Fix/docker compose env vars

[irastorzatobias]

• Fix broken YAML syntax for FALKORDB_MAX_CONNECTIONS and FALKORDB_RETRY_ON_TIMEOUT env vars
• Make Pinecone serverless cloud/region configurable via env vars
• Add defensive import for optional TaskPollResponse
• Update default model from gpt-3.5-turbo to gpt-4.1-nano
• Add vscode to gitignore"

POST: add PINECONE .env vars

---

Note

Medium Risk
Medium risk because it changes the default LLM model used for /api/qa requests and alters Pinecone index-creation parameters via new env vars, which can affect runtime behavior and costs.

Overview
Fixes broken Docker Compose environment variable entries for FALKORDB_MAX_CONNECTIONS and FALKORDB_RETRY_ON_TIMEOUT, and adds a Makefile to standardize common docker compose operations.

Makes Pinecone serverless index creation configurable via PINECONE_SERVERLESS_CLOUD and PINECONE_SERVERLESS_REGION instead of hardcoded values.

Improves modular startup by providing a fallback TaskPollResponse schema when the optional knowledge-graph module isn’t installed, and updates the default QuestionAnswer.model from gpt-3.5-turbo to gpt-4.1-nano. Also ignores VS Code artifacts via .gitignore.

^{Written by Cursor Bugbot for commit ea5b0a5. This will update automatically on new commits. Configure here.}

[cursor]

Cursor Bugbot has reviewed your changes and found 1 potential issue.

^{Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

[cursor]

Inconsistent default model update across sibling classes

Low Severity

The default model in QuestionAnswer was updated to gpt-4.1-nano, but QuestionToLLM in the same file (line 242) still defaults to gpt-3.5-turbo. Both classes are used for user-facing API endpoints (/api/qa/* and /api/ask respectively), and before this change they shared the same default. This creates a new inconsistency where two similar request models silently resolve to different models depending on which endpoint a user hits without specifying a model.