add code for slides

tutorial/app.py

@@ -0,0 +1,135 @@
+from langchain_core.language_models import BaseChatModel
+
+# question = "What are the rat orthologs of human TP53?"
+
+def load_chat_model(model: str) -> BaseChatModel:
+    provider, model_name = model.split("/", maxsplit=1)
+    if provider == "groq":
+        # https://python.langchain.com/docs/integrations/chat/groq/
+        from langchain_groq import ChatGroq
+        return ChatGroq(model_name=model_name, temperature=0)
+    if provider == "openai":
+        # https://python.langchain.com/docs/integrations/chat/openai/
+        from langchain_openai import ChatOpenAI
+        return ChatOpenAI(model_name=model_name, temperature=0)
+    if provider == "ollama":
+        # https://python.langchain.com/docs/integrations/chat/ollama/
+        from langchain_ollama import ChatOllama
+        return ChatOllama(model=model_name, temperature=0)
+    raise ValueError(f"Unknown provider: {provider}")
+
+llm = load_chat_model("groq/llama-3.3-70b-versatile")
+# llm = load_chat_model("openai/gpt-4o-mini")
+# llm = load_chat_model("ollama/mistral")
+
+from langchain_qdrant import QdrantVectorStore
+from langchain_community.embeddings import FastEmbedEmbeddings
+
+vectordb = QdrantVectorStore.from_existing_collection(
+    path="data/qdrant",
+    collection_name="sparql-docs",
+    embedding=FastEmbedEmbeddings(model_name="BAAI/bge-small-en-v1.5"),
+)
+
+retriever = vectordb.as_retriever()
+number_of_docs_retrieved = 5
+
+# retrieved_docs = retriever.invoke(question, k=number_of_docs_retrieved)
+
+from qdrant_client.models import FieldCondition, Filter, MatchValue
+from langchain_core.documents import Document
+
+def retrieve_docs(question: str) -> list[Document]:
+    retrieved_docs = retriever.invoke(
+        question,
+        k=number_of_docs_retrieved,
+        filter=Filter(
+            must=[
+                FieldCondition(
+                    key="metadata.doc_type",
+                    match=MatchValue(value="SPARQL endpoints query examples"),
+                )
+            ]
+        )
+    )
+    retrieved_docs += retriever.invoke(
+        question,
+        k=number_of_docs_retrieved,
+        filter=Filter(
+            must_not=[
+                FieldCondition(
+                    key="metadata.doc_type",
+                    match=MatchValue(value="SPARQL endpoints query examples"),
+                )
+            ]
+        ),
+    )
+    return retrieved_docs
+
+# retrieved_docs = retrieve_docs(question)
+
+# print(f"📚️ Retrieved {len(retrieved_docs)} documents")
+# # print(retrieved_docs)
+# for doc in retrieved_docs:
+#     print(f"{doc.metadata.get('doc_type')} - {doc.metadata.get('endpoint_url')} - {doc.page_content}")
+
+
+from langchain_core.prompts import ChatPromptTemplate
+
+def _format_doc(doc: Document) -> str:
+    """Format a question/answer document to be provided as context to the model."""
+    doc_lang = (
+        "sparql" if "query" in doc.metadata.get("doc_type", "")
+        else "shex" if "schema" in doc.metadata.get("doc_type", "")
+        else ""
+    )
+    return f"<document>\n{doc.page_content} ({doc.metadata.get('endpoint_url')}):\n\n```{doc_lang}\n{doc.metadata.get('answer')}\n```\n</document>"
+
+
+SYSTEM_PROMPT = """You are an assistant that helps users to write SPARQL queries.
+Put the SPARQL query inside a markdown codeblock with the "sparql" language tag, and always add the URL of the endpoint on which the query should be executed in a comment at the start of the query inside the codeblocks.
+Use the queries examples and classes shapes provided in the prompt to derive your answer, don't try to create a query from nothing and do not provide a generic query.
+Try to always answer with one query, if the answer lies in different endpoints, provide a federated query.
+And briefly explain the query.
+Here is a list of documents (reference questions and query answers, classes schema) relevant to the user question that will help you answer the user question accurately:
+{retrieved_docs}
+"""
+prompt_template = ChatPromptTemplate.from_messages([
+        ("system", SYSTEM_PROMPT),
+        ("placeholder", "{messages}"),
+])
+
+# formatted_docs = "\n".join(doc.page_content + "\n" + doc.metadata.get("answer") for doc in retrieved_docs)
+# formatted_docs = f"<documents>\n{'\n'.join(_format_doc(doc) for doc in retrieved_docs)}\n</documents>"
+# prompt_with_context = prompt_template.invoke({
+#     "messages": [("human", question)],
+#     "retrieved_docs": formatted_docs,
+# })
+
+# print(str("\n".join(prompt_with_context.messages)))
+
+# resp = llm.invoke("What are the rat orthologs of human TP53?")
+# print(resp)
+
+# for msg in llm.stream(prompt_with_context):
+#     print(msg.content, end='')
+
+
+import chainlit as cl
+
+@cl.on_message
+async def on_message(msg: cl.Message):
+    retrieved_docs = retrieve_docs(msg.content)
+    formatted_docs = f"<documents>\n{'\n'.join(_format_doc(doc) for doc in retrieved_docs)}\n</documents>"
+    async with cl.Step(name=f"{len(retrieved_docs)} relevant documents") as step:
+        # step.input = msg.content
+        step.output = formatted_docs
+
+    prompt_with_context = prompt_template.invoke({
+        "messages": [("human", msg.content)],
+        "retrieved_docs": formatted_docs,
+    })
+    final_answer = cl.Message(content="")
+    for resp in llm.stream(prompt_with_context):
+        await final_answer.stream_token(resp.content)
+    await final_answer.send()

