Upload 4 files

.env

@@ -0,0 +1 @@
+GOOGLE_API_KEY = "AIzaSyA1_dEl52RWxEH197REEcDrYFvVlE5GCJU" # gemini

app.py

@@ -0,0 +1,29 @@
+import streamlit as st
+import main
+
+st.set_page_config(
+    page_title="Historical Event Explainer",
+    page_icon="🏛️",
+    layout="wide"
+)
+
+st.title("ChronoQuery: Explore History with AI 🕰️")
+st.markdown("Unlock the past — ask about any historical event and get instant, AI-powered insights.")
+
+with st.form(key="query_form"):
+    user_query = st.text_input("Enter a historical event or question:")
+    submit_button = st.form_submit_button(label="Submit")
+
+if submit_button and user_query:
+    with st.spinner("⏳ Fetching historical insights..."):
+        try:
+            result = main.final_chain.invoke(user_query)
+            st.write(result)
+        except Exception as e:
+            st.error("⚠️ Something went wrong. Please try again.")
+            st.exception(e) 
+
+
+
+
+

main.py

@@ -0,0 +1,96 @@
+from langchain_community.retrievers import WikipediaRetriever
+from langchain_core.prompts import PromptTemplate
+from langchain_core.output_parsers import StrOutputParser
+from langchain_google_genai import ChatGoogleGenerativeAI
+from langchain_community.tools import DuckDuckGoSearchRun
+from langchain_core.tools import tool
+from langchain.schema.runnable import RunnableParallel, RunnablePassthrough
+from dotenv import load_dotenv
+
+# loading the environment key
+load_dotenv()
+
+# custom tool
+@tool
+def concat_data(docs : list , search_results: str) -> str:
+    """
+    takes all the page content from the docs and search_results. then combines them and returns them
+    """
+    data = ""
+    for doc in docs:
+        data += "\n\n" + doc.page_content
+
+    data += "\n\n" + search_results
+
+    return data
+
+# model
+model = ChatGoogleGenerativeAI(model="gemini-1.5-flash")
+
+# prompt template
+main_template = PromptTemplate(
+    template = """
+    You are a historical assistant. Based on the following context, answer the user's question or just summarize the topic if it's not a question. Do not use * in the output. \n\n
+    Context:
+    {context}
+    \n\n\n
+    user's question:
+    {question}
+
+    """,
+    input_variables = ["context", "question"]
+)
+
+wiki_template = PromptTemplate(
+    template="""
+You are an expert at identifying the core topic of a user's historical question.
+Your task is to extract and return only the **specific topic or event** the query refers to — no extra explanation.
+
+Examples:
+- Query: who killed maharana pratap?  
+  Output: maharana pratap
+
+- Query: when did india china war happen?  
+  Output: india china war
+
+- Query: who caused world war 2?  
+  Output: world war 2
+
+Now do the same for the following:
+Query: {query}
+Output:""",
+    input_variables=["query"]
+)
+
+# output parser
+parser = StrOutputParser()
+
+# web search tool (built in tool)
+search_tool = DuckDuckGoSearchRun()
+
+# wikipedia retriever (built in tool)
+retriever = WikipediaRetriever(top_k_results=4, lang="en")
+
+# chains
+data_chain = RunnableParallel({
+    'docs': wiki_template | model | parser | retriever,
+    'search_results': search_tool
+})
+
+question_chain = RunnablePassthrough()
+
+mid_chain = RunnableParallel({
+    'context' : data_chain | concat_data,
+    'question' : question_chain
+})
+
+final_chain = mid_chain | main_template | model | parser
+
+# query = "wars between china and india"
+# output = final_chain.invoke(query)
+# print(output)
+
+
+
+
+

requirements.txt

@@ -0,0 +1,7 @@
+langchain-community 
+langchain-google-genai 
+google-ai-generativelanguage 
+faiss-cpu 
+wikipedia 
+duckduckgo-search
+streamlit