simple__inner-2 > a"
 
-with sync_playwright() as p:
-    browser = p.chromium.launch(headless=True)
-    page = browser.new_page()
-    page.goto(url)
-    page.wait_for_timeout(3000)
-    links_elements = page.query_selector_all(selector_links)
-    links = [element.get_property("href").json_value() for element in links_elements] # .get_attribute("href")
-    filepath = "links_tue.csv"
-    #with open(filepath, "w") as f:
-        #f.write("\n".join(links))
-    page.locator("button:has-text(\"wednesday\")").click()
-    links_elements = page.query_selector_all(selector_links)
-    links = [element.get_property("href").json_value() for element in links_elements] # .get_attribute("href")
-    filepath = "links_wed.csv"
-    #with open(filepath, "w") as f:
-        #f.write("\n".join(links))
-    
-    browser.close()
 
 #TODO
 #1 save the structure [{"time":"...", "activities":[{"name_of_activity":"...", "speaker_name": "...", general_topic": "topic name", "room_number":"..". "url":"...", "date":"...", "time":"time"}]}