move requirements and app to root, rewrite basic app

.github/workflows/hugging_face.yml

@@ -18,5 +18,5 @@ jobs:
       - name: Push to hub
         env:
           HF_TOKEN: ${{ secrets.HF_TOKEN }}
-        run: git push https://AIhackathons:$HF_TOKEN@huggingface.co/spaces/AIhackathons/docverifyrag main
+        run: git push https://salgadev:$HF_TOKEN@huggingface.co/spaces/salgadev/docverifyrag main
 

app.py

@@ -1,219 +1,22 @@
-import time
-import streamlit as st
-from PyPDF2 import PdfReader
-from langchain.text_splitter import CharacterTextSplitter
-from langchain.embeddings import OpenAIEmbeddings, HuggingFaceInstructEmbeddings
-from langchain.vectorstores import FAISS
-from langchain.chat_models import ChatOpenAI
-from langchain.memory import ConversationBufferMemory
-from langchain.chains import ConversationalRetrievalChain
-import os
-import pickle
-from datetime import datetime
-from backend.generate_metadata import generate_metadata, ingest
-
-
-css = '''
-<style>
-.chat-message {
-    padding: 1.5rem; border-radius: 0.5rem; margin-bottom: 1rem; display: flex
-}
-.chat-message.user {
-    background-color: #2b313e
-}
-.chat-message.bot {
-    background-color: #475063
-}
-.chat-message .avatar {
-  width: 20%;
-}
-.chat-message .avatar img {
-  max-width: 78px;
-  max-height: 78px;
-  border-radius: 50%;
-  object-fit: cover;
-}
-.chat-message .message {
-  width: 80%;
-  padding: 0 1.5rem;
-  color: #fff;
-}
-'''
-bot_template = '''
-<div class="chat-message bot">
-    <div class="avatar">
-        <img src="https://i.ibb.co/cN0nmSj/Screenshot-2023-05-28-at-02-37-21.png" 
-        style="max-height: 78px; max-width: 78px; border-radius: 50%; object-fit: cover;">
-    </div>
-    <div class="message">{{MSG}}</div>
-</div>
-'''
-user_template = '''
-<div class="chat-message user">
-    <div class="avatar">
-        <img src="https://i.ibb.co/rdZC7LZ/Photo-logo-1.png">
-    </div>    
-    <div class="message">{{MSG}}</div>
-</div>
-'''
-
-
-def get_pdf_text(pdf_docs):
-    text = ""
-    for pdf in pdf_docs:
-        pdf_reader = PdfReader(pdf)
-        for page in pdf_reader.pages:
-            text += page.extract_text()
-    return text
-
-
-def get_text_chunks(text):
-    text_splitter = CharacterTextSplitter(
-        separator="\n",
-        chunk_size=1000,
-        chunk_overlap=200,
-        length_function=len
-    )
-    chunks = text_splitter.split_text(text)
-    return chunks
-
-
-def get_vectorstore(text_chunks):
-    embeddings = OpenAIEmbeddings()
-    # embeddings = HuggingFaceInstructEmbeddings(model_name="hkunlp/instructor-xl")
-    vectorstore = FAISS.from_texts(texts=text_chunks, embedding=embeddings)
-    return vectorstore
-
-
-def get_conversation_chain(vectorstore):
-    llm = ChatOpenAI()
-    # llm = HuggingFaceHub(repo_id="google/flan-t5-xxl", model_kwargs={"temperature":0.5, "max_length":512})
-
-    memory = ConversationBufferMemory(
-        memory_key='chat_history', return_messages=True)
-    conversation_chain = ConversationalRetrievalChain.from_llm(
-        llm=llm,
-        retriever=vectorstore.as_retriever(),
-        memory=memory
-    )
-    return conversation_chain
-
-
-def handle_userinput(user_question):
-    response = st.session_state.conversation({'question': user_question})
-    st.session_state.chat_history = response['chat_history']
-
-    for i, message in enumerate(st.session_state.chat_history):
-        # Display user message
-        if i % 2 == 0:
-            st.write(user_template.replace("{{MSG}}", message.content), unsafe_allow_html=True)
-        else:
-            print(message)
-            # Display AI response
-            st.write(bot_template.replace("{{MSG}}", message.content), unsafe_allow_html=True)
-
-            # THIS DOESNT WORK, SOMEONE PLS FIX
-            # Display source document information if available in the message
-            if hasattr(message, 'source') and message.source:
-                st.write(f"Source Document: {message.source}", unsafe_allow_html=True)
-
-
-def safe_vec_store():
-    # USE VECTARA INSTEAD
-    os.makedirs('vectorstore', exist_ok=True)
-    filename = 'vectores' + datetime.now().strftime('%Y%m%d%H%M') + '.pkl'
-    file_path = os.path.join('vectorstore', filename)
-    vector_store = st.session_state.vectorstore
-
-    # Serialize and save the entire FAISS object using pickle
-    with open(file_path, 'wb') as f:
-        pickle.dump(vector_store, f)
-
-
-def main():
-    st.set_page_config(page_title="Doc Verify RAG", page_icon=":hospital:")
