Upload 3 files

app.py

@@ -0,0 +1,153 @@
+import os
+import openai
+import gradio as gr
+#import nest_asyncio
+import time
+import asyncio
+#nest_asyncio.apply()
+
+from llama_index.embeddings.openai import OpenAIEmbedding
+from llama_index.core.node_parser import SentenceSplitter
+from llama_index.llms.openai import OpenAI
+from llama_index.core import Settings
+
+Settings.llm = OpenAI(model="gpt-3.5-turbo-0125")
+# change to Huggingface embedding model 
+Settings.embed_model = OpenAIEmbedding(model="text-embedding-ada-002")
+Settings.node_parser = SentenceSplitter(chunk_size=1024, chunk_overlap=128)
+Settings.num_output = 512
+Settings.context_window = 3900
+
+
+from llama_index.core import (
+    VectorStoreIndex,
+    StorageContext,
+    PromptTemplate,
+    load_index_from_storage
+)
+
+from llama_index.readers.file import PyMuPDFReader
+
+from theme import CustomTheme  
+
+system_prompt = (
+    "You are a helpful assistant in the Bavarian ministry of science and education. "
+)
+
+context = (
+    "Context information is below. \n"
+    "----------------------\n"
+    "{context_str}\n"
+    "----------------------\n"
+    "Given the context information and not prior knowledge, "
+    "If you don't know the answer, tell the user that you can't answer the question - DO NOT MAKE UP AN ANSWER. "
+    "Do not make up your own answers, refer only from the given information. "
+    "Your answers use correct grammar and your texting style is casual. "
+    "Always be friendly, always reply in German! "
+)
+
+prompt = (
+    "Context information is below. \n"
+    "----------------------\n"
+    "{context_str}\n"
+    "----------------------\n"
+    "Given the context information and not prior knowledge, "
+    "If you don't know the answer, tell the user that you can't answer the question - DO NOT MAKE UP AN ANSWER. "
+    "Do not make up your own answers, refer only from the given information. "
+    "Your answers use correct grammar and your texting style is casual. "
+    "Always be friendly, always reply in German! "
+
+)
+
+prompt_template = PromptTemplate(prompt)
+
+
+
+# check if storage already exists
+if not os.path.exists("./storage"):
+    # load the documents and create the index
+    #documents = SimpleDirectoryReader("data").load_data()
+    loader = PyMuPDFReader()
+    documents = loader.load(file_path="./data/Rahmenvereinbarung-2023-2027_ohne-Unterschrift.pdf")
+    index = VectorStoreIndex.from_documents(documents)
+    # store it for later
+    index.storage_context.persist()
+else:
+    # load the existing index
+    storage_context = StorageContext.from_defaults(persist_dir="./storage")
+    index = load_index_from_storage(storage_context)
+
+chat_engine = index.as_chat_engine(
+    chat_mode= "context", system_prompt=system_prompt, context_template=context)
+
+query_engine = index.as_query_engine(streaming=True)
+#query_engine = index.as_query_engine(similarity_top_k=5)
+query_engine.update_prompts(
+    {"response_synthesizer:text_qa_template": prompt_template}
+)
+
+
+default_text="Ich beantworte Fragen zur Rahmenvereinbarung Hochschulen 2023 - 2027 gemäß Art. 8 Abs. 1 BayHIG. Wie kann ich helfen?"
+
+bot_examples = [
+    "Was sind die 3 zentralen Themen im Text?",
+    "Erstelle jeweils eine Zusammenfassung zu den zentralen Themen",
+    "Welche Vereinbarungen wurden zwischen den Universitäten und dem Ministerium getroffen?",
+    "Wie unterscheiden sich die Vereinbarungen der Universitäten von den Vereinbarungen der Hochschulen für angewandte Wissenschaften?",
+    "Welche Maßnahmen sind zum Ausbau der Wissenschaftskommunikation vorgesehen?",
+]
+
+submit_button = gr.Button(
+    value="Ask me",
+    elem_classes=["ask-button"],
+)
+
+def response(message, history):
+    loop = asyncio.new_event_loop()
+    asyncio.set_event_loop(loop)
+    
+    histories = chat_engine.chat_history
+    answer = chat_engine.stream_chat(message, chat_history=histories)
+
+    output_text = ""
+    for token in answer.response_gen:
+        time.sleep(0.1)
+
+        output_text += token
+        yield output_text
+
+    #return str(answer)
+
+
+def main():
+    openai.api_key = os.environ["OPENAI_API_KEY"]
+    custom_theme = CustomTheme()
+
+    desc = "[Rahmenvereinbarung Hochschulen 2023 - 2027 gemäß Art. 8 Abs. 1 BayHIG](https://www.stmwk.bayern.de/download/22215_Rahmenvereinbarung_inkl_Unterschriften.pdf%C2%A0)"
+
+    # default_text noch einbauen
+    chatbot = gr.Chatbot(
+        layout='bubbles',
+        #height=600,
+        value=[[None, default_text]]
+    )
+
+    chat_interface = gr.ChatInterface(
+        fn=response,
+        retry_btn=None,
+        undo_btn=None,
+        title="MUC.DAI Chatbot",
+        submit_btn=submit_button,
+        clear_btn=None,
+        theme=custom_theme,
+        chatbot=chatbot,
+        description=desc,
+        css="style.css",
+        examples=bot_examples,
+    )
+
+    chat_interface.launch(inbrowser=True, debug=True)
+
+
+if __name__ == "__main__":
+    main()

