Merge pull request #197 from wiltshirek/main

Neo4J integration.

.gitignore

@@ -5,4 +5,8 @@ book.txt
 lightrag-dev/
 .idea/
 dist/
+env/
+local_neo4jWorkDir/
+neo4jWorkDir/
+ignore_this.txt
 .venv/

Dockerfile

@@ -0,0 +1,56 @@
+FROM debian:bullseye-slim
+ENV JAVA_HOME=/opt/java/openjdk
+COPY --from=eclipse-temurin:17 $JAVA_HOME $JAVA_HOME
+ENV PATH="${JAVA_HOME}/bin:${PATH}" \
+    NEO4J_SHA256=7ce97bd9a4348af14df442f00b3dc5085b5983d6f03da643744838c7a1bc8ba7 \
+    NEO4J_TARBALL=neo4j-enterprise-5.24.2-unix.tar.gz \
+    NEO4J_EDITION=enterprise \
+    NEO4J_HOME="/var/lib/neo4j" \
+    LANG=C.UTF-8
+ARG NEO4J_URI=https://dist.neo4j.org/neo4j-enterprise-5.24.2-unix.tar.gz
+
+RUN addgroup --gid 7474 --system neo4j && adduser --uid 7474 --system --no-create-home --home "${NEO4J_HOME}" --ingroup neo4j neo4j
+
+COPY ./local-package/* /startup/
+
+RUN apt update \
+    && apt-get install -y curl gcc git jq make procps tini wget \
+    && curl --fail --silent --show-error --location --remote-name ${NEO4J_URI} \
+    && echo "${NEO4J_SHA256}  ${NEO4J_TARBALL}" | sha256sum -c --strict --quiet \
+    && tar --extract --file ${NEO4J_TARBALL} --directory /var/lib \
+    && mv /var/lib/neo4j-* "${NEO4J_HOME}" \
+    && rm ${NEO4J_TARBALL} \
+    && sed -i 's/Package Type:.*/Package Type: docker bullseye/' $NEO4J_HOME/packaging_info \
+    && mv /startup/neo4j-admin-report.sh "${NEO4J_HOME}"/bin/neo4j-admin-report \
+    && mv "${NEO4J_HOME}"/data /data \
+    && mv "${NEO4J_HOME}"/logs /logs \
+    && chown -R neo4j:neo4j /data \
+    && chmod -R 777 /data \
+    && chown -R neo4j:neo4j /logs \
+    && chmod -R 777 /logs \
+    && chown -R neo4j:neo4j "${NEO4J_HOME}" \
+    && chmod -R 777 "${NEO4J_HOME}" \
+    && chmod -R 755 "${NEO4J_HOME}/bin" \
+    && ln -s /data "${NEO4J_HOME}"/data \
+    && ln -s /logs "${NEO4J_HOME}"/logs \
+    && git clone https://github.com/ncopa/su-exec.git \
+    && cd su-exec \
+    && git checkout 4c3bb42b093f14da70d8ab924b487ccfbb1397af \
+    && echo d6c40440609a23483f12eb6295b5191e94baf08298a856bab6e15b10c3b82891 su-exec.c | sha256sum -c \
+    && echo 2a87af245eb125aca9305a0b1025525ac80825590800f047419dc57bba36b334 Makefile | sha256sum -c \
+    && make \
+    && mv /su-exec/su-exec /usr/bin/su-exec \
+    && apt-get -y purge --auto-remove curl gcc git make \
+    && rm -rf /var/lib/apt/lists/* /su-exec
+
+
+ENV PATH "${NEO4J_HOME}"/bin:$PATH
+
+WORKDIR "${NEO4J_HOME}"
+
+VOLUME /data /logs
+
+EXPOSE 7474 7473 7687
+
+ENTRYPOINT ["tini", "-g", "--", "/startup/docker-entrypoint.sh"]
+CMD ["neo4j"]

README.md

@@ -161,6 +161,39 @@ rag = LightRAG(
 ```
 </details>
 
+
+<details>
+<summary> Using Neo4J for Storage </summary>
+
+* For production level scenarios you will most likely want to leverage an enterprise solution
+* for KG storage. Running Neo4J in Docker is recommended for seamless local testing.  
+* See: https://hub.docker.com/_/neo4j
+
+
+```python
+export NEO4J_URI="neo4j://localhost:7687"
+export NEO4J_USERNAME="neo4j"
+export NEO4J_PASSWORD="password"
+
+When you launch the project be sure to override the default KG: NetworkS
+by specifying kg="Neo4JStorage".
+
+# Note: Default settings use NetworkX
+#Initialize LightRAG with Neo4J implementation. 
+WORKING_DIR = "./local_neo4jWorkDir"
+
+rag = LightRAG(
+    working_dir=WORKING_DIR,
+    llm_model_func=gpt_4o_mini_complete,  # Use gpt_4o_mini_complete LLM model
+    kg="Neo4JStorage", #<-----------override KG default
+    log_level="DEBUG"  #<-----------override log_level default
+)
+```
+see test_neo4j.py for a working example.
+</details>
+
+
+
 <details>
 <summary> Using Ollama Models </summary>
 

get_all_edges_nx.py

@@ -0,0 +1,34 @@
+import networkx as nx
+
+G = nx.read_graphml('./dickensTestEmbedcall/graph_chunk_entity_relation.graphml')
+
+def get_all_edges_and_nodes(G):
+    # Get all edges and their properties
+    edges_with_properties = []
+    for u, v, data in G.edges(data=True):
+        edges_with_properties.append({
+            'start': u,
+            'end': v,
+            'label': data.get('label', ''),  # Assuming 'label' is used for edge type
+            'properties': data,
+            'start_node_properties': G.nodes[u],
+            'end_node_properties': G.nodes[v]
+        })
+
+    return edges_with_properties
+
+# Example usage
+if __name__ == "__main__":
+    # Assume G is your NetworkX graph loaded from Neo4j
+
+    all_edges = get_all_edges_and_nodes(G)
+
+    # Print all edges and node properties
+    for edge in all_edges:
+        print(f"Edge Label: {edge['label']}")
+        print(f"Edge Properties: {edge['properties']}")
+        print(f"Start Node: {edge['start']}")
+        print(f"Start Node Properties: {edge['start_node_properties']}")
+        print(f"End Node: {edge['end']}")
+        print(f"End Node Properties: {edge['end_node_properties']}")
+        print("---")