Upload folder using huggingface_hub

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+Zigistry.pdf filter=lfs diff=lfs merge=lfs -text

.gitignore

@@ -0,0 +1,3 @@
+__pycache__/
+PERSIST_DIR/
+.gradio/

README.md

@@ -1,12 +1,12 @@
 ---
 title: Zigister
-emoji: ⚡
-colorFrom: red
-colorTo: blue
+emoji: 🦀
+colorFrom: yellow
+colorTo: yellow
 sdk: gradio
 sdk_version: 5.16.0
-app_file: app.py
-pinned: false
+app_file: main.py
+pinned: true
 ---
 
 Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference

Zigistry.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dbedcfc405d05e37a1ebcf60c2e167555c80ef88901bc835e5e97e2045c09d31
+size 517396

main.py

@@ -0,0 +1,90 @@
+"""
+Main script to run the LLM chatbot using the Zigistry framework
+
+This script performs the following tasks:
+    1. Perform pre-requisites check
+    2. Configure LLM and embedding model
+    3. Data ingestion
+    4. Query handling
+    5. Launch the chatbot interface
+"""
+
+# --- Importing required libraries ---
+from zigistry import constants
+from zigistry import pre_requisite
+import gradio as gr
+from llama_index.core import (
+    StorageContext,
+    load_index_from_storage,
+    VectorStoreIndex,
+    SimpleDirectoryReader,
+    ChatPromptTemplate,
+    Settings,
+)
+from llama_index.llms.huggingface_api import HuggingFaceInferenceAPI
+from llama_index.embeddings.huggingface import HuggingFaceEmbedding
+
+# -------------------------------------
+
+
+# --- Perform pre-requisites and pre-requisite check ---
+pre_requisite.performPreRequisites()
+
+if not pre_requisite.CorrectlyCompletedPreRequisites():
+    exit(1)
+# ------------------------------------------------------
+
+
+# --- Configure LLM and embedding model ---
+Settings.llm = HuggingFaceInferenceAPI(
+    model_name=constants.LLM,
+    tokenizer_name=constants.LLM,
+    context_window=3000,
+    token=constants.HF_TOKEN,
+    max_new_tokens=512,
+    generate_kwargs={"temperature": constants.TEMPERATURE},
+)
+
+Settings.embed_model = HuggingFaceEmbedding(model_name=constants.EMBEDDER)
+# -----------------------------------------
+
+
+def data_ingestion():
+    """
+    Ingest data from the input files and create an index
+    """
+    documents = SimpleDirectoryReader(input_files=constants.FILES).load_data()
+    index = VectorStoreIndex.from_documents(documents)
+    index.storage_context.persist(persist_dir=constants.PERSIST_DIR)
+
+
+def handle_query(query):
+    """
+    Handle the query and return the response
+    """
+    storage_context = StorageContext.from_defaults(persist_dir=constants.PERSIST_DIR)
+    index = load_index_from_storage(storage_context)
+    text_qa_template = ChatPromptTemplate.from_messages(constants.LLM_RESPONSE_FORMAT)
+    query_engine = index.as_query_engine(text_qa_template=text_qa_template)
+
+    answer = query_engine.query(query)
+    if hasattr(answer, "response"):
+        return answer.response
+    if isinstance(answer, dict) and "response" in answer:
+        return answer["response"]
+    return "Sorry, I couldn't find an answer."
+
+
+if __name__ == "__main__":
+    data_ingestion()
+
+    # --- Launch the chatbot interface ---
+    demo = gr.Interface(
+        fn=handle_query,
+        inputs="text",
+        outputs="text",
+        title="LLM Chatbot",
+        flagging_mode="never",
+    )
+    demo.launch()
+    # -------------------------------------

requirements.txt

@@ -0,0 +1,6 @@
+gradio
+llama_index
+llama-index-llms-huggingface
+llama-index-llms-huggingface-api
+llama-index-embeddings-huggingface
+llama-index-embeddings-instructor

zigistry/constants.py

@@ -0,0 +1,39 @@
+import os
+from dotenv import load_dotenv
+
+print("Loaded environment variables.")
+load_dotenv()
+
+PERSIST_DIR = "PERSIST_DIR"
+LLM = "google/gemma-1.1-7b-it"
+EMBEDDER = "BAAI/bge-small-en-v1.5"
+HF_TOKEN = os.getenv("HF_TOKEN")
+TEMPERATURE = 0.1
+FILES = ["Zigistry.pdf"]
+
+LLM_RESPONSE_FORMAT = [
+    (
+        "user",
+        """You are Zigister - Q&A Assistant
+
+You are Zigister, a Q&A assistant created by Zigistry. Your primary role is to provide accurate and relevant answers based on the given context and instructions.
+
+Remember to always provide clear and concise responses to user queries. If you are unsure about an answer, it is better to admit it than to provide incorrect information.
+
+When asked on data, try responding with proper github links whenever possible.
+
+Remember, https://github.com/ + the full_name is the github link.
+
+When asked about your creator or organization, respond with:
+
+"I was created by Zigistry.dev, A place where you can find all the libraries and programs that suits your Zig lang needs, making it more accessible and easier to use for all."
+Also, zigistry.dev is the website with the latest data for packages and programs.
+
+For all other inquiries, ensure your responses align with the provided context. If a question falls outside your scope, politely guide the user to ask within the relevant context.
+        Context:
+        {context_str}
+        Question:
+        {query_str}
+""",
+    )
+]

zigistry/pre_requisite.py

@@ -0,0 +1,21 @@
+from zigistry import constants
+import os
+from logging import error
+
+
+def CorrectlyCompletedPreRequisites() -> bool:
+    for i in constants.FILES:
+        if not os.path.exists(i):
+            error(f"File {i} does not exist")
+            return False
+    if not os.path.exists(constants.PERSIST_DIR):
+        error(f"Directory {constants.PERSIST_DIR} does not exist")
+        return False
+    if constants.HF_TOKEN == None:
+        error("Hugging Face token not set")
+        return False
+    return True
+
+
+def performPreRequisites():
+    os.makedirs(constants.PERSIST_DIR, exist_ok=True)