Docs/Skills and code/Project documents

README.md

Markdown51 lines2.5 KBProject documents

Mentioned on: page 02, page 14

YouTube Ecosystem Intelligence

A configurable research pipeline for discovering YouTube channels, collecting channel/video evidence, profiling editorial patterns, qualifying channels against a target ecosystem, and ranking them in five complementary research queues.

New ecosystem setup

1. Install Node.js 22 or later.
2. Run npm install.
3. Edit only config.js for the new ecosystem:
   • ecosystem.name and description
   • includeTerms and excludeTerms
   • adjacentTerms
   • discovery.topicSeeds and discovery.adjacentSeeds
   • qualification definitions for core, adjacent, unknown, and irrelevant
4. Optionally replace ecosystem.referenceChannels in config.js. These manual references are optional and are not required for discovery or ranking.
5. Configure the LLM environment variables already used by the project (OLLAMA_HOST, OLLAMA_MODEL) if using LLM stages.
6. Start with a fresh database for a new ecosystem: remove or archive data/ecosystem.db; do not reuse an old ecosystem database unless you intentionally want a mixed corpus.
7. Run npm run discover, then enrich/profile/analyze as needed. For expansion, run npm run discovery:v2. LLM query expansion is opt-in with --llm.
8. Review reports and data health before interpreting rankings. Unknown/missing evidence is not zero.

Typical first pass from the project root:

npm ci
npm run discover
npm run enrich
npm run channel-metadata
npm run profile:channels
npm run profile:videos
npm run video-signals
npm run analyze:corpus
npm run normalize:temporal
npm run vocabulary:build
npm run features:build
npm run videos:embed
npm run priorities:build
npm run qualification:llm
npm run discovery:v2
npm run discovery:v2:enrich
npm run discovery:v2:score
npm run queues:rate

Use --llm only when deliberately enabling LLM-generated expansion hypotheses, for example npm run discovery:v2 -- --llm. Some stages are optional or may need rerunning after new catalogue data is fetched; use each stage's --help where available.

See QUEUE-DEFINITIONS.md for the five queue purposes. npm run test:all runs the local test suite.

Discovery seed dependence

The second discovery phase has deterministic corpus-led routes and config-led topic/adjacent seeds. It can expand without human channel seeds. ecosystem.referenceChannels remains an optional, explicit input only for the reference/inspiration queue; they do not bypass qualification or automatically receive a high score. LLM-generated queries are disabled by default and can be enabled with --llm.

# YouTube Ecosystem Intelligence

A configurable research pipeline for discovering YouTube channels, collecting channel/video evidence, profiling editorial patterns, qualifying channels against a target ecosystem, and ranking them in five complementary research queues.

## New ecosystem setup

1. Install Node.js 22 or later.
2. Run `npm install`.
3. Edit **only `config.js`** for the new ecosystem:
   - `ecosystem.name` and `description`
   - `includeTerms` and `excludeTerms`
   - `adjacentTerms`
   - `discovery.topicSeeds` and `discovery.adjacentSeeds`
   - qualification definitions for `core`, `adjacent`, `unknown`, and `irrelevant`
4. Optionally replace `ecosystem.referenceChannels` in `config.js`. These manual references are optional and are not required for discovery or ranking.
5. Configure the LLM environment variables already used by the project (`OLLAMA_HOST`, `OLLAMA_MODEL`) if using LLM stages.
6. Start with a fresh database for a new ecosystem: remove or archive `data/ecosystem.db`; do not reuse an old ecosystem database unless you intentionally want a mixed corpus.
7. Run `npm run discover`, then enrich/profile/analyze as needed. For expansion, run `npm run discovery:v2`. LLM query expansion is opt-in with `--llm`.
8. Review reports and data health before interpreting rankings. Unknown/missing evidence is not zero.

Typical first pass from the project root:

```bash
npm ci
npm run discover
npm run enrich
npm run channel-metadata
npm run profile:channels
npm run profile:videos
npm run video-signals
npm run analyze:corpus
npm run normalize:temporal
npm run vocabulary:build
npm run features:build
npm run videos:embed
npm run priorities:build
npm run qualification:llm
npm run discovery:v2
npm run discovery:v2:enrich
npm run discovery:v2:score
npm run queues:rate
```

Use `--llm` only when deliberately enabling LLM-generated expansion hypotheses, for example `npm run discovery:v2 -- --llm`. Some stages are optional or may need rerunning after new catalogue data is fetched; use each stage's `--help` where available.

See `QUEUE-DEFINITIONS.md` for the five queue purposes. `npm run test:all` runs the local test suite.

## Discovery seed dependence

The second discovery phase has deterministic corpus-led routes and config-led topic/adjacent seeds. It can expand without human channel seeds. `ecosystem.referenceChannels` remains an optional, explicit input only for the reference/inspiration queue; they do not bypass qualification or automatically receive a high score. LLM-generated queries are disabled by default and can be enabled with `--llm`.

Shown as it is in the repository at build time (2026-11-03); the home directory appears as <home>. If this page and the file differ, the file is right. Back to the catalogue.