PlanOpticon

planopticon / CLAUDE.md

Source Rendered

Blame History Raw 64 lines

1	`# PlanOpticon`
2
3	`Video analysis and knowledge extraction CLI. Processes recordings into structured knowledge graphs with entities, relationships, and insights.`
4
5	`## Knowledge Graph Query Skill`
6
7	PlanOpticon can build and query knowledge graphs from video content. If you see `knowledge_graph.db` or `knowledge_graph.json` files in the workspace, you can query them to understand what was discussed.
8
9	`### Auto-detection`
10
11	`Look for these files (checked automatically):`
12	- `knowledge_graph.db` — SQLite graph database (preferred)
13	- `knowledge_graph.json` — JSON export (fallback)
14
15	Common locations: project root, `results/`, `output/`, `knowledge-base/`.
16
17	`### Quick commands`
18
19	```bash
20	`# Show graph stats (entity/relationship counts)`
21	`planopticon query`
22
23	`# List entities filtered by type`
24	`planopticon query "entities --type technology"`
25	`planopticon query "entities --type person"`
26
27	`# Search entities by name`
28	`planopticon query "entities --name python"`
29
30	`# See what connects to an entity`
31	`planopticon query "neighbors Alice"`
32
33	`# List relationships`
34	`planopticon query "relationships --source Alice"`
35
36	`# Natural language (requires API key)`
37	`planopticon query "What technologies were discussed?"`
38	`planopticon query "Who are the key people mentioned?"`
39
40	`# Output as JSON or Mermaid diagram`
41	`planopticon query --format json stats`
42	`planopticon query --format mermaid "neighbors Alice"`
43
44	`# Interactive REPL`
45	`planopticon query -I`
46	```
47
48	`### When to use`
49
50	- Direct mode (`stats`, `entities`, `neighbors`, `relationships`): No API key needed. Fast, deterministic. Use for structured lookups.
51	- Agentic mode (natural language questions): Requires an API key (`OPENAI_API_KEY`, `ANTHROPIC_API_KEY`, etc.). Use when the user asks open-ended questions about the content.
52
53	`### Python API`
54
55	```python
56	`from video_processor.integrators.graph_query import GraphQueryEngine`
57	`from video_processor.integrators.graph_discovery import find_nearest_graph`
58
59	`path = find_nearest_graph()`
60	`engine = GraphQueryEngine.from_db_path(path)`
61	`result = engine.stats()`
62	`print(result.to_text())`
63	```
64

PlanOpticon

Keyboard Shortcuts