Skip to content

The agent engineering platform.

update llm graph transformer documentation (#27905)

Tomaz Bratanica3bbbe6a86ca
diff --git a/docs/docs/how_to/graph_constructing.ipynb b/docs/docs/how_to/graph_constructing.ipynb
index e9b6a01417..79b9e1463f 100644
--- a/docs/docs/how_to/graph_constructing.ipynb
+++ b/docs/docs/how_to/graph_constructing.ipynb
@@ -44,6 +44,9 @@
"name": "stdout",
"output_type": "stream",
"text": [
+ "\n",
+ "\u001b[1m[\u001b[0m\u001b[34;49mnotice\u001b[0m\u001b[1;39;49m]\u001b[0m\u001b[39;49m A new release of pip is available: \u001b[0m\u001b[31;49m24.0\u001b[0m\u001b[39;49m -> \u001b[0m\u001b[32;49m24.3.1\u001b[0m\n",
+ "\u001b[1m[\u001b[0m\u001b[34;49mnotice\u001b[0m\u001b[1;39;49m]\u001b[0m\u001b[39;49m To update, run: \u001b[0m\u001b[32;49mpip install --upgrade pip\u001b[0m\n",
"Note: you may need to restart the kernel to use updated packages.\n"
]
}
@@ -105,7 +108,7 @@
"os.environ[\"NEO4J_USERNAME\"] = \"neo4j\"\n",
"os.environ[\"NEO4J_PASSWORD\"] = \"password\"\n",
"\n",
- "graph = Neo4jGraph()"
+ "graph = Neo4jGraph(refresh_schema=False)"
]
},
{
@@ -149,8 +152,8 @@
"name": "stdout",
"output_type": "stream",
"text": [
- "Nodes:[Node(id='Marie Curie', type='Person'), Node(id='Pierre Curie', type='Person'), Node(id='University Of Paris', type='Organization')]\n",
- "Relationships:[Relationship(source=Node(id='Marie Curie', type='Person'), target=Node(id='Pierre Curie', type='Person'), type='MARRIED'), Relationship(source=Node(id='Marie Curie', type='Person'), target=Node(id='University Of Paris', type='Organization'), type='PROFESSOR')]\n"
+ "Nodes:[Node(id='Marie Curie', type='Person', properties={}), Node(id='Pierre Curie', type='Person', properties={}), Node(id='University Of Paris', type='Organization', properties={})]\n",
+ "Relationships:[Relationship(source=Node(id='Marie Curie', type='Person', properties={}), target=Node(id='Pierre Curie', type='Person', properties={}), type='MARRIED', properties={}), Relationship(source=Node(id='Marie Curie', type='Person', properties={}), target=Node(id='University Of Paris', type='Organization', properties={}), type='PROFESSOR', properties={})]\n"
]
}
],
@@ -191,8 +194,8 @@
"name": "stdout",
"output_type": "stream",
"text": [
- "Nodes:[Node(id='Marie Curie', type='Person'), Node(id='Pierre Curie', type='Person'), Node(id='University Of Paris', type='Organization')]\n",
- "Relationships:[Relationship(source=Node(id='Marie Curie', type='Person'), target=Node(id='Pierre Curie', type='Person'), type='SPOUSE'), Relationship(source=Node(id='Marie Curie', type='Person'), target=Node(id='University Of Paris', type='Organization'), type='WORKED_AT')]\n"
+ "Nodes:[Node(id='Marie Curie', type='Person', properties={}), Node(id='Pierre Curie', type='Person', properties={}), Node(id='University Of Paris', type='Organization', properties={})]\n",
+ "Relationships:[Relationship(source=Node(id='Marie Curie', type='Person', properties={}), target=Node(id='Pierre Curie', type='Person', properties={}), type='SPOUSE', properties={}), Relationship(source=Node(id='Marie Curie', type='Person', properties={}), target=Node(id='University Of Paris', type='Organization', properties={}), type='WORKED_AT', properties={})]\n"
]
}
],
@@ -209,6 +212,44 @@
"print(f\"Relationships:{graph_documents_filtered[0].relationships}\")"
]
},
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "To define the graph schema more precisely, consider using a three-tuple approach for relationships. In this approach, each tuple consists of three elements: the source node, the relationship type, and the target node."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 7,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "Nodes:[Node(id='Marie Curie', type='Person', properties={}), Node(id='Pierre Curie', type='Person', properties={}), Node(id='University Of Paris', type='Organization', properties={})]\n",
+ "Relationships:[Relationship(source=Node(id='Marie Curie', type='Person', properties={}), target=Node(id='Pierre Curie', type='Person', properties={}), type='SPOUSE', properties={}), Relationship(source=Node(id='Marie Curie', type='Person', properties={}), target=Node(id='University Of Paris', type='Organization', properties={}), type='WORKED_AT', properties={})]\n"
+ ]
+ }
+ ],
+ "source": [
+ "allowed_relationships = [\n",
+ " (\"Person\", \"SPOUSE\", \"Person\"),\n",
+ " (\"Person\", \"NATIONALITY\", \"Country\"),\n",
