Skip to content

The agent engineering platform.

docs: Improve neo4j semantic templates (#23939)

Nada Amin3f1cf00d97f6
diff --git a/templates/neo4j-semantic-ollama/README.md b/templates/neo4j-semantic-ollama/README.md
index bfe52dea24..552f5552ea 100644
--- a/templates/neo4j-semantic-ollama/README.md
+++ b/templates/neo4j-semantic-ollama/README.md
@@ -39,12 +39,26 @@ NEO4J_USERNAME=<YOUR_NEO4J_USERNAME>
NEO4J_PASSWORD=<YOUR_NEO4J_PASSWORD>
```
+Typically for a local Ollama installation:
+
+```shell
+export OLLAMA_BASE_URL="http://127.0.0.1:11434"
+```
+
## Populating with data
If you want to populate the DB with an example movie dataset, you can run `python ingest.py`.
The script import information about movies and their rating by users.
Additionally, the script creates two [fulltext indices](https://neo4j.com/docs/cypher-manual/current/indexes-for-full-text-search/), which are used to map information from user input to the database.
+As an aternative, you can use the demo neo4j recommendations database:
+```shell
+export NEO4J_URI="neo4j+s://demo.neo4jlabs.com"
+export NEO4J_USERNAME="recommendations"
+export NEO4J_PASSWORD="recommendations"
+export NEO4J_DATABASE="recommendations"
+```
+
## Usage
To use this package, you should first have the LangChain CLI installed:
@@ -65,7 +79,7 @@ If you want to add this to an existing project, you can just run:
langchain app add neo4j-semantic-ollama
```
-And add the following code to your `server.py` file:
+And, from within the project, add the following code to your `app/server.py` file, replacing the `add_routes(app, NotImplemented)` section:
```python
from neo4j_semantic_ollama import agent_executor as neo4j_semantic_agent
@@ -83,7 +97,7 @@ export LANGCHAIN_API_KEY=<your-api-key>
export LANGCHAIN_PROJECT=<your-project> # if not specified, defaults to "default"
```
-If you are inside this directory, then you can spin up a LangServe instance directly by:
+If you are inside the top-level project directory, then you can spin up a LangServe instance directly by:
```shell
langchain serve