Procedure Finder

Operators find the right procedure from a description.

Industry	Level	Difficulty	Recommended mode
Energy & utilities	1 of 5	Beginner	flat

Coming soon

A complete, runnable Procedure Finder reference implementation is in progress. The architecture and core code below already work with SrvDB 0.3.

The problem

• Procedures are numerous.
• Control rooms need speed.
• Field sites lack connectivity.

How SrvDB powers Procedure Finder

1. Procedures embedded by step.
2. Descriptions retrieve procedures.
3. Offline copies for field use.
4. Updates via upsert.

What changes

Today	With Procedure Finder on SrvDB
Binder searches	Instant procedures
Connectivity required	Offline
Outdated copies	Current only

How it works

1. Ingest procedures.
2. Embed.
3. Search.
4. Distribute snapshots.
5. Update.

SrvDB features it uses

• Exact search (flat): every query compares against every stored vector, so results are exact and explainable.
• upsert: inserts new items and replaces changed ones, so the knowledge base stays current.
• Fully offline: SrvDB runs in-process and never opens a network connection.
• Crash-safe persist(): after it returns, data survives a crash or power loss and reopens exactly as committed.

Use cases

• Switching procedures
• Plant start-up
• Emergency steps
• Inspection routines
• Safety rules
• Training

Core code

embed() is your embedding model (for example sentence-transformers, or a function that turns your records into feature vectors).

import json
import srvdb

db = srvdb.Database("./procedure-finder-db", dimension=384)

def ingest(items):
    """items: list of (id, text, metadata dict)"""
    db.upsert(
        ids=[i for i, _, _ in items],
        embeddings=[embed(t) for _, t, _ in items],
        metadatas=[json.dumps(m) for _, _, m in items],
    )
    db.persist()                      # crash-safe checkpoint

def find_procedure(description, k=5):
    hits = db.search(embed(description), k)
    return [(json.loads(db.get(i)), round(score, 3)) for i, score in hits]

Limits and honest notes

• Official procedures remain authoritative.

Build roadmap

1. Prototype: load a sample of real data with flat mode and check that searches return what an expert expects.
2. Measure: if you move to an approximate mode, compare its results with flat on the same queries (recall).
3. Pipeline: add continuous upsert and decide when to call persist().
4. Harden: test restarts and crashes, add logging, review every output path with domain experts.
5. Deploy: ship the database directory with the application or service.

All Energy & utilities projects

Level	Project	What it does
1 · Beginner	Procedure Finder (this page)	Operators find the right procedure from a description.
2 · Intermediate	Outage Analyzer	Finds similar past outages and their causes.
3 · Advanced	Inspection Image Matcher	Matches drone and field inspection images to known defects.
4 · Expert	Field Operations Agent	An offline agent for crews with procedures, asset history and notes.
5 · Frontier	Grid Analog Forecaster	Finds historical analogues of current weather and load patterns.

See also: Use cases · Live examples (coming soon)