tutorial/index.py

@@ -0,0 +1,56 @@
+import os
+
+from langchain_qdrant import QdrantVectorStore
+from langchain_community.embeddings import FastEmbedEmbeddings
+from langchain_core.documents import Document
+from sparql_llm import SparqlExamplesLoader, SparqlVoidShapesLoader
+
+
+# List of endpoints that will be used
+endpoints: list[dict[str, str]] = [
+    {
+        # The URL of the SPARQL endpoint from which most informations will be extracted
+        "endpoint_url": "https://sparql.uniprot.org/sparql/",
+        # If VoID description or SPARQL query examples are not available in the endpoint, you can provide a VoID file (local or remote URL)
+        "void_file": "../packages/sparql-llm/tests/void_uniprot.ttl",
+        # "examples_file": "uniprot_examples.ttl",
+    },
+    {
+        "endpoint_url": "https://www.bgee.org/sparql/",
+    },
+    {
+        "endpoint_url": "https://sparql.omabrowser.org/sparql/",
+    }
+]
+
+
+# Get documents from the SPARQL endpoints
+docs: list[Document] = []
+for endpoint in endpoints:
+    print(f"\n  🔎 Getting metadata for {endpoint['endpoint_url']}")
+    queries_loader = SparqlExamplesLoader(
+        endpoint["endpoint_url"],
+        examples_file=endpoint.get("examples_file"),
+        verbose=True,
+    )
+    docs += queries_loader.load()
+
+    void_loader = SparqlVoidShapesLoader(
+        endpoint["endpoint_url"],
+        void_file=endpoint.get("void_file"),
+        verbose=True,
+    )
+    docs += void_loader.load()
+
+os.makedirs('data', exist_ok=True)
+
+vectordb = QdrantVectorStore.from_documents(
+    docs,
+    path="data/qdrant",
+    collection_name="sparql-docs",
+    force_recreate=True,
+    embedding=FastEmbedEmbeddings(
+        model_name="BAAI/bge-small-en-v1.5",
+        # providers=["CUDAExecutionProvider"], # Uncomment this line to use your GPUs
+    ),
+)

tutorial/pyproject.toml

@@ -0,0 +1,19 @@
+[project]
+name = "tutorial-sparql-agent"
+version = "0.0.1"
+requires-python = "==3.12.*"
+
+dependencies = [
+    "sparql-llm >=0.0.4",
+    "langchain >=0.3.14",
+    "langchain-community >=0.3.17",
+    "langchain-openai >=0.3.6",
+    "langchain-groq >=0.2.4",
+    "langchain-ollama >=0.2.3",
+    "langchain-qdrant >=0.2.0",
+    "qdrant-client >=1.13.0",
+    "fastembed >=0.5.1",
+    # "fastembed-gpu >=0.5.1", # Optional GPU support
+    "chainlit",
+    "langgraph >=0.2.73",
+]

tutorial/slides/index.html

@@ -38,7 +38,7 @@
 
     <style>
       /* Improve text size and codeblocks */
-      .reveal section p,pre code {
+      .reveal section p,li,pre code {
         font-size: 0.6em;
       }
       .reveal section h2 {
@@ -55,7 +55,6 @@
       }
       .reveal section a code {
         background: #8b743d;
-        color: #dcdcdc;
       }
     </style>
   </body>

tutorial/slides/public/slides.md

@@ -1,6 +1,12 @@
-## Setup
+## Introduction
+
+In this tutorial, you'll learn how to build an LLM-powered app that assists in writing SPARQL queries, step by step.
+
+As we progress, you'll be provided with code snippets to gradually construct the system. Note that some earlier code may need to be modified or removed to prevent redundancy, ensuring a clean and efficient implementation.
+
+---
 