-    st.write(css, unsafe_allow_html=True)
-    if "openai_api_key" not in st.session_state:
-        st.session_state.openai_api_key = False
-    if "openai_org" not in st.session_state:
-        st.session_state.openai_org = False
-    if "classify" not in st.session_state:
-        st.session_state.classify = False
-    def set_pw():
-        st.session_state.openai_api_key = True
-    st.subheader("Your documents")
-    # OPENAI_ORG_ID = st.text_input("OPENAI ORG ID:")
-    OPENAI_API_KEY = st.text_input("OPENAI API KEY:", type="password",
-                                   disabled=st.session_state.openai_api_key, on_change=set_pw)
-    if st.session_state.classify:
-        pdf_doc = st.file_uploader("Upload your PDFs here and click on 'Process'", accept_multiple_files=False)
-    else:
-        pdf_docs = st.file_uploader("Upload your PDFs here and click on 'Process'", accept_multiple_files=True)
-        filenames = [file.name for file in pdf_docs if file is not None]
-    if st.button("Process"):
-        with st.spinner("Processing"):
-            if st.session_state.classify:
-                # THE CLASSIFICATION APP
-                st.write("Classifying")
-                plain_text_doc = ingest(pdf_doc.name)
-                classification_result = generate_metadata(plain_text_doc)
-                st.write(classification_result)
-            else:
-                # NORMAL RAG
-                loaded_vec_store = None
-                for filename in filenames:
-                    if ".pkl" in filename:
-                        file_path = os.path.join('vectorstore', filename)
-                        with open(file_path, 'rb') as f:
-                            loaded_vec_store = pickle.load(f)
-                raw_text = get_pdf_text(pdf_docs)
-                text_chunks = get_text_chunks(raw_text)
-                vec = get_vectorstore(text_chunks)
-                if loaded_vec_store:
-                    vec.merge_from(loaded_vec_store)
-                    st.warning("loaded vectorstore")
-                if "vectorstore" in st.session_state:
-                    vec.merge_from(st.session_state.vectorstore)
-                    st.warning("merged to existing")
-                st.session_state.vectorstore = vec
-                st.session_state.conversation = get_conversation_chain(vec)
-        st.success("data loaded")
-
-
-    if "conversation" not in st.session_state:
-        st.session_state.conversation = None
-    if "chat_history" not in st.session_state:
-        st.session_state.chat_history = None
-
-    st.header("Doc Verify RAG :hospital:")
-    user_question = st.text_input("Ask a question about your documents:")
-    if user_question:
-        handle_userinput(user_question)
-    with st.sidebar:
-
-        st.subheader("Classification Instrucitons")
-        classifier_docs = st.file_uploader("Upload your instructions here and click on 'Process'", accept_multiple_files=True)
-        filenames = [file.name for file in classifier_docs if file is not None]
-
-        if st.button("Process Classification"):
-            st.session_state.classify = True
-            with st.spinner("Processing"):
-                st.warning("set classify")
-                time.sleep(3)
-
-
-        # Save and Load Embeddings
-        if st.button("Save Embeddings"):
-            if "vectorstore" in st.session_state:
-                safe_vec_store()
-                # st.session_state.vectorstore.save_local("faiss_index")
-                st.sidebar.success("saved")
-            else:
-                st.sidebar.warning("No embeddings to save. Please process documents first.")
-
-        if st.button("Load Embeddings"):
-            st.warning("this function is not in use, just upload the vectorstore")
-
-
-if __name__ == '__main__':
-    main()
+import streamlit as st
+import io
+import tempfile
+
+from scripts import generate_metadata, ingest
+
+
+st.title('PDF to Text Converter')
+st.write('This app converts a PDF file to plain text.')
+
+uploaded_file = st.file_uploader("Choose a PDF file", type=["pdf","txt"])
+
+if uploaded_file is not None:
+    try:
+        file_ext = uploaded_file.name.split('.')[-1].lower()
+        pdf_file = io.BytesIO(uploaded_file.read())
+        docs = ingest(pdf_file, file_ext)
+        metadata = generate_metadata(docs)
+        st.write('## Converted Text')
+        st.write(metadata)
+    except Exception as e:
+        st.error(f'Error: {e}')