style.css

@@ -0,0 +1,34 @@
+.gap.panel {
+  padding: 0 !important;
+}
+/*
+div {
+  background-image: url("https://mediapool.hm.edu/media/mucdai/keyvisuals_1/mucdai_keyvisual_16_portrait_m.jpg") !important;
+}
+*/
+
+
+div.gap > .stretch {
+  display: none !important;
+}
+div.gap.panel > div.gr-group {
+  position: absolute;
+  bottom: 0;
+}
+h1 {
+  font-size: 24px !important;
+}
+.ask-button {
+  background-color: var(--color-accent);
+  font-weight: bold;
+  letter-spacing: 0.1rem;
+}
+.title {
+  color: var(--block-label-text-color) !important;
+}
+.gallery-item {
+  color: var(--block-label-text-color) !important;
+}
+div.message-wrap {
+  margin-bottom: 32px !important;
+}

theme.py

@@ -0,0 +1,62 @@
+from gradio.themes.soft import Soft
+from gradio.themes.utils import fonts
+
+
+class CustomTheme(Soft):
+
+    def __init__(self):
+        super().__init__(
+            font=fonts.GoogleFont("Roboto")
+        )
+
+        white = "#FFFFFF"
+        purple = "#2F1009D"
+        red = "#FC5555"
+
+        primary = white
+        secondary = "#e6e6e6"
+        panel_color = red
+        accent = purple
+        accent_soft = "#49637a28"
+        
+        primary_dark = "#121212"
+        secondary_dark = "#242424"
+        panel_color_dark = red
+        accent_dark = purple
+        accent_soft_dark = "#101727"
+        text_color_dark = white
+
+        super().set(
+            # LIGHT MODE
+            body_background_fill=primary,
+            background_fill_secondary=primary,
+            panel_background_fill=panel_color,
+            border_color_primary=primary,
+            block_background_fill=secondary,
+            block_border_color=primary,
+            block_label_background_fill=primary,
+            input_background_fill="#DADFE6",
+            input_border_color=secondary,
+            button_secondary_background_fill=accent,
+            button_secondary_text_color=white,
+            color_accent_soft=accent_soft,
+            border_color_accent_subdued=white,
+
+            # DARK MODE
+            body_background_fill_dark=primary_dark,
+            background_fill_secondary_dark=secondary_dark,
+            panel_background_fill_dark=secondary_dark,
+            border_color_primary_dark=primary_dark,
+            block_background_fill_dark=secondary_dark,
+            block_border_color_dark=secondary_dark,
+            block_label_background_fill_dark=primary_dark,
+            block_label_text_color_dark=text_color_dark,
+            input_background_fill_dark=panel_color_dark,
+            input_border_color_dark=secondary_dark,
+            button_primary_background_fill_dark=accent_dark,
+            button_primary_text_color_dark=primary_dark,
+            color_accent_soft_dark=accent_soft_dark,
+            border_color_accent_subdued_dark=accent_soft_dark,
+
+            #block_radius="15px",
+        )