-This tutorial will guide you through creating an LLM-based app step by step, you will be given pieces of code to build up the system, while we advance through the tutorial you will need to comment out or delete some previous pieces of code to avoid useless repetitions.
+## Setup
 
 [Install `uv`](https://docs.astral.sh/uv/getting-started/installation/) to easily handle dependencies and run scripts
 
@@ -13,6 +19,8 @@ GROQ_API_KEY=gsk_YYY
 OPENAI_API_KEY=sk-proj-YYY
 ```
 
+> You can get a [free API key on groq.com](https://console.groq.com/keys) after logging in with GitHub or Google. This gives you access to [various open-source models](https://groq.com/pricing/) with a limit of 6k tokens per minute.
+
 ---
 
 ## Setup dependencies
@@ -26,7 +34,6 @@ version = "0.0.1"
 requires-python = "==3.12.*"
 dependencies = [
     "sparql-llm >=0.0.4",
-    "langgraph >=0.2.73",
     "langchain >=0.3.14",
     "langchain-community >=0.3.17",
     "langchain-openai >=0.3.6",
@@ -36,6 +43,7 @@ dependencies = [
     "qdrant-client >=1.13.0",
     "fastembed >=0.5.1",
     "chainlit >=2.2.1",
+    "langgraph >=0.2.73",
 ]
 ```
 
@@ -205,6 +213,8 @@ vectordb = QdrantVectorStore.from_documents(
 )
 ```
 
+> To use GPU you will need to replace the `fastembed` dependency with `fastembed-gpu`
+
 ---
 
 ## Provide context to the LLM
@@ -257,8 +267,6 @@ prompt_with_context = prompt_template.invoke({
     "messages": [("human", question)],
     "retrieved_docs": formatted_docs,
 })
-
-
 ```
 
 ---
@@ -272,11 +280,11 @@ from langchain_core.documents import Document
 
 def _format_doc(doc: Document) -> str:
     """Format our question/answer document to be provided as context to the model."""
-    doc_lang = ""
-    if "query" in doc.metadata.get("doc_type", ""):
-        doc_lang = "sparql"
-    elif "schema" in doc.metadata.get("doc_type", ""):
-        doc_lang = "shex"
+    doc_lang = (
+        "sparql" if "query" in doc.metadata.get("doc_type", "")
+        else "shex" if "schema" in doc.metadata.get("doc_type", "")
+        else ""
+    )
     return f"<document>\n{doc.page_content} ({doc.metadata.get('endpoint_url', '')}):\n\n```{doc_lang}\n{doc.metadata.get('answer')}\n```\n</document>"
 
 formatted_docs = f"<documents>\n{'\n'.join(_format_doc(doc) for doc in retrieved_docs)}\n</documents>"
@@ -335,17 +343,14 @@ async def on_message(msg: cl.Message):
     retrieved_docs = retrieve_docs(msg.content)
     formatted_docs = f"<documents>\n{'\n'.join(_format_doc(doc) for doc in retrieved_docs)}\n</documents>"
     async with cl.Step(name=f"{len(retrieved_docs)} relevant documents") as step:
-        step.input = msg.content
         step.output = formatted_docs
-
     prompt_with_context = prompt_template.invoke({
         "messages": [("human", msg.content)],
         "retrieved_docs": formatted_docs,
     })
     final_answer = cl.Message(content="")
     for resp in llm.stream(prompt_with_context):
-        if resp.content:
-            await final_answer.stream_token(resp.content)
+        await final_answer.stream_token(resp.content)
     await final_answer.send()
 ```
 
@@ -359,4 +364,8 @@ uv run chainlit run app.py
 
 ## Creating more complex "agents"
 
-e.g. reactive agent that can loop over themselves and use tools using [LangGraph](https://langchain-ai.github.io/langgraph/#)
+e.g. reactive agent that can loop over themselves using [LangGraph](https://langchain-ai.github.io/langgraph/#):
+
+- To validate a generated query
+- To use tools
+