fix(bootstrap): add Qdrant collection to registry with Qwen3 default

[POWERFULMOVES]

Summary

• Add qdrant service to bootstrap registry with QDRANT_COLLECTION and QDRANT_URL
• Default collection: pmoves_chunks_qwen3 (3072d Qwen3-4b embeddings)
• Ensures make brand-defaults sets the correct collection for the standardized embedding stack
• Root cause: extract-worker was using legacy pmoves_chunks (384d) causing dimension mismatch with TensorZero Qwen3 embeddings

Test plan

• JSON validates
• After merge: make brand-defaults sets QDRANT_COLLECTION=pmoves_chunks_qwen3 in env.shared
• Extract-worker recreate picks up new collection

Generated with Claude Code

Summary by CodeRabbit

• Chores
  • Added Qdrant Vector Database service configuration with support for customizable collection and connection settings via environment variables.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

Run ID: e8a96f5e-bd44-4e6d-b64b-08e603655b50

📥 Commits

Reviewing files that changed from the base of the PR and between 46df243 and ce615a5.

📒 Files selected for processing (1)

• pmoves/bootstrap/registry.json

---

📝 Walkthrough

Walkthrough

A new Qdrant Vector Database service is added to the bootstrap registry with two environment variables: QDRANT_URL (defaulting to http://qdrant:6333) and QDRANT_COLLECTION (defaulting to pmoves_chunks_qwen3 for Qwen3-4b embeddings).

Changes

Cohort / File(s)	Summary
Qdrant Service Registration pmoves/bootstrap/registry.json, pmoves/env.shared	Added Qdrant Vector Database service entry with id "qdrant" and two new environment variable declarations: QDRANT_URL (url-typed, http endpoint) and QDRANT_COLLECTION (collection name for vector storage).

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~8 minutes

Possibly related PRs

• feat(embeddings): standardize on Qwen3-4b embedding model #1086: Converges on the same QDRANT_COLLECTION configuration value (pmoves_chunks_qwen3), indicating coordinated integration of Qdrant vector database functionality.

Poem

🐰 A vector hop through time and space,
Qdrant joins the registry race,
Collections nested, URLs set,
Configuration—no regret! ✨

🚥 Pre-merge checks | ✅ 2 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Description check	⚠️ Warning	The description covers the summary and test plan, but is missing required sections: Testing commands/output documentation and Required Checks checklist.	Add the Testing section with actual commands run, and complete the Required Checks checklist with CHIT Contract Check and documentation update confirmations.

✅ Passed checks (2 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title accurately and specifically describes the main change: adding Qdrant with Qwen3 default collection to the bootstrap registry.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch fix/qdrant-collection-bootstrap

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: ce615a55cd

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

Align Qdrant default with active embedding pipeline

Setting QDRANT_COLLECTION to pmoves_chunks_qwen3 as a global bootstrap default forces all compose services that use ${QDRANT_COLLECTION:-...} to share that collection, overriding their per-service defaults (for example pmoves/docker-compose.yml for extract-worker and hi-rag-gateway-v2). Those services can still generate non-Qwen vectors by default/fallback (pmoves/libs/providers/embedding.py defaults to gemma embedding models, and pmoves/services/hi-rag-gateway-v2/app.py falls back to all-MiniLM-L6-v2), which leads to dimension mismatch at query/ingest time and can trigger failures or collection recreation in ensure_qdrant_collection (app.py around lines 466-483). This makes a fresh bootstrap susceptible to broken retrieval unless additional model env vars are updated in lockstep.

Useful? React with 👍 / 👎.