cityops_copilot_agent_memory_ollama.ipynb builds a
field copilot for the inspectors who keep a city’s infrastructure safe: bridges, substations,
pipelines, water-treatment plants and seawalls. When an inspector writes up what they see on
site, the copilot suggests a likely diagnosis and points to that asset’s history. Its memory
spans inspectors and visits, so a new inspector at Harbor Bridge starts from what the previous
one observed, recommended and graded.
It runs on Oracle AI Database 26ai and Oracle AI Agent Memory (oracleagentmemory 26.4), with
in-database ONNX embeddings and llama3.2 on a local Ollama server.
| Part | What you build |
|---|---|
| 1 | the Oracle connection: one converged database for domain data and memory |
| 2 | the in-database embedder and the OracleAgentMemory client |
| 3 | threads, verbatim messages and automatic memory extraction from inspection narratives |
| 4 | structured findings in a VECTOR(384) table with an HNSW index, searched with filters in one query |
| 5 | scoping: private turns per inspector, knowledge shared per asset |
| 6 | call_copilot, end to end, and the cross-inspector handoff |
Appbook: apps/appbooks/cityops_copilot is the interactive
companion. Each Part becomes a page of diagrams that runs this notebook’s code live, with a data
explorer for every row it writes.
| Path | What it is |
|---|---|
cityops_copilot_agent_memory_ollama.ipynb |
the notebook, saved with the outputs of a full run |
data/ |
maintenance_logs.json and inspection_reports.json |
diagrams/ |
the notebook’s diagrams (SVG, with their Mermaid source) |
requirements.txt |
the package versions the notebook was run with |
.env.example |
settings; copy to .env (the appbook reads the same file) |
cd notebooks/agent_memory/cityops_copilot
python3 -m venv .venv && source .venv/bin/activate
pip install -r requirements.txt
cp .env.example .env # then set DBPASSWORD
ollama pull llama3.2
jupyter lab cityops_copilot_agent_memory_ollama.ipynb
The notebook expects an Oracle AI Database 26ai user (prism by default) with the
ALL_MINILM_L12_V2 ONNX embedding model in its schema, on localhost:1522/FREEPDB1. With the
Free container, these steps match how the Oracle LiveLabs image sets up its database:
docker run -d --name oracle-free -p 1522:1521 -e ORACLE_PWD='<password>' \
container-registry.oracle.com/database/free:latest
# once the log says DATABASE IS READY TO USE:
docker exec oracle-free bash -c 'mkdir -p /opt/oracle/dmdump && curl -fL -o /opt/oracle/dmdump/all_MiniLM_L12_v2.onnx \
https://adwc4pm.objectstorage.us-ashburn-1.oci.customer-oci.com/p/iPX9W0MZeRkwJKWdFmdJCemmN-iKAl_bFvNGYLW7YqIrw4kKsukL24J2q93Beb9S/n/adwc4pm/b/OML-ai-models/o/all_MiniLM_L12_v2.onnx'
docker exec -i oracle-free sqlplus -s / as sysdba <<'SQL'
alter session set container = FREEPDB1;
create user prism identified by "<password>" default tablespace users quota unlimited on users;
grant connect, resource, db_developer_role, create mining model, create property graph to prism;
grant execute on dbms_vector to prism;
grant execute on dbms_vector_chain to prism;
create or replace directory dm_dump as '/opt/oracle/dmdump';
grant read, write on directory dm_dump to prism;
alter session set current_schema = prism;
begin
dbms_vector.load_onnx_model(
directory => 'DM_DUMP',
file_name => 'all_MiniLM_L12_v2.onnx',
model_name => 'ALL_MINILM_L12_V2',
metadata => json('{"function":"embedding","embeddingOutput":"embedding","input":{"input":["DATA"]}}'));
end;
/
SQL
Install Ollama and pull llama3.2. The notebook reads OLLAMA_BASE_URL
(default http://localhost:11434).