+ " (\"Person\", \"WORKED_AT\", \"Organization\"),\n",
+ "]\n",
+ "\n",
+ "llm_transformer_tuple = LLMGraphTransformer(\n",
+ " llm=llm,\n",
+ " allowed_nodes=[\"Person\", \"Country\", \"Organization\"],\n",
+ " allowed_relationships=allowed_relationships,\n",
+ ")\n",
+ "llm_transformer_tuple = llm_transformer_filtered.convert_to_graph_documents(documents)\n",
+ "print(f\"Nodes:{graph_documents_filtered[0].nodes}\")\n",
+ "print(f\"Relationships:{graph_documents_filtered[0].relationships}\")"
+ ]
+ },
{
"cell_type": "markdown",
"metadata": {},
@@ -229,15 +270,15 @@
},
{
"cell_type": "code",
- "execution_count": 7,
+ "execution_count": 8,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
- "Nodes:[Node(id='Marie Curie', type='Person', properties={'born_year': '1867'}), Node(id='Pierre Curie', type='Person'), Node(id='University Of Paris', type='Organization')]\n",
- "Relationships:[Relationship(source=Node(id='Marie Curie', type='Person'), target=Node(id='Pierre Curie', type='Person'), type='SPOUSE'), Relationship(source=Node(id='Marie Curie', type='Person'), target=Node(id='University Of Paris', type='Organization'), type='WORKED_AT')]\n"
+ "Nodes:[Node(id='Marie Curie', type='Person', properties={'born_year': '1867'}), Node(id='Pierre Curie', type='Person', properties={}), Node(id='University Of Paris', type='Organization', properties={}), Node(id='Poland', type='Country', properties={}), Node(id='France', type='Country', properties={})]\n",
+ "Relationships:[Relationship(source=Node(id='Marie Curie', type='Person', properties={}), target=Node(id='Poland', type='Country', properties={}), type='NATIONALITY', properties={}), Relationship(source=Node(id='Marie Curie', type='Person', properties={}), target=Node(id='France', type='Country', properties={}), type='NATIONALITY', properties={}), Relationship(source=Node(id='Marie Curie', type='Person', properties={}), target=Node(id='Pierre Curie', type='Person', properties={}), type='SPOUSE', properties={}), Relationship(source=Node(id='Marie Curie', type='Person', properties={}), target=Node(id='University Of Paris', type='Organization', properties={}), type='WORKED_AT', properties={})]\n"
]
}
],
@@ -264,12 +305,71 @@
},
{
"cell_type": "code",
- "execution_count": 8,
+ "execution_count": 9,
"metadata": {},
"outputs": [],
"source": [
"graph.add_graph_documents(graph_documents_props)"
]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "Most graph databases support indexes to optimize data import and retrieval. Since we might not know all the node labels in advance, we can handle this by adding a secondary base label to each node using the `baseEntityLabel` parameter."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 10,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "graph.add_graph_documents(graph_documents, baseEntityLabel=True)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "Results will look like:\n",
+ "\n",
+ "![graph_construction3.png](../../static/img/graph_construction3.png)\n",
+ "\n",
+ "The final option is to also import the source documents for the extracted nodes and relationships. This approach lets us track which documents each entity appeared in."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 11,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "graph.add_graph_documents(graph_documents, include_source=True)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "Graph will have the following structure:\n",
+ "\n",
+ "![graph_construction4.png](../../static/img/graph_construction4.png)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "In this visualization, the source document is highlighted in blue, with all entities extracted from it connected by `MENTIONS` relationships."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {},
+ "outputs": [],
+ "source": []
}
],
"metadata": {
@@ -288,7 +388,7 @@
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
- "version": "3.10.1"
+ "version": "3.11.5"
}
},
"nbformat": 4,
diff --git a/docs/static/img/graph_construction3.png b/docs/static/img/graph_construction3.png
new file mode 100644
index 0000000000..3d89b1162e
Binary files /dev/null and b/docs/static/img/graph_construction3.png differ
diff --git a/docs/static/img/graph_construction4.png b/docs/static/img/graph_construction4.png
new file mode 100644
index 0000000000..4b892cd652
Binary files /dev/null and b/docs/static/img/graph_construction4.png differ