backend/generate_metadata.py

@@ -1,102 +0,0 @@
-import os
-import io
-import argparse
-import json
-import openai
-import sys
-from dotenv import load_dotenv
-from langchain_community.document_loaders import TextLoader
-from langchain_community.document_loaders import UnstructuredPDFLoader
-from langchain_community.embeddings.fake import FakeEmbeddings
-from langchain_text_splitters import RecursiveCharacterTextSplitter
-
-load_dotenv()
-
-
-import io
-
-def ingest(file_obj, file_ext='pdf'):
-    if file_ext == 'pdf':
-        loader = UnstructuredPDFLoader(file_obj)
-    elif file_ext == 'txt':
-        loader = TextLoader(file_obj)
-    else:
-        raise NotImplementedError('Only .txt or .pdf files are supported')
-
-    # transform locally
-    documents = loader.load()
-    text_splitter = RecursiveCharacterTextSplitter(chunk_size=1000, chunk_overlap=0,
-    separators=[
-        "\n\n",
-        "\n",
-        " ",
-        ",",
-        "\uff0c",  # Fullwidth comma
-        "\u3001",  # Ideographic comma
-        "\uff0e",  # Fullwidth full stop
-        # "\u200B",  # Zero-width space (Asian languages)
-        # "\u3002",  # Ideographic full stop (Asian languages)
-        "",
-    ])
-    docs = text_splitter.split_documents(documents)
-
-    return docs
-
-
-def generate_metadata(docs):
-    prompt_template = """
-    BimDiscipline = ['plumbing', 'network', 'heating', 'electrical', 'ventilation', 'architecture']
-
-    You are a helpful assistant that understands BIM documents and engineering disciplines. Your answer should be in JSON format and only include the filename, a short description, and the engineering discipline the document belongs to, distinguishing between {[d.value for d in BimDiscipline]} based on the given document."
-
-    Analyze the provided document, which could be in either German or English. Extract the filename, its description, and infer the engineering discipline it belongs to. Document:
-    context="
-    """     
-    # plain text     
-    filepath = [doc.metadata for doc in docs][0]['source']
-    context = "".join(
-        [doc.page_content.replace('\n\n','').replace('..','') for doc in docs])
-
-    prompt = f'{prompt_template}{context}"\nFilepath:{filepath}'
-
-    #print(prompt)
-    
-    # Create client
-    client = openai.OpenAI(
-        base_url="https://api.together.xyz/v1",
-        api_key=os.environ["TOGETHER_API_KEY"],
-        #api_key=userdata.get('TOGETHER_API_KEY'),    
-    )
-
-    # Call the LLM with the JSON schema
-    chat_completion = client.chat.completions.create(
-        model="mistralai/Mixtral-8x7B-Instruct-v0.1",        
-        messages=[
-            {
-                "role": "system",
-                "content": f"You are a helpful assistant that responsds in JSON format"                
-            },
-            {
-                "role": "user",
-                "content": prompt                                
-            }
-        ]
-    )
-
-    return json.loads(chat_completion.choices[0].message.content)    
-
-
-if __name__ == "__main__":
-    parser = argparse.ArgumentParser(description="Generate metadata for a BIM document")
-    parser.add_argument("document", metavar="FILEPATH", type=str,
-                        help="Path to the BIM document")
-
-    args = parser.parse_args()
-
-    if not os.path.exists(args.document) or not os.path.isfile(args.document):
-        print("File '{}' not found or not accessible.".format(args.document))
-        sys.exit(-1)
-
-    docs = ingest(args.document)
-    metadata = generate_metadata(docs)
-    print(metadata)