Update app.py

app.py

@@ -10,20 +10,20 @@ llm = HuggingFaceEndpoint(
     repo_id="mistralai/Mistral-7B-Instruct-v0.3",
     task="text-generation",
     max_new_tokens=128,
-    temperature=0.7,
+    temperature=0.5,
     do_sample=False,
 )
 
 template_classify = '''
-You are a topic detector bot. Your task is to determine the main topic of given text phrase.
+You are an irrelevant content detector bot.
+In social media, there are lots of bot accounts and they produce irrelevant contents on company hashtags.
+You will detect it as Relevant or Irrelevant.
 
-Answer general main topic not specific words.
-Your answer does not contain specific information from given text.
-Answer just one general main topic. Do not answer two or more topic.
-Answer shortly with two or three word phrase. Do not answer with long sentence.
-If you do not know the topic just answer as General.
+Company name: {COMPANY_NAME}
+Company sector: {COMPANY_SECTOR}
+About Company: {ABOUT_COMPANY}
 
-What is the main topic of given text?:
+Detect following text as RELEVANT OR IRRELEVANT:
 
 <text>  
 {TEXT}  
@@ -32,47 +32,26 @@ What is the main topic of given text?:
 convert it to json format using 'Answer' as key and return it.
 Your final response MUST contain only the response, no other text.
 Example:
-{{"Answer":["General"]}}
+{{"Answer":["RELEVANT"]}}
 '''
 
-"""
-template_json = '''
-Your task is to read the following text, convert it to json format using 'Answer' as key and return it.
-<text>
-{RESPONSE}
-</text>
-
-Your final response MUST contain only the response, no other text.
-Example:
-{{"Answer":["General"]}}
-'''
-"""
-
 json_output_parser = JsonOutputParser()
 
 # Define the classify_text function
-def classify_text(text):
+def classify_text(text, company_name_input, company_sector_input, about_company_input):
     global llm
 
     start = time.time()
-    lang = detect(text)
-
-    language_map = {"tr": "turkish",
-                    "en": "english",
-                    "ar": "arabic",
-                    "es": "spanish",
-                    "it": "italian",
-                    }
-    try:
-        lang = language_map[lang]
-    except:
-        lang = "en"
 
     prompt_classify = PromptTemplate(
         template=template_classify,
-        input_variables=["LANG", "TEXT"]
+        input_variables=["TEXT", "COMPANY_NAME", "COMPANY_SECTOR", "ABOUT_COMPANY"]
     )
-    formatted_prompt = prompt_classify.format(TEXT=text, LANG=lang)
+    formatted_prompt = prompt_classify.format(TEXT=text,
+                                              COMPANY_NAME=company_name_input
+                                              COMPANY_SECTOR=company_sector_input,
+                                              ABOUT_COMPANY=about_company_input
+                                             )
     classify = llm.invoke(formatted_prompt)
 
     '''
@@ -97,12 +76,15 @@ def gradio_app(text):
 
 def create_gradio_interface():
     with gr.Blocks() as iface:
+        company_name_input = gr.Textbox(label="Enter Company Name")
+        company_sector_input = gr.Textbox(label="Enter Company Sector")
+        about_company_input = gr.Textbox(label="Enter Information About Company")
         text_input = gr.Textbox(label="Text")
-        output_text = gr.Textbox(label="Detected Topics")
+        output_text = gr.Textbox(label="Result")
         time_taken = gr.Textbox(label="Time Taken (seconds)")
-        submit_btn = gr.Button("Detect topic")
+        submit_btn = gr.Button("Detect")
 
-        submit_btn.click(fn=classify_text, inputs=text_input, outputs=[output_text, time_taken])
+        submit_btn.click(fn=classify_text, inputs=[company_name_input, company_sector_input, about_company_input, text_input], outputs=[output_text, time_taken])
 
     iface.launch()