feat: init project

app.py

@@ -12,65 +12,49 @@ from llama_index.llms.llama_cpp.llama_utils import (
 from llama_index.core.memory import ChatMemoryBuffer
 import nltk
 
-
-def import_documents():
-    urls = [
-        "https://www.zatsit.fr/",
-        "https://www.zatsit.fr/collaborer-avec-zatsit/",
-        "https://fr.linkedin.com/company/zatsit",
-        "https://www.zatsit.fr/contact/",
-        "https://blog.zatsit.fr/blog/green-exploitation-miniere",
-        "https://blog.zatsit.fr/blog/bundlephobia-reduire-javascript",
-        "https://blog.zatsit.fr/blog/gemini-vertex-ai",
-        "https://blog.zatsit.fr/blog/asyncapi-3-is-out",
-        "https://blog.zatsit.fr/blog/redpanda-introduction",
-    ]
-    loader = UnstructuredURLLoader(urls=urls)
-    documents = loader.load_data()
-    return documents
-
-
-def create_embed_model():
-    # create embed model from HuggingFace
-    return HuggingFaceEmbedding(model_name="thenlper/gte-large")
-
-
-def create_store_index(documents, embed_model):
-    # create vector store index
-    return VectorStoreIndex.from_documents(documents, embed_model=embed_model)
-
-
-def create_llm():
-    model_url = "https://huggingface.co/TheBloke/Mistral-7B-Instruct-v0.2-code-ft-GGUF/resolve/main/mistral-7b-instruct-v0.2-code-ft.Q2_K.gguf"
-    return LlamaCPP(
-        # You can pass in the URL to a GGML model to download it automatically
-        model_url=model_url,
-        # optionally, you can set the path to a pre-downloaded model instead of model_url
-        model_path=None,
-        temperature=0.1,
-        max_new_tokens=256,
-        # llama2 has a context window of 4096 tokens, but we set it lower to allow for some wiggle room
-        context_window=3900,
-        # kwargs to pass to __call__()
-        generate_kwargs={},
-        # kwargs to pass to __init__()
-        # set to at least 1 to use GPU
-        model_kwargs={"n_gpu_layers": 1},
-        # transform inputs into Llama2 format
-        messages_to_prompt=messages_to_prompt,
-        completion_to_prompt=completion_to_prompt,
-        verbose=True,
-    )
-
-
 # download punkt
 nltk.download('punkt')
 nltk.download('punkt_tab')
 
-documents = import_documents()
-embed_model = create_embed_model()
-index = create_store_index(documents=documents, embed_model=embed_model)
-llm = create_llm()
+urls = [
+    "https://www.zatsit.fr/",
+    "https://www.zatsit.fr/collaborer-avec-zatsit/",
+    "https://fr.linkedin.com/company/zatsit",
+    "https://www.zatsit.fr/contact/",
+    "https://blog.zatsit.fr/blog/green-exploitation-miniere",
+    "https://blog.zatsit.fr/blog/bundlephobia-reduire-javascript",
+    "https://blog.zatsit.fr/blog/gemini-vertex-ai",
+    "https://blog.zatsit.fr/blog/asyncapi-3-is-out",
+    "https://blog.zatsit.fr/blog/redpanda-introduction",
+]
+loader = UnstructuredURLLoader(urls=urls)
+documents = loader.load_data()
+
+embed_model = HuggingFaceEmbedding(model_name="thenlper/gte-large")
+
+# create vector store index
+index = VectorStoreIndex.from_documents(documents, embed_model=embed_model)
+
+model_url = "https://huggingface.co/TheBloke/Mistral-7B-Instruct-v0.2-code-ft-GGUF/resolve/main/mistral-7b-instruct-v0.2-code-ft.Q2_K.gguf"
+llm = LlamaCPP(
+    # You can pass in the URL to a GGML model to download it automatically
+    model_url=model_url,
+    # optionally, you can set the path to a pre-downloaded model instead of model_url
+    model_path=None,
+    temperature=0.1,
+    max_new_tokens=256,
+    # llama2 has a context window of 4096 tokens, but we set it lower to allow for some wiggle room
+    context_window=3900,
+    # kwargs to pass to __call__()
+    generate_kwargs={},
+    # kwargs to pass to __init__()
+    # set to at least 1 to use GPU
+    model_kwargs={"n_gpu_layers": 1},
+    # transform inputs into Llama2 format
+    messages_to_prompt=messages_to_prompt,
+    completion_to_prompt=completion_to_prompt,
+    verbose=True,
+)
 
 
 def querying(query, history):