Description
The project currently relies on TigerGraph Cloud (using the pyTigerGraph library) to persist and query city graph data. While TigerGraph is a powerful graph database, relying on it for local development creates several friction points for contributors:
- It requires credentials/tokens for TigerGraph Cloud (the hardcoded token in
config.py is set to expire).
- Local testing is dependent on external internet access and server uptime of the TigerGraph Cloud instance.
- Setup is complex and prone to permission/network issues on corporate or institutional networks.
Replacing the database layer (or offering a parallel backend) using a lightweight SQL database like PostgreSQL (with optional PostGIS for location querying) would make the project much easier to set up, run offline, and containerize.
Affected Files
backend_python/config.py — connection settings and db flagsbackend_python/custom_db/tigergraph_client.py— client wrapper class and helper functionsbackend_python/custom_db/__init__.py — database module importsbackend_python/requirements.txt — dependencies list (remove pyTigerGraph, add psycopg2-binary or asyncpg)
-data/load_to_tg.py & data/tigergraph_loader.py — initial mock data ingestion scripts
Expected Behavior
- Support for a relational database backend (like PostgreSQL) to store intersection nodes, edges/segments, and incidents.
- A database configuration flag (e.g.,
USE_POSTGRES or DB_TYPE="postgres") in config.py.
- Relational equivalents for all high-level database helper functions exposed in
custom_db/__init__.py:
get_client()upsert_incident_to_graph(payload: dict)get_zone_data(zone_id: str)update_zone_danger_score(zone_id: str, danger_score: float)
- A data migration script (e.g.,
data/load_to_postgres.py) to seed the PostgreSQL database from local raw JSON graph data.
Suggested Implementation
- Design a relational schema that mimics the graph's nodes (vertices) and edges as tables (e.g.,
intersections and road_segments with fields for danger_score, risk_score, coordinates, and connected nodes).
- Implement a
PostgresClient that inherits from or matches the interface of TigerGraphClient, and use factory pattern or conditional imports in custom_db/__init__.py depending on the active environment config.
- Add setup steps and schema migrations to the README.
Description
The project currently relies on TigerGraph Cloud (using the
pyTigerGraphlibrary) to persist and query city graph data. While TigerGraph is a powerful graph database, relying on it for local development creates several friction points for contributors:config.pyis set to expire).Replacing the database layer (or offering a parallel backend) using a lightweight SQL database like PostgreSQL (with optional PostGIS for location querying) would make the project much easier to set up, run offline, and containerize.
Affected Files
backend_python/config.py— connection settings and db flagsbackend_python/custom_db/tigergraph_client.py— client wrapper class and helper functionsbackend_python/custom_db/__init__.py— database module importsbackend_python/requirements.txt— dependencies list (removepyTigerGraph, addpsycopg2-binaryorasyncpg)-
data/load_to_tg.py&data/tigergraph_loader.py— initial mock data ingestion scriptsExpected Behavior
USE_POSTGRESorDB_TYPE="postgres") inconfig.py.custom_db/__init__.py:get_client()upsert_incident_to_graph(payload: dict)get_zone_data(zone_id: str)update_zone_danger_score(zone_id: str, danger_score: float)data/load_to_postgres.py) to seed the PostgreSQL database from local raw JSON graph data.Suggested Implementation
intersectionsandroad_segmentswith fields fordanger_score,risk_score, coordinates, and connected nodes).PostgresClientthat inherits from or matches the interface ofTigerGraphClient, and use factory pattern or conditional imports incustom_db/__init__.pydepending on the active environment config.